P R O C E E D I N G S

(8:25:23 a.m.)

DR. NOLLER: Everyone take their seats, please. We have a very full day so I want to get started exactly on time. My name is Ken Noller, and I'd like to call the meeting to order. This is the Meeting of Obstetrics and Gynecology Devices Panel. I request that everyone in attendance please sign in. If you have not done so, please go out and sign in at the front desk now.

I also note for the record that the voting members present constitute a quorum as required by 21 CFR Part 14. I'm going to ask the panel members to introduce themselves. Let's start at this end, please.

MS. MOONEY: Mary Lou Mooney. I'm the Vice President of Clinical Regulatory and Quality for SenoRx, and I'm the Industry Rep to the panel.

MS. LUCKNER: Kleia Luckner, Hospital Administrator, Toledo, Ohio, and I am the Consumer Rep.

DR. D'AGOSTINO: Ralph D'Agostino from Boston University, Biostatistician.

DR. BRILL: Andrew Bill. I am a Professor OB-GYN, University of Illinois.

DR. HILLARD: Paula Hillard, Professor of OB-GYN and Pediatrics, University of Cincinnati.

DR. DIAMOND: Michael Diamond, Professor OB-GYN, Wayne State University, Detroit Michigan.

DR. ROBERTS: Anne Roberts, Professor of Radiology, University of California - San Diego.

DR. NOLLER: I'm Ken Noller, Professor and Chair of Tufts University OB-GYN.

DR. WHANG: I'm Joyce Whang. And I'm an FDA Reviewer and the Executive Secretary for this panel.

DR. BAILEY: I'm Mike Bailey. I'm also a Reviewer in the OB-GYN Devices group, and I'm an Assistant Executive Secretary.

DR. BROWN: Hi. Carol Brown, I'm a Panel Member. I am an Assistant Professor at Cornell Weill Medical College, OB-GYN and a GYN Oncologist at Memorial Sloan-Kettering Cancer Center.

DR. CRUM: I'm Larry Crum from the University of Washington. I'm Director of the Center for Industrial and Medical Ultrasound at the University of Washington.

DR. JANIK: Grace Janik, Clinical Professor at the Medical College of Wisconsin, Reproductive Endocrinologist.

DR. SAMULSKI: Thad Samulski, Duke University Medical Physics.

DR. HAYES: Evelyn Hayes, Professor of Nursing, University of Delaware.

DR. ASCHER: Susan Ascher, Radiologist, Georgetown University Hospital.

DR. WOOD: Bradford Wood, Interventional Radiologist, National Institutes of Health.

MS. BROGDON: I'm Nancy Brogdon. I'm not a member of the panel. I'm the Director of FDA's Division of Reproductive, Abdominal, and Radiological Devices.

DR. SOLOMON: Steve Solomon from Department of Radiology, Johns Hopkins.

DR. NOLLER: Thank you. For the press, the FDA press contact is Colin Pollard who is sitting here in the front row. I don't expect that we'll have any super controversial outbursts today, but we would like everyone to please be courteous, turn off your cell phones, and if you have anything to say, wait until you're recognized and then come to the table. For the audience and the panel members I will recognize people before they speak. Our Executive Secretary has some things to read into the Minutes.

DR. WHANG: There will be OB-GYN Devices Panel on July 26^th and 27^th, so the remaining panel meeting date for this year is October 25^th to 26^th.

We are pleased to introduce a new voting member to this panel, Dr. Paula Hillard of the Department of Obstetrics and Gynecology and the Department of Pediatrics at the University of Cincinnati, College of Medicine.

Today we will have eight temporary voting members, Drs. Ascher, Brill, Crum, D'Agostino, Janik, Roberts, Samulski and Wood. And I will now read into the record the appointments to temporary voting status signed by Daniel Schultz, M.D., the Acting Director for the Center of Devices and Radiological Health.

"Pursuant to the authority granted under the Medical Devices Advisory Committee Charter dated October 27^th, 1990, and amended August 18^th, 1999, I appoint the following individuals as voting members of the Obstetrics and Gynecology Devices Panel for this meeting on June 3^rd, 2004; Susan M. Ascher, M.D., Andrew I. Brill, M.D., Lawrence A. Crum, Ph.D., Ralph B. D'Agostino, Ph.D., Grace M. Janik, M.D., Kenneth E. Najarean, M.D., Anne C. Roberts, M.D., Thaddeus V. Samulski, Ph.D., Bradford J. Wood, M.D.

For the record, these people are special government employees and are consultants to this panel. They have undergone the customary conflict of interest review, and they have reviewed the material to be considered at this meeting."

I will now read the conflict of interest statement for this meeting. "The following announcement addresses conflict of interest issues associated with this meeting, and is made a part of the record to preclude even the appearance of an impropriety. To determine if any conflict existed, the Agency reviewed the submitted agenda, and all financial interests reported by the committee participants. The Conflict of Interest statutes prohibit special government employees from participating in matters that could affect their or their employer's financial interests. However, the Agency has determined that participation of certain members and consultants, the need for whose services outweighs the potential conflict of interest involved is in the best interest of the government. Therefore, full waivers have been granted for Dr. Susan Ascher and Anne Roberts, and limited waivers have been granted for Drs. Michael Diamond and Steven Solomon for their interest in firms that could potentially be affected by the panel's recommendations.

Dr. Ascher's waiver involves a contract to her employer funded for less than $100,000 per year with a competing firm. Dr. Roberts' waiver involves a stockholding in a competing firm in which the value is between $15,001 and $25,000. Dr. Diamond's limited waiver involves a contract to his institution for the sponsor study in which he had no involvement in data generation or analysis, and for which total funding to the institution was less than $100,000. Dr. Solomon's limited waiver involves a contract to his institution for the sponsor study in which he had no involvement in data generation or analysis, and for which funding to the institution is unknown.

The waivers of Dr. Ascher and Dr. Roberts allow them to participate fully in today's deliberations. The limited waivers for Dr. Diamond and Dr. Solomon allow them to participate in the panel discussions, but exclude them from voting.

Copies of these waivers may be obtained from the Agency's Freedom of Information Office, Room 12A-15 of the Parklawn Building. We would like to note for the record that the Agency took into consideration other matters regarding Drs. Diamond and Solomon. They reported current interests with firms at issue, but in matters that are not related to today's agenda. The Agency has determined, therefore, that these individuals may participate fully in the panel's deliberations.

In the event that the discussions involve any other products or firms not already on the agenda for which an FDA participant has a financial interest, the participant should excuse him or herself from such involvement and the exclusion will be noted for the record.

With respect to other participants, we ask in the interest of fairness that all persons making statements or presentations disclose any current or previous financial involvement with any firm whose products they may wish to comment on.

Transcripts for today's meeting are available from Neal R. Gross and Company of Washington, D.C., at (202) 234-4433, and videos are available from FDA Live at (301) 984-0001, or FDA Advisory Committee.com at (800) 627-8171.

Any presenters to the panel who have not already done so should provide FDA with a hard copy of your remarks, including overheads. Michelle Byrnes will collect these from you at the podium.

DR. NOLLER: Thank you. First, I'd like to introduce Colin Pollard, Chief of Obstetrics and Gynecology Devices Branch of the Food and Drug Administration.

DR. POLLARD: Thank you, Dr. Noller, and I just have a few brief comments to kick off our panel meeting today. I want to welcome all the panel member, and thank you very much for coming from near and far.

Today you'll be looking at a PMA for a high-intensity focused ultrasound system, really a new surgical modality that uses conventional MR Imaging for pre-op treatment planning and MR thermal mapping, really a new feature of MR technology for interactive treatment feedback. And treatment of uterine fibroids is the very first indication that's coming before this center in a PMA. The technology, obviously, looks capable of many other clinical applications and the center is currently working on a plan to optimize our regulatory review approach.

As you'll hear later in our presentation, we put together something of a designer review team for this PMA drawing from all parts of our center, especially from the technical side, and as we look around the table here I see several familiar faces, but lot of new faces. And really, we've put together something of a designer panel, as well, and so we're very much looking for your input.

This PMA, the center granted expedited review to based on unique features and advantages. The FDA review is still ongoing, but we consider it quite appropriate at this stage to hear the panel input even as we continue to work our way through many review issues. And finally, that we feel this technology pushes the traditional envelope of clinical management, and if the panel gets to that point, we'll be definitely looking for input regarding training and labeling, and credentialing and that sort of thing.

So with those initial comments, Dr. Noller, I turn it back to you. Thank you.

DR. NOLLER: Thank you. Let me just ask that Drs. Miller and Weeks introduce themselves, please.

DR. MILLER: Hugh Miller from Arizona.

DR. NOLLER: And what do you do, Dr. Miller?

DR. MILLER: I'm a Maternal-Fetal Medicine Specialist.

MR. WEEKS: Jonathan Weeks from Louisville, Kentucky, Maternal-Fetal Medicine, Norton Health Care.

DR. NOLLER: Thank you both. First, let me ask before we open the public hearing, is there anyone present who will be speaking, request speaking at this session of the public hearing? All right. So if there's no one at this time, I will not read the conflict statement then. We'll move right ahead to the presentation by the sponsor.

I'd like to introduce Rob Newman from InSightec. The sponsor has been granted one hour and 15 minutes for their presentations. I ask the panel members to hold all questions until the end of the presentation.

MR. NEWMAN: Good morning, Chairman Noller, and thank you very much, ladies and gentlemen of the panel and the audience. I'm Rob Newman. I'm from InSightec in Dallas, Texas. My trip here has been paid for by my company. I am a member of the sponsor.

I'd like to introduce other members of our team today. Dr. Elizabeth Stewart is an Associate Professor of Gynecology from Harvard Brigham & Women's Hospital. Dr. Clare Tempany is a Professor of Radiology at Brigham & Women's. They are Co-PIs of the study. Dr. Stewart is the Lead PI for the study.

Also, Kobi Vortman, the President of InSightec is here with us today. Karin Coyne, Senior Research Scientist from MEDTAP International, who has helped us with the quality of life work, and some of the biostatistics. Kathy McDermott from MedTrials, our CRO. We also have a guest here, Dr. Bobbie Gostout, who is Assistant Professor from the Mayo Clinic in Rochester, and Dr. Gina Hesley, an Instructor of Radiology, who are Co-PIs from the Mayo Clinical site.

This is an outline of our discussion today. I'll give a brief introduction. Dr. Stewart will talk about, in general, an overview of uterine fibroids and their application here. I'll give a brief overview of the device description. If you had a chance to review the video in the package, I think that will cover some of it, and there was quite a bit of material in the panel package, so I won't go over that in any great detail.

Dr. Clare Tempany will talk about a review of MR anatomy, and what's commonly seen on MR that may be a little bit more than what some of you see in a regular clinical practice. She'll also discuss the treatment development process. Dr. Stewart will talk about clinical design trial results. I'll cover some elements of training, in addition to what was in the panel package, and then Dr. Stewart will summarize.

The indications for use for this device is its for use in pre or peri-menopausal women with symptomatic fibroids. The fibroids to be treated must be visible on non-contrast MRI and should enhance on contrast MR.

Outside the U.S., the system has received CE Mark in Europe in 2002. Its commercially available in Europe, Israel and Japan. In the U.S., the only applications are investigational. We have treated approximately 600 women worldwide for uterine fibroids. And I'd like to introduce Dr. Stewart, who will introduce the topic.

DR. STEWART: Mr. Chairman, panel members and guests, my travel expenses were paid by InSightec today. As Mr. Newman said, I serve as a Clinical Trial Investigator for the company, and am a Consultant for the company, but abide by the Harvard Medical School ethical guidelines that limit consulting when an investigator is involved in clinical research.

I want to start today by talking about the important problem of uterine fibroids. As everyone in this room probably knows, this is a very important clinical problem for women. That are very common tumors and the prevalence rates vary from anywhere from about 20 percent of women to being affected, to more recent estimates looking at high-risk populations by ultrasound where the prevalence of clinically detectible fibroids appears to be in the range of 75 percent.

Most of the discussion regarding uterine fibroids centers around cost and the costs are substantial for a healthcare system. It's estimated that the cost for hysterectomy alone in the U.S. along per year is in excess of $2 billion. This is really the tip of the iceberg because it doesn't even start to take into account other surgical options, non-surgical options, medical options and various alternative treatments that women seek to try to control their symptoms.

I think it's important to note also that there is information regarding productivity in women with menorrhagia and so this is probably an under-estimate for the kind of women that we're seeing in our study who have clinically significant fibroids. The estimation from 2000 was that lost productivity due to menorrhagia or excessive menstrual flow is in the range of $1,600 per woman per year.

I think it's important to realize that fibroids are a common source of morbidity for women. They cause a lot of symptoms that tend to cluster in several different areas. Menorrhagia or excessive menstrual flow is an extremely important problem due to fibroids. And for women, this really limits their ability to carry out their work or their interactions with their families. There are many women that spend up to two weeks every month with significant menstrual bleeding, and there are many women who have such significant menstrual bleeding that they cannot attend to any other activity for an hour or more without having to stop to deal with changes in sanitary protection.

Pain and discomfort are significant symptoms related to uterine fibroids. Many clinically significant fibroids are in the range of a three, four, five month pregnant uterus, and this gives women significant symptoms in terms of urinary frequency, urgency bladder discomfort, pelvic discomfort.

These symptoms have been shown to significantly impair health-related quality of life, and in several studies there have been demonstrations that women with uterine fibroids have significantly lower health-related quality of life than population norms. Uterine fibroids have also been linked to time away from work and other activities that are important to the economic system. And the Rand Corporation estimated that medical therapy may fail to control the symptoms in approximately two-thirds of women, so we do need better therapies for uterine fibroids.

There are treatment options for uterine fibroids, but I think if you look at the range of options available for uterine fibroids in contrast to the woman who has a normal uterus and her options available for endometrial ablation, the contrast is clear. This panel has approved many devices for endometrial ablation, and many of those are restricted to women who have a structurally normal uterus.

Hysterectomy is a good solution for uterine fibroids. It is very effective in solving the symptoms, but it does have a significant morbidity associated with it, and a significant time away from work and family. For many women today to have the six week recovery for a major surgery is something that they cannot incorporate into their work and their family.

Myomectomy is an option for women who have a desire to retain their uterus but want resolution of their fibroid symptoms. Clearly, there are some women that are amenable to minimally invasive Myomectomies if the fibroid is in the right position at the serosal or the mucosal surface. However, again this modality is a surgical modality, and can have significant recovery associated with it.

Uterine Artery Embolization has been an important option that has been added in the past decade for women with uterine fibroids. It has significantly decreased recovery time and fewer complications than hysterectomy. However, this modality is associated with pain and fever post-operatively, and there's increasing attention to the fact that there is an age-related impairment of ovarian function and this may be particularly an issue for certain groups of women.

Thermally ablative therapies have been tried previously for uterine fibroids. Many people have had experience with either myolysis or cryomyolsis, and there's a small experience with RF-ablation. These techniques have not really made it into the general gynecologist armamentarium, probably because of a lack of thermal monitoring.

With these prior therapies, there was no gauging of temperature, and so you couldn't tell had you established a sufficient temperature to destroy the tissue. If not, you probably decreased your efficacy. Or if you exceeded the temperature goal, potentially you were injuring normal tissue and causing problems with adhesions or other follow-up.

Again, we do have drug therapies, but they tend to fall into two broad categories. Drugs such as oral contraceptives and progestins are widely used to control fibroid symptoms, but they tend to not be efficacious in the long-term.

On the other hand, GnRH agonists are very effective, but their side effects are significant, and their cost is significant, and so these drugs really haven't been great long-term choices for women with uterine fibroids.

We see that there's a spectrum of options available for uterine fibroids that for women with severe disease or who require a definitive solution, hysterectomy is still a choice. But many women are sitting down here with expected management and dealing with significant levels of symptomatology because they fear the surgical invasiveness of the other options, or because they cannot, again, take the time and the recovery that's necessary to undergo a very invasive option.

We think MRI guided focused ultrasound surgery will be a very important option to offer women. It will give them the symptom relief that they require with significantly less invasiveness than many of the other options.

There are several unique things that are important to know about MRI guided focused ultrasound. It is a non-invasive, rather than a minimally invasive surgery. There is no surgical incision. There is no probe that goes into the fibroid. It is able to be accomplished as an out-patient procedure. Again, it serves the uterus and is uterine sparing.

The other important issue is that it is a fibroid-specific therapy. Unlike something like uterine artery embolization that targets the entire uterus, the fibroid is specifically targeted so that there is no impact on the myometrium or the endometrium.

Again, the real time feedback on temperature gives you precise thermal ablation, and this is very important both for the optimization of safety and efficacy. Again, you can know that your temperature is getting to a therapeutic level and causing tissue destruction, and yet remaining in a safe range. And we have found that this procedure does not preclude or complicate future treatment options.

I will return the presentation to Mr. Newman, who will talk a little bit more about the device.

MR. NEWMAN: Thank you, Dr. Stewart. I'll just briefly review some of the key points of the device itself. As I said, much of this material is in the panel package, so I won't belabor the issues.

MR guided focused ultrasound is really a combination of two things, the idea of focused ultrasound as a source of thermal energy, and MR to plan and control the treatment in progress. There's two main components; one is the patient table and the electronics that's attached to the MR system. In the top of the patient table is the transducer, and there's a water bath here. The patient lies on top of that. The energy is transmitted through the abdominal wall and focuses on a point inside the body.

Out next to the operator console of the MR is the control console for the focused ultrasound. Here's the regular MR console, and they sit side-by-side so that you can see your work on both systems during the treatment. Once the treatment begins, all of the control of the treatment and all the observation of the patient images is done from the ExAblate workstation. Next slide, please.

Just a brief history of focused ultrasound. Although this may be one of the first times that many of you have heard about it, focused ultrasound is a technology that's been around for a long time. There are publications as early as the 1930s. We didn't invent this. We're just kind of the latest people to carry on a long line of research in this. The Fry Brothers in the 40s and 50s did a lot of work on this looking at focused ultrasound in the brain and other places in the body.

Lele carried on work looking at several tumors. There's also some carried on with some work using focused ultrasound for acoustic hemostasis and other applications. In 1993, Hynynen, Cline and others wrote the first paper on the combination of MR with focused ultrasound using MR for the thermal imaging. And then 1995 to present is ExAblate, the development of the device we're discussing today.

The transducer, a little of the physics of the transducer. The transducer lies here. The energy passes through the skin and intervening tissue to focus at a point. The energy is highly focused. It's kind of like taking a magnifying glass and focusing the sun's energy, so you can put your hand above the magnifying glass, below the magnifying glass, and it isn't until you get right at the point that the energy is highly concentrated.

The density in the far field, the energy is attenuated and absorbed along the beam path so while it's highly concentrated here, it falls off with distance in the far field.

The focused ultrasound energy propagates through tissue and skin. It's blocked by air, such as in bowel or the rectum behind the focal point, and it's absorbed by bone, such as the pubic bone or the sacrum in the far field.

This is just brief picture showing the patient lies on top of it, so here's the transducer underneath. Here's the overall beam path for this entire volume, and in the blue is the focal path for a single sonication.

The one thing that's different about this as a source of thermal energy is we ablate one small piece at a time, as opposed to a cryoprobe where you create a large two, or three, or four centimeter lesion in one go, or RF ablation. We build up the treatment from a series of these individual sonications that are approximately 25 by 25 by 10 millimeters, so you're ablating about a half a cubic centimeter at a time. A single sonication takes about 20 seconds, and the target is to raise the tissue in-between 65 and 85 degrees Centigrade. If you raise tissue above 57 degrees Centigrade for one second, it's ablative.

There's a rapid fall-off. It's a highly focused transducer, so there's a rapid fall-off with distance, so just a very few millimeters away from the focal spot you're back at normal body temperature. There's also a combining effect here because tissue's sensitivity to temperature is very time-dependent, so when you look at one pixel, it's the time temperature, it's the product of time and temperature that dictates whether you've had ablation or reversal heating of that point. So when do a treatment, the physician draws a region of treatment around the area to be treated, and then the system tiles it, if you will, with a series of these jellybeans, so that basically you draw a region of treatment and the system figures out how many jellybeans are in the jar. So how many will it take to completely cover this volume, so you can do a single layer, you can do multiple layers, and here's what it looks like in the horizontal plane, or looking down on it from above.

The treatment is controlled by MR thermometry. You're doing MR continously throughout the treatment, so it isn't like you do a single planning image or a stereotactic plan before. You're using MR continuous throughout the energy delivery, so approximately every three seconds you're acquiring an MR image. The accuracy in vivo in fibroid tissue is about 3 degrees Centigrade. And what we're doing is we're measuring change in temperature with MR. We can't measure absolute temperature, but we measure change relative to body core temperature.

We take the information. We use this to tell us -- we can see the focal spot. We can tell where the energy is being delivered in three dimensions, and we can quantify the temperature to get this time/temperature information from each pixel.

At the end of each sonication, using this time/temperature information, we can calculate the volume of tissue that exceeded the dose, and we can use this information to plan the next sonication.

This is just a quick picture showing during a single sonication we're acquiring an image every three seconds here over a 15 second sonication in this case, so you can see at 1.7 seconds, you can see the spot start to show up on the MR image. At the end we take this information, calculate the volume of tissue that was ablated, and we can draw one of these time versus temperature histograms here or maps, so we can where this little red cursor - it's hard to see in this slide - but there's a little cursor here, and you can see the time/temperature history out to 98 seconds.

If we would move that cursor somewhere here in the background away from the focal point, you'll just see some bouncing around, plus or minus a few degrees of normal body temperature.

We've done extensive thermal modeling in both 2D and 3D looking at a simulation of energy along the beam path to quantify the absorption of the energy, and to look at the tissue characteristics. We've done this to explore boundary conditions, to look at what type of dosimetry would be appropriate for maximum effectiveness, and to minimize thermal damage outside the treatment volume. And to really simulate things that we can't really do in vivo to look at kind of worst case scenarios of energy delivery and absorption that we wouldn't be able to do in vivo.

Pre-clinical evaluation was extensive where we looked at transducer designs, looking at transducer power, verification of ability to control the focal spot. There's a lot of work done in cavitation. Some of you may be familiar with focused ultrasound in other applications, such as lithotripsy. In that application, you're trying to generate cavitation. The whole point is to generate a very high energy shockwave to shatter a stone, such as a kidney stone. In our application, we only want thermal effects, and we want no cavitational effects, so there's a lot of design in the system and the use, limitations on the use to make sure that we limit our effects to thermal effects.

There was testing of the biocompatibility and a lot of animal testing in both -- for both our system and in the literature. There are several dozen publications over the last 15 years on the use of thermal imaging in MR.

Next I'd like to introduce Dr. Clare Tempany, who will give us an overview of MR anatomy for treatment planning.

DR. NOLLER: If I could interrupt for just a second. Our support personnel, could we have the temperature turned down a little bit, whoever is doing that.

DR. TEMPANY: Thank you. Good morning, Mr. Chairman, panel members, and guests. My name is Clare Tempany. I, too, am a Clinical Trial Investigator at the Brigham. My trip and accommodations have been paid for by the company. I work as a consultant like Dr. Stewart for the company, and work within the Harvard Medical School Guidelines for Conflict of Interest and Ethics in Research.

What I'd like to do for you today is two things. I'd like to introduce you to the MR imaging anatomy, display of anatomy and pathology that's used in this trial. It's used routinely in clinical imaging today, and then walk you through a typical clinical treatment.

Female pelvic MRI has become a very powerful diagnostic tool, and it's been available now to us in radiology for over 15 years. It exclusively displays the female pelvic anatomy as you see in these what are called T2-weighted images for you on this slide. On the left you see a sagittal view, and on your right is a coronal view. And on the left, you can see the anatomy of the uterus and cervix displayed with the substructure of the zonal architecture of the uterus displayed with the layers delineated for you. And on the right you see the same thing with the ovary on either side.

Many of you are more familiar perhaps with pelvic ultrasound, and these are images of patients with fibroids where you can see an enlarged uterus here in the center, and then you see a slightly different appearing uterus in the right side here. The texture and tissue characterization of ultrasound is somewhat limited to either solid or cystic, where we can see the differences here with the cystic component on the right.

A little bit of MR anatomy and how we visualize these fibroids before we determine whether they're eligible for treatment or not, selected images here. Now we're going to walk through several planes just to show you the display of the anatomy, and on the left you can see an axial view with the patient lying prone. The blue line represents the sagittal image on your right, and all of the relevant structures will be labeled, obviously, but you can see a very typical uterine leiomyoma sitting here in the center. It's classically a typical one that has a very low signal intensity or it's black, and it has a very sharp border. This is what we call a cookie-cutter sharp border which delineates and differentiates this from say adenomyosis, which will not have such a sharp border.

The coronal plane here you can now see nicely posteriorally as delineated up here on the blue line, but way in deep at the back of the pelvis here and the woman is standing in front of us, you can see the sacral nerves coming down here posteriorally, coming down along the lateral aspect of the pelvis to exit through the sacrosciatic notch. And as we come forward, you can see more anteriorally now. We're coming into the uterus. We see the large fibroid. We can see its relationship to the bladder. It's very easy to understand some of the symptomatology this patient has experienced when you see images like this with a large uterine fibroid pressing on the bladder.

Now axial planes in a typical treatment position now with the patient is lying again prone, and you can see the uterine fibroid sitting here. We see the anterior skin there, and you can see the direction of the beam as you will see in a minute. And there's the fibroid. These are the anterior rectus muscles here anteriorally, and posteriorally we see the fat, the bowel, and the sacral nerves.

We've learned a lot about uterine fibroid or leiomyoma imaging over the years with MRI, and done many pathological correlation studies, and have determined that there are many types of fibroids, as you've known in the clinical world for many years. And these can be seen and characterized well in MRI. And to just summarize some of them here for you where you see about five different varieties described.

The top two are probably the typical ones that we would treat in this trial, or have treated in this trial, and these consist of the classic leiomyoma which is a fiber muscular stroma. It's of low signal in every imaging sequence we have. In other words, it's black, it's easy to see.

A different type is what we call hypercellular, where it's also known as the white fibroid. It appears at high-signal intensity on T2-weighted images. Those are the ones we've treated.

The other group will represent ones that we wouldn't treat, which are non-enhancing leiomyomas basically, once that have already undergone spontaneous degeneration or necrosis in vivo, and obviously, of varying patterns also.

Just to show you some more examples of the range of the types of appearances of fibroids, here's a woman who's had very significant fibroid burden. Everything with an F on it is clearly a fibroid here. This is a coronal T2-weighted image, as if she's standing in front of us, her urinary bladder in white, and you see how this may appear like a five-month gravida uterus.

On the right side we see a different patient with multiple fibroids and an unusual appearing one here posteriorally that's already undergone degeneration. This is a large cystic degenerated leiomyoma, and we know it's degenerated because we have post contrast images here in the middle that's after the injection of intravenous Gadolinium, and it shows no evidence of enhancement or it stays black with no perfusion; thus, indicating that it's necrotic. So let's just move into a treatment process now.

Much of that imaging will occur prior to the patient's being determined as eligible for the trial, and we have identified, selected the patient and identified the target treatment, and this is what now happens on the day. So starting the night before, the patient will receive written guidelines about the therapy and what to expect during the treatment. She will review that. She will have prepared the abdominal wall, removing abdominal hair from the umbilicus down to below the pubic bone. This is important because we want the skin to be as smooth as possible, and to not interfere with any coupling or cause decoupling of the ultrasound beam as it will transmit through the skin.

She remains NPO from midnight because we use intravenous conscious sedation, and clearly don't want to have any problems with food. So we have the patient then come in the next morning. I meet with the patient. We review the treatment guidelines with her. We review what sort of sensations or experiences she may feel during the treatment. We develop the communication ritual to tell her when we're going to do a sonication, she tells me what she feels, and we sort of discuss all of that communication issue before we go in the room at all.

I also then consent her for administration of intravenous conscious sedation per our hospital guidelines. Once that is done, the IV line is sited, the Foley Catheter is placed. We use a Foley Catheter clearly to control the bladder during the procedure to make sure that the bladder stays empty. As you know, when the bladder fills, the uterus moves, and treating a moving target is clearly difficult, so we use the Foley Catheter to control that.

At the same time in parallel, the room check is going on. There's a phantom checking of the system occurring, and after all of that is done, the patient then comes into the room, is positioned on the table in the coil, her vital signs monitoring devices are placed in position, obviously her pulse ox, blood pressure cuff, et cetera. The nurse will remain in the room with her at all times, and both she and the nurse will have a small little sonic button in their hands which will allow them to terminate an individual sonication should the patient experience an unusual severe pain. She has full control of the therapy itself at the time.

So here's just some pictures. You can see this is the MRI magnet, this is the table, patient sitting getting ready to go into position. She then turns over and lies prone, positioning the pelvis over the transducer. The transducer, as you've seen already, is in the table surrounded by degassed water so she lowers the skin down onto the water bath basically with a gel pad also in side it, and she makes direct contact with the skin into the water.

A little bit about our conscious sedation and monitoring of the patient during the procedure. We use standard intravenous conscious sedation medications at our site. We use Versed and Fentanyl. These are administered to provide a combination of both analgesia and sedation. It's clearly important that the patient's anxiety and any claustrophobia that she may be experiencing in the magnet be aided by the administration of these medications. Patients, obviously, may experience positional pain lying on their stomach in the magnet for the duration of the procedure. And again, the analgesic effect is useful for that. And we obviously want to try to reduce any pain from sonication so we use the Fentanyl.

Typical doses that have been used in the procedures, and these are the total doses, range from as little as 25 mics of Fentanyl to 250 mics, to .25 to 5 of Versed. These are both intravenously. We also give patients an oral non-steroidal anti-inflammatory at the very beginning of the procedure. Usually, typically 75 milligrams of Voltaren has been used.

The medication then is given as required. Before we start any treatment, we will give a very small incremental initial dose of Versed and Fentanyl, and then depending on how the patient is feeling, responding during the therapy, we will give further doses during the procedure, so that's why the ranges are quite wide here. Some patients require very little, some patients require a little bit more.

So let's just start now with treatment planning. The patient is positioned on the table, and you can see the transducer. And this is a good positioning on your left here, as opposed to the one on the right where the transducer is too high. And you can see this is a very large field of view image here. The uterus is really too low, and we would have to angle too steeply to treat that, so we clearly can readjust the transducer and the patient at this stage before we start going any further.

We will then take three planes of pelvic MRI images, as you've just seen, an axial, sagittal, and coronal to again define our target, to allow me to draw the contour to target volume superimposed on the fibroid at that point, and those images are coming up in a minute. We'll show you how we do that.

Just to remind you that in the trials, we have protocol treatment guidelines. Single treatments initially were limited to 120 minutes. The maximal thermal dose per fibroid was limited to less than 100 CCs, and you could see treatment for all fibroids, if more than one was treated, was a total of 150 Ccs.

We have a maximum of four fibroids that could be treated in any one setting.

The protocol treatment guidelines delineate a little bit further in detail here for you, and the schema on the right really explains it all nicely. The large black circle is a fibroid. The smaller one on the inside the region of treatment, the ROT, is the circle that I would draw as the sub-volume in the fibroid. We have to work with the guidelines, obviously, remaining within 15 millimeters of the outer serosal lining, and 15 millimeters of the endometrial lining. And so this clearly restricted somewhat the volume of the fibroid that we could actually treat during the initial safety and efficacy evaluations.

Here are some pictures of the same thing. You can see the treatment plan. Now the sonication grid has been overlaid and you can see these jellybeans, as Mr. Newman has already referred to, and you can see them overlaid here on the images. Before we do anything now, the next thing to do is to walk through each of these sonications and determine is the beam path going to be safe, and will it remain within the guidelines. So we work through this system here where we see the beam path on each and every one of these. And there are some images now, just to show you how that's done. You can see the passage of the beam going through in green here, and the focal point is delineated there on the sagittal view, the axial, and the coronal.

What can be in the way? Well, things that certainly can be in the way that we can identify relatively easily are things like scars that would be in the skin from prior surgeries, clearly things that are in the skin such as scar can cause defocusing of the ultrasound beam as it's passing through, cause local heating of the skin, and something that we try to avoid at all costs. And so it's fairly simple to do this, we simply identify the scar ahead of time, and then using the roll and tilt mechanism of the transducer, we can angle around that area.

The same thing with bowel loops. Bowel loops are relatively easy to see here. You can see them on this T2-weighted image. They're the dark structures up at the top, and you can see again, we can angle either up and around the bowel loop, or if necessary, just simply not treat that area, clearly not go anywhere close to the bowel loop. The sonication can simply be deleted.

Same thing here if we're looking at the distal field. We can evaluate the location of the sciatic nerves, and we can determine whether or not the beam is going to pass through, and angle and roll, and tilt again to avoid it.

Okay. So now we're ready to go. The first thing we do is the geometric accuracy, and so this is when a low powered sonication will be delivered, and the very first set of images will come up as that's being delivered, and this is a cropped down view here of the face map, and you can see that we're determining the accuracy; first of all, the visibility of the sonication. Can I see it at all? And you can hardly see it because it's covered by the red cross, but underneath that little red cross is a white dot, and that's the first sonication that's being delivered. And the initial assessment is good, now I see it. Is it in the right place? And so here you can that it's off by about 5 millimeters, so we will readjust all of the anatomical and adjust the geometric alignment so that the green overlies the red, and that they are absolutely concurrent.

The next step then is to move into a therapeutic sonication dose, so we increase the power up to typically 100, 140 watts, and we start the actual procedure with therapeutic doses being delivered. We compare this as it's going. We modify the treatment parameters as necessary. As you'll see in a minute, we're constantly looking at the feedback mechanisms of the thermal imaging, to determine if we've achieved a therapeutic dose or not.

Throughout the procedure I'm in constant communication with the patient. This is very important because there's a one-on-one communication between myself, the patient, and the nurse in the room, but I will talk to her, tell her we're about to start a sonication at the beginning of one, and then at the end of that ask her if she's experienced any sensations, and if she has any concerns.

These are some examples now of the typical dose profile. On the left, you see a sagittal view or a long axis of a sonication, so you see the jellybean shape. This is a short axis view where you see it on end. And then these are three incidents of things that could happen, so if you look at the bottom left, we have a sonication that's achieving a thermal dose that's probably too hot. The temperature, you probably can't read that, I'm sure I can't either - it's 100 degrees is what that one has reached, so clearly, that's a little too hot. So what we do in that situation is to back down on the power before we go any further, so we wouldn't continue to treat without changing parameters once we've seen that.

Similarly, in the opposite direction, the next one demonstrates a sonication that achieved a 50 degree temperature, and that's too cold, so the first step then would be to increase the power to bring it up to a therapeutic dose, which we like to see between 60 and 80 degrees.

Calcifications occur in fibroids, as you know, very frequently. They can be small punctate little pieces of calcium within a fibroid, or you can have a very more densely, heavily calcified one. The latter patient doesn't usually get into the trial because we can identify that in imaging, and a dense rim of calcification precludes treatment using this treatment modality. But small punctate calcifications are impossible to see ahead of time, and this is what may happen.

The ultrasound beam will be reflected off the sonication, will simply not achieve any therapeutic dose, so we simply move onto the next location, delete that sonication, so to speak, and don't treat that specific area. So an overview of the treatment cycle is seen here for you for an individual sonication. Before anything happens, the MR scanning starts. Then the sonication is delivered, then a tissue cooling period occurs, and at all times the images are being acquired, and this total is about 2 minutes. So this is an extremely interactive treatment.

As you've seen already, the entire beam path is checked prior to delivery of sonication, irregularities of skin, bowel, and beam path are evaluated. We have multiple tools available to avoid critical structures, things that we would not want to have the beam pass through, and we use each and every image to modify the next sonication, so it's a very iterative process, so we're learning from the last sonication what to do for the next sonication. And we do this with the MR imaging that's continuously occurring during the procedure.

So there are some safety issues, obviously, where motion is a problem, if patients were to move during the procedure, as I've already said, heat treating a moving target is not good. So we obviously have prevented that now by the Foley Catheter placement with the bladder being controlled. We also obviously coach the patient that she should not move her pelvis. She's lying prone, and if any of you have ever had an MRI scan, you know that we strap patients in on the table, that there's a coil around it, so it's fairly restrictive. There's not a lot of room to maneuver and to move around, so these are things obviously to our advantage.

We also use the restraint strap which is strapped around the outer pelvis to hold the patient on the table. The sedation somewhat helps also, but clearly she can still move if she really so desires. We monitor this with both sets of images. The real time images being acquired during the sonication are very easy to see motion, because it's like watching a movie. You're sort of seeing a cine loop, so to speak, so you can see changes if she was to move her skin or her spine.

We also place fiducials at the beginning of the imaging sequences, and these are the little red marks you see here. And those are monitored carefully, as well, to ensure that they don't change in position over the procedure.

The outcome assessment while the patient is still on the table, essentially as we're going we're developing this blue map in the center here which represents all of the therapeutic sonications that have been delivered, and have been therapeutic; in other words, reached the goal temperature delivery. And so the blue is the area that we will expect to see the necrosis. And at the end of the procedure, we confirm this by injecting Gadolinium and evaluating the necrotic tissue. And as you can see, nicely maps. The blue area here is now seen as the black area here, which is the non-enhancing or necrotic tissue.

Again, some other images from the end of a treatment. Typical treatments look like this. They can range from relatively small sub-volumes to slightly larger volume here, with the areas of necrosis seen in the area of treatment. And I thank you for your attention, and pass the podium back to Dr. Stewart, who will continue with the clinical trial design.

DR. STEWART: Thank you. In moving on to clinical trials in fibroids, that can be quite a daunting task. As the Duke Evidence-Based Practice report has shown on, despite the fact of a wealth of clinical experience with uterine fibroids, this isn't a lot of good evidence on which to base therapy.

We were fortunate in going into our feasibility study having information from an in vitro model using a rodent model, and using ultrasound guided high-intensity focused ultrasound that showed treatment with this energy modality was feasible for uterine fibroids. And we wanted to get several important things out of our feasibility study.

First of all, we wanted to make sure that this was a safe treatment for women. We also wanted to confirm our targeting accuracy. As Clare has discussed, the feedback we get from the MR is important, and we are depending on the non-enhancing volume representing the tissue that we have successfully ablated, so we did want to get pathologic confirmation of this ablation. And this is actually something that hasn't been done with previous therapies, such as myolysis, cryomyolysis, or even uterine artery embolization. And we wanted to take this information to help refine our pivotal trial design.

The study design was that it was an open trial for women who were scheduled for hysterectomy. They were to undergo MRI guided focused ultrasound three to thirty days in advance of their hysterectomies. In your panel packet, it appears that there are two distinct studies that our center and St. Mary's in London has described in one area, and then the other three sites are described in another area. However, because women were reluctant to go through treatment and hysterectomy, recruitment in the original cohort suffered, and so as time went on, these other sites began recruiting patients, as well. And then, in fact, the Israeli National Health Service made hysterectomy optional for that group of patients. They felt that it was unethical to require women to undergo this therapy and then not have the option of opting out of definitive therapy, so our trial design changed somewhat midstream, but we followed all of these patients, and reported them together.

We were able to confirm our pathological information, and this is a diagram from our manuscript. This is the treated fibroid, and this is the pre-MR imaging that shows Gadolinium going throughout the fibroid indicating good perfusion. This is the post-treatment Gadolinium MRI where you see a large area of non-enhancement. And then this is the hysterectomy specimen. You can see on gross examination that there is a clear lesion that corresponds to the targeted area. And on microscopic exam, there appear to be coagulative necrosis corresponding to this area.

We were also able to confirm that there is a relationship between the targeted volume, the non-enhancing volume, and the pathologically correlated area of tissue destruction. In this particular fibroid from our St. Mary's site, you see the thermal dose volume in A, the B is a little bit bigger, the non-enhanced volume, and the pathologic area confirmed more closely to this non-perfused volume.

We did find that the non-perfused volume in general over-estimated the amount of tissue destruction, but we found that in all cases the area of targeting was confined to the treated fibroid. There was one case where microscopic evidence of sonication was seen at the serosal border; however, in retrospect, it appears that that was incorrectly targeted. This was one of the cases where the bladder filled and the target moved, and is a reason why we adopted a Foley Catheter with our pivotal trial treatment.

So we were able to confirm pathologically that the tissue that we thought we destroyed was destroyed. We also were very pleased with our results in terms of patient treatment. All but one patient were able to be treated as an out-patient. There was a single hospitalization overnight for control of nausea. There was no post embolization syndrome. There, in fact, was very little pain in women undergoing this protocol. And most of the patients that we saw were not even taking over-the-counter medications at the time we saw them within 72-hours of their treatment.

The one safety issue that we did see in this initial protocol was there, there was a significant incidence of infection seen post-hysterectomy. They did not occur between the focused ultrasound and the hysterectomy, but following the hysterectomy. And we stopped the trial at the time we saw the first three infections. We reviewed our procedures. At that time, we did change our protocol to institute prophylactic antibiotics. And once those were instituted, we didn't see further significant infections. And our pivotal protocol did not have prophylactic antibiotic use.

We also used the information in the trial to mitigate the adverse events we saw. We found early on that paying attention to the skin in various forms was important. Initially, patients were not shaving and there were small skin burns at the area where there may have been loss of coupling of the ultrasound to the skin. We also, again, found the importance of mapping the scars, and incorporating those into treatment planning. Because scar tissue is very similar to fibroid tissue, some of the energy would stop at that point and patients would be uncomfortable, and so we used a lot of the information from this feasibility study to define the optimal treatment protocol to embark on our pivotal study.

The major issue when embarking on the pivotal study was the selection of a control group, and there are always issues with picking the perfect control group. And it's especially important, I think, to put this in the context of the times. At the time that the selection was going on, it was December, 2001. Although uterine artery embolization today might appear to be the best alternative, as a control group, this was not really possible at that time. There were no embolic agents that had received FDA approval at that time. And with extensive negotiations with the FDA and the investigators, we looked at the other alternatives. And we felt that looking at a surgical option would really give us important safety information. It was important to have a contemporaneously recruited control group, and not to depend in historical controls.

Again, abdominal myomectomy in many ways appears to be an important option. The issue for this group of patients was that many of them may not be symptomatic. They would be pursuing treatment to attain fertility. They would also tend to be younger than the symptomatic patients that we were seeing. And with our group, we specifically wanted to recruit women with a threshold level of symptomatology. Therefore, we decided that although no control group was perfect, that abdominal hysterectomy would be the best alternative.

With our knowledge of difficulties in recruitment and our pivotal study, and also information we were gaining from the experience with uterine artery embolization trials, at this time many groups were trying to perform randomized trials between conventional surgical therapies and uterine artery embolization. And no one succeeded in having sufficient enrollment, so in that group of patients there were generally case series or parallel controls. And again, this is the study design that we settled on.

The hysterectomy group and the focused ultrasound group were enrolled in parallel. They met the same inclusion and exclusion criteria, and both received the same six month follow-up. We also chose to separate the sites TAH and focused ultrasound so that you did not have investigator bias channeling good prognosis patients into focused ultrasound, and bad prognosis patients into hysterectomy, so the sites were all separated. And with our power calculations we found that a 3-2 ratio would give us the desired number of patient's in each arm.

The inclusion criteria included women who were not pursuing future pregnancy. We felt it was not ethical to treatment women who desired future fertility until we had information regarding the efficaciousness of this treatment. They were all pre-menopausal or peri-menopausal women. They did have both clinical exam and MRI consistent with fibroids. The fibroids needed to be visible on contrast MR, and feasible for treatment.

We also chose to have a minimum symptom severity score, so that they had to score over 40 points on a scale of 100 to be included in this protocol. The exclusion criteria were fairly obvious. Women who could not undergo MR were not included. Women with excessive uterine sizes in excess of 24 weeks, or women that were too heavy to fit in to the MRI equipment were excluded. We also excluded anyone with an undiagnosed pelvic mass, or other worrisome pelvic pathology.

The primary hypothesis for our pivotal study was that we would see at least a 10 point improvement in the uterine fibroid symptom and quality of life symptom severity score. This is the only validated quality of life score specific for uterine fibroids. And we felt that in our treated group, we would have at least 50 percent of our patients achieving this goal.

We realized that the treatment modality would likely not be as effective as hysterectomy given the limitations, but we felt that this was an important landmark in demonstrating the efficacy.

We also evaluated several important secondary hypotheses. We wanted to look at the significant clinical complications in both arms to compare safety. We wanted to look at the trajectory of recovery, and also the costs involved.

For those of you not familiar with the uterine fibroid symptom and quality of life measure, this again is a disease-specific validated measure. It was developed specifically for uterine fibroids and it has two different parts. The symptom severity score, which you'll see in this presentation referred to as the SSS, has eight questions that relate to the fibroid specific symptoms, pain, bleeding and bulk.

There is also a component to the health-related quality of life which has six different sub-scales as is common with all quality of life questionnaires. And this questionnaire was developed from an ethnically diverse set of focus groups to really get input of fibroid patients, and what they felt their significant symptoms were.

Also during the validation process, this was correlated with the SF-36, which is really the standard measurement of quality of life, as well as a menorrhagia questionnaire indicating its comportance with symptoms of menstrual blood loss.

For those of you not familiar with the questionnaire, you'll see that the symptom severity score addresses issues such as heavy bleeding during your menstrual period, passing blood clots. It also looks at bulk related symptoms, feelings of tightness and pressure, frequency of urination or nocturia or feeling fatigued. And patients are asked to rate their symptoms on a five point Likert scale from not at all to a very great deal.

This is data from the initial validation of this questionnaire that you'll see the two parts are divided here to the symptom severity score, and these are the sub-scales of the health related quality of life. One of the first things you'll notice is that there's an inverse relationship between them. For symptom severity score, the women in blue who are women with uterine fibroids, have a higher score, so higher scores mean higher symptoms. Whereas, with the health-related quality of life, the normal women tend to have higher scores and impaired related quality of life is reflected in a lower score.

It's also interesting to note the absolute levels of the symptom severity score. In this study, looking at women with symptomatic fibroids, the mean score was 44; whereas, the mean score for normal women was 23 or about a 20 point difference between the two groups.

This clinical difference was the primary reason we selected our 10 point difference between treatment success and treatment failure; that if 20 points represents the difference between women with fibroids and normal, getting 50 percent relief of symptoms appears to be an appropriate clinical end-point.

There were also standard methodologic reasons to choose this. That 10 points is very similar to the standard deviation in the population. The standard error of the mean and gives a moderate effect size, as well.

We did not depend only on one outcome. We also looked at additional efficacy measures. We used the SF-36 which gives standard health-related quality of life. We looked at several measures of disability days, some assessment of an overall treatment effect, and also patient's treatment satisfaction.

This is a schematic drawing of the pivotal study design that at the screening visit we perform the MR prior to the treatment, as well as the symptom screening. A hematocrit ruled out serious anemia, and during the treatment visit we again got information regarding symptomatology.

We took seriously that this was a new technology, and that there wasn't a lot of experience with follow-up, so we have everyone come back for a physical exam within a week so that we would not miss important issues that arose, so patients came back and did have a hematocrit and a physical exam at that time.

The one month and the three month follow-up were generally by phone, but then there was a full visit at six months with a physical exam and MR exam, and again complete testing.

The pivotal study design was originally designed to have outcomes at six months. However, later we have extended follow-up so that we're now seeing patients who are continuing on at 12 months, 24 months and 36 months. And again, getting information on quality of life, as well as MR exams at that time.

We wanted to try to capture significant clinical complications, and what we did at this time was we went to the literature. The paper by Dicker, et al, arose out of the collaborative study of sterilization. And they felt it was important at that time to try to define characteristics that could be used to compare treatment.

We used their criteria, but tried to update it both for the change and length of stay that has occurred since the 1970s, and also some of the differences that we would potentially see with this new therapy included as additions or things like discharged going to a rehabilitation facility, discharged with either a catheter or a drain, or also various interventional treatments that may not qualify under their definition of surgical procedures.

While this would seem to favor picking up complications from hysterectomy, I think it's important to remember that if there had been inappropriate targeting and significant injury of adjacent structures, these complications would have been seen and picked up if the treatment had significant side effects in that way.

So moving onto the results of the trial, as we talked about earlier for the pivotal trial, there were separate sites for hysterectomy and MRI guided focused ultrasound. There were three U.S. sites and several through Europe and Israel. There were also hysterectomy groups and about half of the enrollment for both arms came from the U.S., and half from out of the U.S.

There was fairly equal distribution of patients through the sites. There wasn't a primary site that contributed all of the patients. And we looked at the demographics between the patients undergoing focused ultrasound, and the patients undergoing hysterectomy. We knew that since this was not a randomized trial, there were likely to be some differences. We did find them similar in age, and fairly typical for women with fibroids. There was a statistically different finding in body mass index with the women undergoing hysterectomy being somewhat heavier. And both groups of women had significantly elevated symptom severity scores.

As you'll recall in the validation study, the women with fibroids typically had scores in the 40s, and both of our groups this mean score was over 60. And again, there was a difference between these two groups with women who had elected definitive therapy for hysterectomy having a somewhat higher score.

There were more black women in the hysterectomy group. Again, probably a relationship of site selection, but all women in both groups were pre-menopausal by and large.

There were some differences in co-morbidities. The women undergoing hysterectomy were more likely to have diabetes and hypertension, and the women undergoing focused ultrasound were more likely to have thyroid disease. As you'll see later on, we looked at these differences between the focused ultrasound group and the hysterectomy group to see if these differences affect the treatment outcome.

We did perform an intention to treat analysis, so that every patient who received focused ultrasound is included, and so our denominator in the slide you'll see is 109 patients. There were three withdrawals from the study less than six months, and 11 patients were non-evaluable. We did, however, do calculations for both evaluable patients and intention to treat patients, and they were similar.

The characteristics of the fibroid patients were consistent with women who had symptomatic fibroids. The average uterine volume was approximately 600 Ccs, but there were clearly a number of women who had uteruses in the range of 1,000 cubic centimeters or more. The average total fibroid load, meaning calculating the volume of the fibroids without the myometrium was in the range of 300 to 400 cubic centimeters. And patients had an average of two to three fibroids, but as many as 12. And although one to four fibroids could be treated during this protocol, in a average most women got one treated.

We excluded from treatment fibroids that were amendable to either hysteroscopic or laparscopic myomectomy, so although these say submucosal and subserosal, they were probably more accurately classified as partially submucosal or partially subserosal with a large intramural component. And we also looked at differences in location when we assessed treatment outcome.

So when looking at the treatment parameters again with the intention to treat patients, the baseline fibroid volume was about 300 Ccs. The non-perfused volume at the end of treatment was in the range of 68 cubic centimeters, so we had approximately 24 percent of the fibroid that had been treated during this protocol. That at six months, there had been a decrease in size from about 330 to 295. This percentage of shrinkage is similar to the non-perfused volume. Again, it's not a large absolute number, but it is proportional to the amount targeted for treatment.

Looking at our primary efficacy and the symptom severity score, again we hypothesized that at least 50 percent of our patients would have a 10 point improvement. We were substantially in excess of that. Over 70 percent of our patients reached this targeted improvement, and this was statistically significant.

We also found that in fact the symptom severity score at entry was in the range of 60. By three months there was already clear evidence of a treatment effect with a mean treatment level going down to 41, and then some continued improvement between three months and six months. And you can also see here, this is the criteria we set for entry, so many women at the three or the six month time point would not have had symptoms sufficient to qualify for enrollment if they had come at that point in time to seek treatment.

This is the distribution of changes in symptom severity score, so again this line indicates the threshold for success, or 10 points or more. These are the patients who had no improvement, or one to ten points of improvement, so everyone from here over is a treatment success.

The mean patient improvement, however, was about two and a half times what we had predicted and the mean treatment improvement was approximately 24 points. There were, however, some patients who improved as much as 60 points in symptom severity.

When we turn to look at the health-related quality of life subscales, these parallel the changes that we saw in symptom severity score. Because of the inverse relationship these lines go up rather than down, so again you see a significant change or marked change between baseline and three months, and then some improvement from three to six months.

We use the SF-36 to be able to compare more accurately the patients between the focused ultrasound and the hysterectomy arm. What we see again in the focused ultrasound group is the same pattern of improvement that already at one month you're seeing improvement in some scales, continued improvement at three months, and stabilization from three to six months. In contrast, the women who underwent hysterectomy had marked impairment in some of their functioning at one month, and it took them three months to six months to get back to where they were and, in fact, to note improvement following the treatment.

The significant difference is in terms of disability between the two groups. I think it's important to note not only the differences between the groups, but the absolute level for the focused ultrasound patients. When looking at the days -- this is follow-up at one month following treatment. There were only 1.4 days of missed work on average for the women in the focused ultrasound group. Whereas, women undergoing hysterectomy clearly has much more short-term disability with 18 days. And parallel the days that women with focused ultrasound were kept from their normal activities averaged about three days. And they again spent only about a day and a half in bed, so these numbers demonstrate the significant improvement and short-term recovery seen with this treatment.

We also looked at resource utilization through six months. Because of our different sites in different countries we didn't bring this down to dollars, but looked at encounters with the healthcare system. This takes into account not only all of the scheduled study visits for the MRI guided focused ultrasound patients, but for those patients that elected additional therapy, or went on to additional procedures. All of those resource utilizations are captured.

We found that there was a significantly different length of stay. Only 1 percent of our focused ultrasound patients stayed more than five hours post treatment. They also had substantially fewer provider encounters, fewer additional procedures, and fewer diagnostic tests.

We looked at a logistic regression model to see if our baseline differences affected outcomes, so in the model we included not only the things that differed between our groups, such as race and BMI, but also looked at other variables of interest, such as age, country of treatment, fibroid location, percent non-perfused volume. And the only predictor of success was baseline symptom severity score. In other words, the most highly symptomatic patients were the patients that improved the most.

We also looked at patient satisfaction and asked patients were they satisfied with their treatment, was it effective in eliminating their symptoms, and would you recommend this to a friend? And again, over 70 percent of women answered affirmatively to these three questions.

We did continue to follow patients beyond the pivotal study, and attempted to bring patients in for follow-up between six and twelve months. Again, we start with our intent to treat population of 109. We found that 91 patients continued on, 9 patients declined to be included in the follow-up. They had enrolled for a six month trial, and elected not to come back, and 9 were withdrawn, which left us with 82 evaluable patients at 12 months.

We found in following this group that 23 patients had gone on to alternative therapies, and four patients had elected and were offered additional focused ultrasound treatments. Both of these groups of patients are then included as treatment failures in our 12-month analysis.

So the original study was, indeed, designed for six month follow-up and we did contact as many patients as we could to come back. Because of the date that we started to do this, there was some lag, so although it's reported as 12-month follow-up, the actual mean follow-up was approximately 14 months. The success rates do not look as promising at this point. If we look at our intent to treat group, there's only approximately a 38 percent success rate. And again, the patients who declined to come back for us to follow-up or chose alternative treatments are included here. And if you look at our evaluable patients, it is slightly higher at 51.

I think what's notable is that there were a substantial number of women who were still improved with the mean treatment being targeted at approximately 20 percent of their fibroid load. The other thing that is interesting about the results at 12 months is that we still could see significant decrements in the treatment parameters as measured by the symptom severity score, so that at baseline again, we're coming in at about 61 points, and going down to points in the mid to high 30s at six months and twelve months.

Part of the issue with the twelve month data may be that fibroid symptoms returned. This is clearly a common problem in the literature, and is well described for myomectomy. Again, it appears to be an issue that may be applicable to uterine artery embolization, as well. But again, many of the studies with uterine artery embolization also have relatively short-term follow-up. And I think that our original treatment parameters were aimed at the maximization of safety and, therefore, may not have optimally targeted the amount of fibroids to get sustained treatment.

However, there were still significant patient satisfaction with treatment success at 12 months. Again you see in blue the six month data, and the twelve month data in yellow, so the patients were still very happy with the treatment option that they had pursued.

Turning our attention to safety, I think it's important, first of all, to note what we did not see; that many devices that are approved have significant complications. In many case series, there have been patient deaths or urgent unintended procedures. There were none of those in this treatment. There were no bowel injuries. There were no hospitalizations for pain control or post embolization syndrome. So compared to some concerns that we had at the beginning, we were very happy that there were not severe safety issues that we encountered.

Looking at a strict definition of adverse events, we found that 19 percent of patients in the focused ultrasound group had no adverse events compared with 1 percent in the total abdominal hysterectomy. We chose for this protocol because of its novel technology to strictly define adverse events more similar to what you would see in a drug study than a typical device study. We knew that this was a device that didn't have clear predicates, and we wanted to make sure that we were not missing adverse events.

However, we found that when we looked at device or procedure-related serious adverse events, we still did very well with only 2 percent of MRI guided focused ultrasound patients having serious adverse events compared to 13 percent in our contemporaneously enrolled group.

We found that the body systems in which adverse events were found to be similar in most cases. On average, women undergoing focused ultrasound had about two adverse events versus four for the total abdominal hysterectomy.

We also wanted to define what we thought prospectively would be device or treatment-related adverse events. We that non-significant events might include fever or pain in the treatment area, swelling or firmness in the treatment area, or minor skin burns. However, we felt that either skin burns that caused ulceration or any kind of nerve damage should be termed significant anticipated events, and that we were especially looking for these events as treatment unfolded.

Again, as we saw in our feasibility study, there was a substantial decrease in the amount of pain patients had both during this procedure and post procedure, compared to some alternative therapies. Interoperatively, the patients reported on average mild discomfort and mild to moderate pain. And then at post procedure, their levels of both pain and discomfort were significantly closer to no pain at all than to mild.

There were patients who had some severe pain during the procedure. And as Dr. Tempany discussed, we do have the ability to redose pain medications during the procedure. Only one patient, however, related her pain as severe post procedure. And again, we found that narcotic use following the procedure was very rare, and that even over-the-counter medication use was rare in the days following treatment.

We wanted to look at adverse events again to see if our baseline differences and the co-morbid conditions or the demographics affected these outcomes. Clearly, there were some co-morbid conditions in the hysterectomy group that may have made them more likely to experience complications. We found, however, that in controlling for this, the odds ratios still showed that there was significantly increased risk of dermatologic, gastrointestinal CNS and pain adverse events in the group undergoing hysterectomy compared to the group undergoing focused ultrasound.

Again, we wanted to look at the significant clinical complications to make sure that we captured significant events, and again use the literature to prospectively define this. We found that the patients undergoing hysterectomy were more likely to have a significant clinical complication with about 46 percent of the group in the hysterectomy group having an adverse event, as opposed to 12 percent in the focused ultrasound group.

One of the interesting comparisons is looking at fever and antibiotic use, given that the patients in the focused ultrasound group did not receive prophylactic antibiotics; whereas, the patients undergoing hysterectomy traditionally did. And still, the incidents of fever and antibiotic use started after the prophylactic antibiotics for presumed infection were lower. The transfusion rate was also low. There were no unintended surgical procedures, no discharges with appliances. There were several rehospitalizations, but none requiring interventional treatment, and no death or life-threatening events.

We also found that there were differences in -- there were significant differences in clinical complications. And our most serious adverse events included device-related adverse events; that we found that there were several instances of leg pain, which again we had identified as an anticipated event. There were also some skin burns, although most were first and second degree burns that resolved easily.

Our most important device-related, and our only device-related SAE involved a patient who had a treatment where there was injury to the sacral nerves. I think this is the case that pointed out to us the importance of having patients talk to us about their pain and discomfort. And that this patient did not receive pain medication at her request, and was noted at post treatment time to have weakness and nerve conduction studies confirmed injury.

However, by 12 months she has resumed a high level of physical activity, and in fact has run a marathon since her treatment. She had significant symptom improvement, and continues to be a part of our study.

We also put in a number of steps to mitigate the risk of nerve damage. As Dr. Tempany talked, there are several things that can be done in the treatment planning and the use of feedback from the patient. Since we instituted these measures, there's been no significant nerve injury, and the incidence of nerve injury has been minimal. So we also did a number of simulations that allowed us to look at this issue. And so again, we had one event since learning from this important case.

Looking at the serious adverse events, again we classified everyone that was hospitalized as having an adverse event, an SAE, even if it was felt not to be device or procedure-related; again, sacral nerve injury, nausea. And then there were four women that went on to additional therapy, which we felt was really progression of disease and not device-related.

There is one complication on commercial treatment that resulted in a patient death. A single patient in one of our outside the U.S. sites had a pulmonary embolism following commercial treatment. This was investigated by the local M&M committee, and it was felt that her death was not related to the procedure. And in fact, in retrospect it turns out she had several important thrombotic risk factors that had not been identified.

We do have a continued access protocol, and have continued to treat patients. It's very similar to our pivotal study with mild changes in the treatment parameters that allow slightly increased treatment time and treatment volume. And we've been enrolling patients in this protocol since April of `03 with 89 patients treated to-date. The adverse events in this group have been significantly less than in our pivotal study, and indicate that our mitigation steps have been successful. And we don't have enough of these patients to six months to comment on efficacy, but the three month efficacy appears similar to the pivotal study.

So in summary, we only had one device-related SAE, and a low incidence of adverse events. We confirmed that this treatment can be safely performed as an out-patient, and have learned from our experience to design a safer study protocol.

We also found that we met our primary efficacy point with a significant margin. We had a much lower symptom severity score than we had predicted, and all of the measures of improvement tend to move together to show patient improvement.

I'll turn the program over to Rob to talk about the training.

MR. NEWMAN: I'd just like to speak briefly to amplify on the information that's in the panel packet about the training program. We believe that this is truly a non-invasive surgical alternative. This is a scalpel of sorts, a non-invasive one, but it is a scalpel. The physician controls the delivery of therapy, and the system provides the ability for real-time interactive control of that looking at the results from the treatment itself.

The system works only a 1.5T MRI system. We believe that this is necessary. It's the current state-of-the-art for pelvic imaging for assessment of anatomy and pathology. And it also gives us the image quality that we need for accurate temperature measurements. These symptoms have a high level of service, and are in wide use throughout the medical community.

This system will only be used under the direct supervision of trained physicians. This is not something that would be used by anybody else. We believe that the gynecology and radiologic expertise is required, and the nursing requirements for these kinds of treatments are similar to what is currently being used in hospitals for regular interventional radiological interventional control, so there's nothing unique about that part of the treatment.

The training for all installations will include the entire team, doctors, the MR technologists and nursing. It's divided into two phases. One is the system operation, the technology side of it. The other part will be the clinical issues, which will be covered by preceptorships at clinical sites involving topics of patient selection, treatment planning, anesthesia, adverse event management and those kinds of things.

First, treatments will be supervised. And on our system, every sonication on our system is recorded and kept in a file, so we have a log of every treatment we've ever done. This allow us both to review prior treatments if you had an adverse event, or if you've had something interesting. It also builds us a continuously growing teaching file that we can use for future sites.

Just a brief overview, the kinds of things would be what you expect we would cover in the classroom part of it on system components, and the physiology, device, protocol development, and we would follow this up with training after the procedures have begun at a specific site.

InSightec has a continued commitment to studying MR guided focused ultrasound. We think that this is -- there's an ongoing process here, a lot we can learn. As we've described before, we have the continued access protocol is in progress. We've treated 89 of 250 patients, and we intend to complete that 250 patients and collect three-year follow-up data on them to look at -- to gather more data on safety and efficacy of the system. And will provide us a lot of information on improvements in treatment planning, and ways to make it more effective. And we also have additional studies ongoing outside the United States, and will include the analysis of that in our development of future features.

DR. STEWART: So in summary, I think we've demonstrated to you that the device that we're presenting has a low risk of serious adverse events. We were very careful to try to capture all events that occurred, and to report as completely as we could to make sure that this novel technology did not have any unintended side effects that we were missing.

One of the important issues with this technology is that it is fibroid-specific. And I think that that has benefits beyond what we've demonstrated today. The risk of complications is significantly lower than hysterectomy. And I think if we had chosen other control groups, we would have probably been able to demonstrate significant differences with other treatment modalities.

We've had a very low incidence of device-related events. And because this technique employs conscious sedation rather than anesthesia, there is also a decreased risk of anesthesia-related events.

We have seen a clinically significant improvement in these patients. Patients are very vocal about voicing their improvement with this treatment, and we have been able to capture that by a number of different modalities. We designed our study and well-exceeded both our primary and our secondary end-points. And to be able to gain this kind of improvement without surgical incision, without major disability I think is a major step forward. The fact that these procedures can be performed as out-patients is important, as is the fact that it preserves the uterus.

Many women, I think, with fibroids tend to live with their symptoms rather than go through some of the treatment options. Some women have significant disability that they put up with day in and day out because of their concerns regarding invasive therapies. And I think MRI guided focused ultrasound surgery gives us an important new choice, and an important choice to help reduce the symptoms of uterine fibroids for women. Thank you.

DR. NOLLER: Thank you. We very much appreciate the sponsor staying within their time limit.

Our next presentation will be by the FDA. By the clock we are using up here, it is now 10:13. We will take a break until 10:30 by this clock, for 17 minutes, and then the FDA will make their presentation.

(Whereupon, the proceedings in the above-entitled matter went off the record at 10:07:53 a.m. and went back on the record at 10:26:03 a.m.)

DR. NOLLER: Okay. We'll reconvene now, please. And again, I'll ask the panel to hold its questions until after the FDA presentation. At that time we will I think have about 30 minutes to formulate and ask some questions. I'd like to introduce Kathryn Daws-Kopp, who will lead us through the FDA presentation.

MS. DAWS-KOPP: Good morning, ladies and gentlemen, distinguished panel members and guests. I'm Kathy Daws-Kopp, the Lead Reviewer for FDA on this PMA. My presentation will give a brief overview --

DR. NOLLER: Excuse me. Turn to the sound up. We can't hear her.

MS. DAWS-KOPP: Okay. Good morning. I'm Kathy Daws-Kopp, the Lead Reviewer for FDA on this PMA. My presentation will give a brief overview of FDA's review process on this PMA to orient you for the remainder of the FDA presentations.

You may notice as we go through our presentation that you'll be hearing some of the same things that the company said. Our intention is to focus on the issues we felt were important in our review of the file.

I'll start off by describing the history of regulatory interactions with the company, and I'll describe components of the device from a regulatory perspective. I'll provide a list of the PMA review team, and briefly discuss what we did in reviewing the PMAs, and I'll follow that with a list of some major issues that are still ongoing with this review, some of which are part of the panel discussion questions.

I'll close with an agenda of the remaining FDA topics and presenters.

This is a brief overview of the history of FDA review on this device. The sponsor first came to FDA with a feasibility study in 2000. That file was reviewed by another branch in FDA, the General Surgery Devices Branch, who consulted with us on the file.

In late 2001, our branch took over review and the sponsor came to us to discuss a pivotal study. The study was given conditional approval in March of 2002, followed by full approval in May. We worked with the company on the protocol, and the study includes as you've heard both U.S. and foreign sites.

In 2003, when they had completed enrollment of the pivotal trial, the sponsor requested permission to conduct a continued access study which allows the company to continue to enroll patients while they're working on preparation of a PMA, and while the PMA review is ongoing.

For a number of reasons, the proposed protocol for the continued access study differ somewhat from the pivotal study. The continued access study was given conditional approval in June of 2003, and full approval in August. We received the PMA submission on January 27^th, 2004, and I'd just like to note that that file received expedited review status.

The ExAblate system is made up of the following basic components; patient table, operator workstation, software, equipment cabinet. The patient table is a standard MR table that has been modified to house the ultrasound transducer and associated equipment, and was already described by the company.

It should be noted that the MR system is a commercially available GE device, the Signa 1.5T MRI system is not commercially approved for thermography at the site. Software in the ExAblate device uses MR information from the GE device for mapping and targeting, as well as these new thermography functions.

This is the indication for use the company has already presented, but we'd like to go over this again. ExAblate is intended for use in pre and peri-menopausal women with symptomatic uterine fibroids. Patients must have a uterine size of less than 24 weeks, and be family complete. The fibroid or fibroids to be treated must be visible on non-contrast MR and should enhance on contrast MR imaging.

This is a list of the review team. As you can see, a number of people have been involved in the review of this PMA application in the areas of clinical, statistical, epidemiology, MRI, ultrasound software, bioresearch monitoring, patient labeling, human factors, and manufacture.

This slide lists the things that we look at during our review. For software and hardware we look at safety and effectiveness. Examples of safety issues for software and hardware include electric shock, EMI shielding, and unintended burns. Examples of effectiveness are adequate targeting and thermal dose delivery.

We specifically look at requirements in testing. We check to see that the device is designed to do what the sponsor or manufacturer says it will do. And we look to see that they do tests that check to see that it works the way it's supposed to.

For bioresearch monitoring, we look at study execution, including recordkeeping and informed consent administration, as examples. For manufacturing, we look at compliance with design controls both included in inspection. Bioresearch monitoring inspects clinical sites, as well as any records related to the conduct of the trial at the sponsor's facility. Manufacturing connects an inspection at the manufacturing facilities.

Bioresearch monitoring inspection is common for clinical trials, but is not required. A pre-approval manufacturing inspection is required. Drs. Corrado and Del Mundo will address clinical and statistical reviews during their presentations.

This is a list of our current major ongoing issues. This is not a comprehensive list of all issues. We are still discussing the thermal accuracy of the system with the company. Dr. Loren Zaremba will discuss this further in his presentation. We're still discussing adverse events that occurred, and appropriate medications to employ in response to these events. This will be discussed further by Dr. Noel Del Mundo. We will also discuss how the treatment in control groups differed, which Dr. Corrado will be discussing in her talk.

A pre-approval inspection is required, as I mentioned. FDA is working to get this inspection completed in a timely manner. Review of the labeling for a device is an integral part of the scientific review; however, we do not complete our review of labeling until we have finished the rest of our review of the file. These last two items, inspection and labeling will not be discussed further by other presenters today.

The rest of FDA's presentation will proceed as follows. Dr. Corrado will provide a summary of the clinical study and results. Dr. Zaremba will discuss the MR thermal mapping review. Bruce Herman will discuss the ultrasound-related review concerns, and Dr. Del Mundo will close FDA's presentation with a safety analysis discussion that will cover what we have considered most significant adverse events. Thank you for your time and attention, and I will now turn the floor over to Dr. Corrado.

DR. CORRADO: Thanks a lot, Kathy. Good morning, everybody. I'm Julia Corrado, and I'm a member of the review team.

You have all already heard about the clinical trials of ExAblate from Dr. Stewart and Dr. Tempany, and I am going to be covering some of the same material, but I'm going to try to give an FDA perspective on that material. And I will try very hard to avoid unnecessary redundancy.

I'm going to be starting with a brief description of the feasibility study. I will then describe in more detail the pivotal clinical study, and the aspects of that study as you see here. And finally, I will give a very, very brief synopsis of the continued access study.

I'd just like to say who the -- normally we don't spend much time talking about the feasibility study at panel meetings, but this one was especially important because it signaled to us a couple of aspects of this treatment that we really wanted to scrutinize closely when it came to the pivotal study.

This feasibility study was prospective. It was non-randomized. It was conducted at two centers, and I'll just digress for a second. Dr. Stewart described five centers. There was an IDE pivotal study that was conducted under FDA approval, and that was conducted at a center in the U.S. and one in Britain. And I'm speaking just about that feasibility study in my next couple of slides.

It was a pre-hysterectomy study. The women who volunteered were scheduled for hysterectomy, but they agreed to undergo the ExAblate procedure approximately a month prior to hysterectomy. And we approved the study for 15 subjects and 13 subjects received treatment.

The objectives were already described by Dr. Stewart. There were, in general, two types of tissue effect that are noted from ExAblate. I won't speak about them further, but there is a thermal coagulative necrosis and then there is an ischemic necrosis. The difference is that the thermal coagulative necrosis is caused by direct heating, and the ischemic necrosis results from lack of blood flow to surrounding tissue following heating.

In the summary of the feasibility study, the pathologist from Brigham & Women's described the tissue effect as follows; that the volume of necrosis was sometimes larger than the treated area. That's a very important point that I'm going to be emphasizing. The treatment effect consists of bland and highly uniform coagulative-type necrosis with relatively sharp outline, scattered interstitial hemorrhage, and variable amounts of acute inflammation consisting mostly of neutrophils.

The next point also should be noted, and that is that the only abnormality noted in the myometrium outside of the fibroid, this was beyond the fibroid capsule, was microscopic coagulative necrosis extending one to two millimeters beyond the fibroid. This is the only case where we saw this effect, that there was a treatment effective beyond the fibroid capsule. But nevertheless, we thought it was important, as I'll describe further.

The purpose of the next slide is to illustrate something I just hinted at, and that is that the volume of effected tissue is different from the thermal dose volume; that is, the volume that was actually targeted. And there are two volumes that we can talk about from the feasibility study. One is the non-profuse volume immediately following treatment. This is on, I believe, T1-weighted images with Gadolinium enhanced MRI. But also from this population, most of these women underwent hysterectomy so we also have volumes from hysterectomy specimens. And what I'd like you to notice here is that there is a consistent -- the non-profuse volume and the volume from histology are consistently greater than the thermal dose volume, which led us to feel that we wanted to be cautious in how the pivotal clinical study was conducted because we did not want to get injuries resulting from tissue necrosis beyond the targeted area.

As always, as we would expect during any kind of a clinical study of an investigational device, problems were encountered during treatment. For example, several patients received what was described as sub-optimal treatment due to excessive fat layers within the beam path. And in one case, the portion of the fibroid that the clinician wanted to treat was too close to intestine, and that limited treatment in that case. In three cases, patients did not receive treatment due to tissue aberration and scar in the beam path that caused the patient to experience pain.

FDA, of course, always looks closely at adverse events and clinical trials, and we saw the following. But before I go into these adverse events, let me just note that despite that enhanced volume effect that I have described, we did not see any evidence of thermal injury to tissue adjacent to the uterine serosa, and this is one of the types of adverse events that we always watch very closely in devices that treat uterine pathology, so we did not see any such adverse events.

What we did see was bleeding post ExAblate, two first degree skin burns, a couple of cases of nausea and vomiting, and some post-hysterectomy adverse events that we would not be able to argue were related to the treatment. They were probably related to the hysterectomy.

As Dr. Stewart mentioned, there is also feasibility data from outside of the United States. And interestingly, in this study although 56 patients received the ExAblate treatment, only four of those patients elected to undergo hysterectomy, and that was as of 14-month follow-up. So there is relatively less hysterectomy data from this feasibility study population.

The next couple of slides I'm not going to spend much time on, but I would just like to say that they demonstrate a trend, at least, towards non-profuse volume being greater than the thermal dose volume, although it was not uniform as it was in the smaller feasibility study conducted at Brigham & Women's and at St. Mary's in London.

In the feasibility studies that were conducted in Israel, again this was not conducted under FDA IDE regulation. However, there was one adverse event that in hindsight we probably under-appreciated at the time, and that was a case of sciatica post treatment. This patient had symptoms as of three weeks following her treatment, which at that time were described as improving, and at that time she was referred to a neurologist. I'm going to at least allude to this adverse event later in my discussion.

Dr. Stewart described the IDE pivotal study of ExAblate, and I'm not going to repeat what she said, nor will I repeat the primary and secondary hypotheses, with the exception that I want to note that the secondary hypothesis here, which was in valuation of the trajectory of recovery in the two treatment groups. The sponsor has already described this. I am not going to talk about that any further.

Now FDA worked with the sponsor on the pivotal study design, and we really perseverated on what we consider to be potential for adverse events with this device. We were nervous about the potential for tissue necrosis of non-targeted tissue; in particular beyond the uterus of the necrosis of tissue up adjacent to the uterus, so we worked with the sponsor to establish a very conservative treatment planning program. And the list of items that contributed to this was already discussed by Dr. Stewart. But, for example, we felt that because this is a very new type of technology combining MRI thermography and focused ultrasound, and because we have seen this volume effect that was greater than the targeted volume, we felt that it would be prudent to begin here with limiting the volume of tissue that could be targeted both within individual fibroids and within the entire uterus.

I also want to add an important point here. Our concern at the time that we were reviewing this IDE for the pivotal study was in what I would call near-term thermal damage. We were not, at the time, sensitive to the fact, or that we might get treatment effects in the far field. That is beyond the area of focus, so that's going to be important later in my talk, and critical when you hear from Dr. Del Mundo later this morning; that we did not appreciate the potential for effect in the far field.

Very generally, I'd like to just reiterate a little bit about the baseline demographics between the two populations. There was no difference in age, essentially. The body mass index was higher in the hysterectomy group. There was a significant difference with respect to race, and with respect to other chronic disease, there were some differences between the ExAblate arm and the control arm. And specifically, women in the control arm had significantly greater prevalence of diabetes mellitus, hypertension, and anemia.

I just want to make a couple of simple points with respect to this slide. First is that you've heard about the symptom severity score of the uterine fibroid symptom quality of life questionnaire. It was the subset of that questionnaire, the symptom severity score that formed the basis for the primary endpoint in the study and the definition of success of the study.

There were, however, two scores that were taken prior to treatment. There was a screening score to determine whether or not a patient was eligible for the study. And then there was a score that's called the baseline score. That's in maroon color on this slide. And that baseline score formed the comparison for the six-month evaluation, so the screening score was only relevant to get into the study. After that, it was the baseline score that was relevant to the study success. But we thought that it would be interesting for you to see that even before treatment, there were subtle differences in the scores that were derived from that questionnaire.

And what this slide simply shows is the distribution. Right in the very middle of the X axis you see the word "unchanged", and right above that "zero." And what this means is the 23 subjects had no change in score between that screening questionnaire and the baseline questionnaire. Beyond that, there is a very roughly equivalent distribution on either side of zero, but nevertheless, you can get the feeling for actual numbers of subjects whose scores changed by a particular range of points prior to treatment. Again, this is all prior to treatment. And we think that this will give you a feel for the stability of the scoring instrument.

Okay. Now I would like to change gears a little bit and talk about study success. And I just want to mention that there are two ways to look at study success. Intent to treat has a very strict definition. Essentially, it is all patients enrolled and treated, and all of these patients must be represented in calculating the percentage success.

Evaluable is a different way of looking at it. The rules for an evaluable analysis are not, frankly, as tight as they are for intent to treat, but what evaluable permits one to do is to not be penalized by counting subjects as failures who one can argue really shouldn't strictly be viewed as failures. But, nevertheless, from the most strict definition standpoint, intent to treat, the percent success was 70.6. Remember that the primary hypothesis was that greater than 50 percent would improve by a score of 10 points or better at six months. And the 70.6 is well within the 95 percent confidence interval. It's well above 50 percent.

And this slide again indicates that for those subjects who achieved success, you can get a feeling for how many of them met that primary endpoint and by how many points. And what this also shows you is that there were some subjects whose scores were considered unchanged, and then some whose scores were considered worse. And just because often we are interested in well, what happened with those failures, I'm just going to try to give you a quick rundown for the population at six months for whose scores worsened and whose were unchanged.

Among patients who were considered to have worsened scores were two treatment failures. These subjects underwent hysterectomy. Two patients withdrew, and subsequently one had hysterectomy. One withdrew. She had an aborted treatment, but nevertheless, she had a worsened score. One subject was lost to follow-up, and had no six month quality of life data. Seven women in this category completed treatment, and they had actual worse scores following treatment at six months.

Unchanged is a little bit different the way we look at that. We asked the sponsor to consider seven patients to have had zero change who were treated with protocol deviations, because we felt that the success rate for -- we feel that the success rate for any study ought to reflect patients who are treated according to the strict rules so that we can write meaningful labeling, so the clinicians can understand what they might expect. And that it would be inappropriate to have the results reflect women who were treated outside the bounds of that treatment protocol, so that explains seven of the subjects with unchanged scores. Two had no six-month data, and one actually had no score change.

And I should have mentioned, we have a patient tree in your day-of folders, and it might help you if you're interested in looking at those numbers more closely in understanding what patients from the original intent-to-treat populations fell into which categories.

Now I'd like to just talk about patient satisfaction. You've heard from the sponsor about overall grouping of patient satisfaction and the ExAblate group in general had high levels of satisfaction. What I wanted to do here was just give you a slightly different perspective, and in the interest of doing that, here you can see satisfied being broken down into three different categories, very moderately in some. And similarly, patients who describe themselves as dissatisfied, what level of dissatisfaction did they experience. And for very satisfied, the point here is that the hysterectomy subjects at six months were significantly more apt to say they were very satisfied, compared to the ExAblate group, which is probably not surprising because the hysterectomy subjects did receive definitive treatment for their fibroids. But nevertheless, we think that it's important to keep perspective when you're thinking in terms of patient satisfaction. There are degrees of satisfaction, and they did differ between the two groups.

Now at 12 months, you've heard that both the intent-to-treat and the evaluable success rates dropped. I should preface this by mentioning that when the study was designed, FDA had a different understanding regarding the length of follow-up than the sponsor had. And it was a misunderstanding. We expected three years of follow-up. The sponsor believed that they were expected to follow the patients through six months, so in all honesty, we informed them very late in the pivotal study that they would be required to follow-up these patients for up to three years. And it created some difficulty for them in terms of tracking down patients, and asking patients if they would continue to participate in the study, so I think it's only fair to mention that. And that helps to put some perspective on the next slide where I talk about some of the women who were not successes at 12 months, in part from an intent-to-treat analysis. If they declined to participate, they're considered failures; withdraw or lost to follow-up, non-evaluables.

Significantly, though, we think it's important to note that 23 women did have alternative treatment as of 12 months, and many of these women had hysterectomy by 12 months. Four of them had a second ExAblate treatment.

You've already seen Dr. Stewart's slide that covers this material. There are a couple of things that I just want to highlight. One is that the percentage of non-perfused volume of 23.6, that was the average percent of non-perfused volume that is immediately following the treatment. And then at six months, the volume of the treated fibroids was measured, and the percent shrinkage of the treated fibroids was 15.3 percent.

In looking at these numbers and thinking about them, it's worthwhile again to think about the fact that the treatment guidelines for the ExAblate procedure only permitted the sponsor to treat up to 33 percent of any individual fibroid, and only up to 100 Ccs in any one fibroid up to 150 Ccs for an entire uterus, so I think that part of what we're seeing there reflects the conditions of the treatment.

You all have in your folder a list of discussion questions that you'll be talking about after lunch, and I want to bring to your attention Discussion Question 1 now, because it relates to some of the things I've been talking about. And it asks you how you view the symptom severity scale of the UFS-QOL as the instrument that was used to measure the primary endpoint in the study, or that was used to determine study success.

I'd also like to draw your attention now to Discussion Question 2, just to kind of plant the seed that this afternoon you're going to be talking about essentially whether the study demonstrated the effectiveness of the procedure. And in making that assessment during that discussion, FDA would request the panel would consider the results from the symptom severity score. And also, couple that with the clinical results of actual volume reduction in making your decisions.

I'm going to change speeds once again now, and briefly talk about safety-related aspects of the procedure. But I want to preface this by saying I am going to ask you to focus at the end of my discussion on two types of adverse events, and they were skin burns and nerve injuries.

As Dr. Stewart described, we worked with the sponsor prospectively before they started the pivotal study to identify what we thought would be a legitimate list of adverse events or complications against which to compare the two study populations. And the result of that comparison, not very surprisingly, showed that there were relatively fewer of the significant clinical complications in the ExAblate group compared to abdominal hysterectomy patients.

There were some other adverse events that are possibly more relevant to the ExAblate procedure, and those are pain and discomfort, in particular, during the procedure, menorrhagia post procedure, urinary symptoms, nausea and vomiting. And then the last two, skin burns and nerve injury that we do believe are unique. We believe that these are unique to this procedure, to ExAblate.

And I just want to conclude my discussion of the safety evaluation in the pivotal study by pointing out that in Discussion Question 6, you're going to be looking at the relative safety of two procedures, ExAblate compared to total abdominal hysterectomy. We just want you to include in your deliberation, or at least address the differences between the study arm and the hysterectomy arm at baseline, because there were some differences, and provide your input to FDA with respect to how comparable these two groups are, and what types of conclusions we might be able to reach or not reach regarding relative safety of ExAblate versus total abdominal hysterectomy.

Now as you're aware, the sponsor has permission to continue to treat patients with the device, although the pivotal clinical study is over. And of the aims of this, in addition to allowing continued access, was to modify the treatment planning in addition to improving long-term follow-up. But we wanted to essentially liberalize the parameters for treatment, and you can see that the sponsor is now allowed to treat up to 50 percent of an individual fibroid volume, as long as it's not a sub-serosal fibroid. And the 15 millimeter margin, in retrospect, didn't make much sense with respect to the endometrium, so that has been eliminated. In addition, the sponsor may perform one additional treatment session. And as you see, increased maximum duration of the treatment.

As of March, 89 patients had been treated. The baseline demographics are similar to the pivotal study, but there are three months safety data available on only 53 to 54 subjects, so it is these subjects who have been followed up for at least three months who are going to be discussed by Dr. Del Mundo.

The preliminary efficacy data are good, 79 percent reported 10 point improvement in their symptom severity score.

In the continued access study to-date, there have been two instances of sonication-induced leg pain, and the significance of these events, whether or not they are nerve injuries, will be described by Dr. Del Mundo.

In conclusion, I have attempted to give you an overview of the effectiveness of the device as seen in the pivotal clinical study, and to a limited degree discuss the safety profile of ExAblate as FDA understands it at this time. I have indicated to you that two types of adverse events appear to be unique to ExAblate, and that is skin burn and nerve injury. And as you recall, there was a nerve injury even in the feasibility population, but at the time, that was a case of sciatica. At the time, we did not appreciate how it might be related to this device.

Before Dr. Del Mundo presents a detailed FDA analysis of those injuries, Loren Zaremba and Bruce Herman are going to discuss for you the physics of MRI thermography and focused ultrasound, and the results of some modeling, and how modeling can help us understand the potential for heating of tissue in the far field using ExAblate. Thank you very much.

DR. ZAREMBA: Good morning. My name is Loren Zaremba. I'm an MR reviewer in the Radiology Branch, Office of Device Evaluation. This morning I will be discussing the role of thermal mapping in the focused ultrasound of uterine fibroids using the ExAblate 2000. I will discuss the advantages and limitations of MR thermal mapping, and the safety and reliability concerns with respect to the use of MR thermal mapping for this intended use.

MR thermal mapping provides three major functions in the ExAblate 2000. First, it allows adjustment of the ultrasound focus location. This is done by obtaining an image of the temperature distribution produced by a low power sonication. Second, it provides a measurement of the temperature distribution during the treatment procedure. Third, it provides feedback which enables the user to adjust the power following a sonication if the temperature was too high or too low.

Temperature measurements are necessary with any type of therapy which uses heating. Previous devices, such as those approved for hypothermia have used temperature probes. The ExAblate uses a new approach, MR thermal mapping, which has the advantages that it does not require surgical implantation. It can be integrated with the MRI system, which is used to visualize the uterine fibroids, and it provides a temperature over the full MRI field of view, not just a few points.

Also, unlike temperature probes, it does not cause heating at the probe tissue interface which can lead to an overestimate of the temperature with those types of probes if they are not corrected.

The limitations of MR thermal mapping are, first, it does not measure the actual temperature, but change in temperature. Second, it cannot make measurements in bone or fat. Third, a very small amount of motion by the patient during the three seconds required for MR thermal mapping can spoil the measurement. Four, MRI thermal mapping has lower time and spatial resolution than temperature probes. And five, calibration can present some difficulties.

I want to direct your attention to the Discussion Question 3, has the sponsor demonstrated the MR thermal mapping provides adequate interoperative feedback during treatment regimen to ensure safe and reliable dosing to the intended fibroid?

With regard to safety, we have these considerations; can temperature measurements be made in all regions of interest? Are they sufficiently accurate? Can they be made in time to allow adjustments? And with respect to reliability, how frequently does thermal mapping fail? If it fails, is adequate backup provided?

With regard to the first safety factor, the ability to measure temperature in all regions of interest, the critical data showed that the ExAblate is capable of measuring temperature in the principal region of interest, the uterine fibroid, and in most surrounding tissues. However, it cannot measure temperature in the sacral nerves due to the fat surrounding these nerves ,and also near the bone tissue interface where heating can be intensified due to the very high ultrasound absorption by the bone.

In a later presentation, Dr. Del Mundo will discuss the adverse events during the clinical trial that may be related to nerve heating. Since MR thermal mapping cannot provide the answer, we must rely on thermal modeling to estimate nerve and bone heating. The next speaker, Bruce Herman, will discuss the modeling and results that have been done by FDA with respect to nerve and bone heating.

Accuracy is relevant to the evaluation of the ExAblate 2000 because if an incorrect reading of temperature is given to the user, they could adjust the power level higher, which could result in injury, or lower, which could result in inadequate treatment. MR thermal mapping measures the temperature change not temperature. The ExAblate assumes that a sufficient time has elapsed following a sonication that the temperature has returned to a baseline of 37 degrees Centigrade. The company recommends that the user wait 90 seconds before initiating the next sonication to allow the temperature to return to baseline. However, this can be adjusted by the user.

If an adequate cool-down time is not selected, the heating induced by the previous sonication will add to that induced by the current sonication, but this will not be shown by the thermal map. This is of particular concern with regard to the sacral nerves because modeling shows that the return to baseline may take longer in the nerve region as will be discussed in the next talk.

A second issue relating to accuracy is temporal or time resolution. This is just the amount of time it takes to make the measurement compared to the heating period. The ExAblate requires a little over 3 seconds to obtain a thermal map. The fibroid can be heated very rapidly, and temperature can rise 10 degrees in the time needed to obtain a thermal map. And the MR thermal mapping, the result is the average temperature rise during measurement peak rather than the peak. For all measurements but the final one prior to the termination of the sonication, the ExAblate assigns the temperature to the midpoint of the measurement interval, which partially corrects for this. However, this is not done for the final sonication.

The third issue relating to accuracy is spatial resolution. In MRI, this is determined by pixel size, which is about 1 millimeter by 2 millimeters for the normal ExAblate field of view. Focused ultrasound produces very high temperature gradients, which means the temperature falls off very rapidly with distance from the focus. Thermal mapping averages the temperature over a pixel, and that could result in an under-estimate of the maximum temperature for a small focal spot, or an under-estimate of the size of the region ablated by sonication. However, for the large group of spot sizes normally used in treatment of uterine fibroids, this is not a serious concern.

The last issue relating to accuracy is calibration. In the case of MR thermal mapping, calibration enables us to translate the observed change in proton resonant frequency with temperature into the temperature change. In the ExAblate, the calibration factor relating the frequency change to temperature change is assumed to 0.009 parts per million per degree C, independent of the amount of temperature rise, tissue type, or thermally induced changes in tissue.

A calibration for MR thermal mapping involves comparing the mapping results with an independent temperature measurement method, such as a thermal couple probe. Ideally, the calibration for the MR thermal mapping used in the ExAblate should be done for uterine fibroids in human subjects in order to derive some indication of the variations between fibroids, subjects, and other conditions of treatment.

We have an in vitro study using tissue samples heated by a water bath, which indicates a variation of 3 degrees. And we have an in vivo calibration in rabbit muscle which indicates a variation of 10 degrees.

One of the purposes of MR thermal mapping is to provide feedback to allow adjustment of the power following a sonication. However, there is a delay in feedback from the thermal mapping. The temperature versus time graph in the thermal ridges for a sonication are not displayed until the sonication is complete. Consequently, a correction cannot be made until the next sonication.

Among our reliability concerns is the fact that a very small amount of patient motion can result in a loss of a thermal map. The temperature measurement cannot be repeated because this would mean resonicating the same spot.

The failure rate for thermal mapping appears to be quite low. InSightec estimates it is only about 4 percent, and since the treatment consists of a large number of sonications, this may not be a concern. The user is instructed to check the fiducial markers to determine if movement was sufficient to affect the treatment.

Another reliability concern is the availability of an alternate means of assessing its effects over the treatment; i.e., a backup method. If the thermal maps are lost, the so-called magnitude images may not be adversely affected. Magnitude images display signal strength and are not as sensitive to motion as thermal images. However, the magnitude images may be of limited usefulness in displaying the effects of the sonication on fibroid tissue. The final confirmation of the effect of the treatment is contrast enhanced image obtained after the treatment.

In summary, MR thermal mapping provides significant advantages over other available technologies in that it is non-invasive, can be integrated into MRI system use to visualize the pathology, provides a thermal map over the full MRI field of view, and does not interact with ultrasound. The major limitation is that it cannot measure temperature in bone or fat, which prevents estimation of the heating of the bone and sacral nerves in the far field. The limitations associated with sensitivity to motion and lower temporal and spatial resolution are not serious. And the calibration of the method can probably be improved with additional studies.

I would now like to turn the discussion over to Bruce Herman, who will describe the thermal modeling that has been done.

MR. HERMAN: Hello. My name is Bruce Herman. I'm a Physicist with the Office of Science and Engineering Laboratories within CDRH. I'll be discussing the thermal effects distal to the focus using the ExAblate as regards possible adverse thermal effects. My presentation will not be an exhaustive discussion of every concept that might affect the thermal modeling, but I hope to give a relevant background and orientation.

I will be discussing the concept of thermal dose, some factors affecting the temperature rise of the sacral nerve, and the bone, and the so-called far field of the ultrasound beam. I'll be talking about the limitations of the knowledge regarding tissue characteristics, which are relevant to modeling the temperature rise in these structures. I'll give a couple of temperature rise simulations, and I'll talk about the limitations of these models.

As Dr. Zaremba mentioned, it's important because magnetic resonance thermal imaging can not determine the temperature rise near bone or in fat which might typically surround nerves, which means that theory, i.e., modeling, plus, of course, clinical trial results are very important to assess the safety.

In most biological systems for temperature rises -- with temperatures above 43 degrees, which corresponds to a 6 degree temperature rise assuming the baseline temperature of 37 in the body, each temperature rise of 1 degree C requires housing the exposure time to achieve the same level of effect. If the temperature is a varying function of time and T-43 is the time necessary to achieve an effect at 43 degrees CI, a 6 degree temperature rise, and the time necessary to achieve an effect for a time varying temperature is given by this integral equation. It's not just the peak temperature that's relevant, but the temperature and the time over which a particular organ sees that temperature, which is relevant to assess the propensity for any type of damage due to an organ or structure.

This gives a report of thermal dose thresholds for cell damage for certain tissue types. We receive for muscle, fat, and fibroid, typically 42 degrees C. The time necessary to see an effect is 14,400 seconds, which corresponds to 240 minutes. At 51 degrees C, though, the time required drops to 56 seconds, 53, 14, 55 degrees Celsius, to 3.5.

With more sensitive structures, such as nerve, colon, or intestine, reported times for effects at 42 degrees C have been between 1,500 and 3,600 seconds. This corresponds to a time of 10 seconds at 51, 2.4 seconds at 53, .6 at 55. Rule of thumb might be for these more sensitive structures, as you get into the lower 50s, you might begin to see some type of damage.

Of course, for all these structures when you get to be above 65 degrees C, the damage occurs almost instantaneously, and you get pretty much ablation almost instantaneously, as has been mentioned in previous presentations.

This slides shows - and it's not drawn to scale - shows the ultrasound transducer focusing the beams with very high intensity within the region of treatment within the fibroid. The beam distal to the focus then spreads out, and the intensity is lowered both due to the spreading-out of the beam, and due to absorption in the intervening layers.

In the far field of the beam, you might have a structure such as sacral nerve surrounded by fat, and it might be near a bone. You can think of the absorption in let's say the sacral nerve as a combination of two types of phenomena; one, direct absorption in the nerve and in the fat surrounding the nerve. And if it's near the bone, then since the bone is such a highly absorbing material, as we'll see, it then will conduct heat to the sacral nerve after absorbing a lot of the ultrasound energy.

As the nerve gets closer to the bone, this phenomena might become prominent, predominant. And as, of course, it gets further away from the bone but closer to the focus with higher intensities, the direct absorption might be the predominant mechanism of temperature rise.

The temperature is a function of, of course, the local intensity, the absorption of ultrasound by the structures, the incidence of the ultrasound beam on the bone. I mention this because if an ultrasound beam is normally incident on the bone, most of the energy is absorbed by the bone. But if it's offset or comes at an angle greater than about 30 degrees to normal, most of the energy is actually reflected and not absorbed, so the bone heating would not be a strong factor.

Of course, the beam restructures size, as we'll see is important. The thermal characteristics of the tissue, such as thermal conductivity, heat capacity are important, and the geometry, the size, how close one structure is to another. I want to emphasize, basically it's a complicated phenomena, multi-parametric phenomenon.

This slide shows the range of reported tissue absorption value. As you can see, for various tissues, there's a wide range of reported values. This is important because for a lot of structures, the direct absorption, the temperature rise due to direct absorption is approximately linear with the absorption value, how much of this energy it absorbs.

Now these red diamonds are the values used by InSightec in their modeling. Now they are commonly accepted values for tissue absorption, do cluster around the red diamonds. But I wanted to emphasize that these values, which will be used in modeling you'll see, come with very large error bars.

We talked about that the size of a structure might be critical in determining the temperature rise. This shows, for one, soft tissue absorption model, that for the stated incident intensity, the temperature rise after 20 seconds is a strong function of the dimensions of the structure, for something on the order of 1 millimeter, .1 centimeters, the temperature rolls over very quickly. And after 20 seconds, the rise is not very great. Whereas, for a large structure, such as the 3 centimeter structure, the temperature stays linear up to 20 seconds, and you get a much higher temperature rise. We use 20 seconds because that is the sonication time used by the ExAblate.

This is a simulation which was actually done by InSightec which shows the temperature rise at a sacral nerve 3 millimeters from the bone, and surrounded on all sides by 3 millimeters of fat. It utilizes a focus-to-bone distance of 40 millimeters, which again, InSightec showed a slide that showed this is the current protocol for use, to keep the bone at least 40 millimeters away from the focus, and it uses a power of 250 acoustic watts. This is the maximum power that this transducer can put out, so this is a worst, worst case.

The red curve shows the temperature right at the sacral nerve without the bone and so gives an indication of the direct absorption, the temperature rise due to direct absorption of energy, and the blue line shows the temperature rise with both contributions, the bone and the direct effect. As you can see, the temperature rise for this case get into the mid-50s or the low 50s. But again, I want to mention that if we assume a higher absorption than was used, this red curve could be quite a bit higher than was assumed in this model.

This next, again an InSightec model, shows the same situation, but here the protocol demanded focused temperature is equal to 85 degrees. Again, the protocols currently demand that the temperature at the ablation focus should be no greater than 85 degrees. And, of course, this would consequently lower the total power needed. And this gives the temperature rise again at that same sacral nerve with and without bone.

As you see, the temperature rises are lower. It goes from 37 to 41 without bone, and goes to about 44 with. But again, I want to emphasize that the if the absorption values used are higher, these temperatures will go up. If the sacral nerve is closer to the focus, let's say it may be 13 millimeters away instead of 3 millimeters away, this bone contribution will be less, but the contribution due to direct absorption could be as much as 75 percent higher. Again, emphasize the complexity of the situation.

Actually, I do want to mention the fact that we see that we have a 20 second sonication time and then a 90 second cool down time. And this model shows that even after the 90 second cool down time in the far field, there's still a significant temperature rise, which may mean that if there is any overlap of two consecutive sonications on the bone or on the structure, you might get some addition of temperature to the second sonication from the first.

This is a model done by CDRH. This shows a temperature rise at the bone tissue interface again for 85 degrees for the focus, and a focal bone distance of 40 millimeters. It assumes no fat between the focus and the bone, so the maximum energy hits the bone. Again, this has arised at the bone tissue interface. As you can see, we have two consecutive pulses, 20 seconds on, 90 seconds cool down. Temperature rises can be very high, so if by chance a nerve structure is right at the bone, it can experience quite high temperatures due to bone heating. Again, a significant temperature rise after the 90 seconds, so you might have a partial additive effect if there's an overlap.

Again, I want to emphasize these last three slides assume normal incidence of the ultrasound beam on the bone, meaning maximal energy absorption by the bone. The current protocols used by ExAblate try to maximize the angle of the beam on the bone to avoid this absorption by the bone. And, of course, these temperature rises would go down consequently if the beam did come in at an angle of 30 degrees or more from the normal.

Okay. The factors that may cause temperature rises higher than those models by CDRH or InSightec could be higher absorption values than assumed, larger structures than assumed, structures closer to bone or focus, possible inaccuracy of the temperature map at the focus. If it reads low than a higher intensity to cause 85 degrees to the focus might be used and is necessary, which would increase the intensity in the far field. Incorrect thermal conductivity or heat capacity, and possible overlap of consecutive sonications.

Now in conclusion, the modeling using generally accepted values for tissues parameters, together with the discussed protocol caveats, predict reasonably that thermal events of adverse effects in the far field should be very rare. But given the range of imported and possible actual variability of tissue values, the individual range of structure geometries, the accuracy of MR, et cetera, this modeling in and of itself does not allow adverse thermal effects to be totally ruled out, which of course means that clinical results take on added importance in assessing the accuracy of the modeling and the actual risk benefit. And I gather that the clinical situation is consistent with these conclusions.

As regards this, Dr. Del Mundo will shortly present specific adverse effects that have occurred during use of the ExAblate, and will discuss how we and InSightec have used this modeling to try to understand and prevent such events. Again, as regards this also, you will be presented with a discussion question for this afternoon. Basically the question is, do the results from the thermal modeling and our understanding of the underlying physics allow sufficient information to understand the etiology of the injuries that occurred in the study and, of course, to mitigate their occurrence? Dr. Del Mundo will now give his presentation.

DR. DEL MUNDO: Thank you, Bruce, and good morning. I'm Noel Del Mundo, Medical Officer in the OB-GYN Devices Branch. I will be presenting the safety analysis of sonication-related adverse events that occurred during the pivotal trial.

I will focus on the description of the types and severity of skin burns and nerve injuries that occurred during the pivotal trial. I will provide the analysis of possible causes of the sonication-related adverse events, and I will then go through the list of possible mitigations to prevent each type of adverse event, and I will provide the preliminary safety results from the continued access study in which the mitigations were implemented.

Of the adverse events that Dr. Corrado had previously summarized for you, the most notable sonication-related adverse events were skin burns and nerve injuries. IN all, there were five first or second degree skin burns during the pivotal trial. Improper acoustic coupling between the skin and the gel pad can result in undesired heating of the skin. In other words, any areas between the skin and the transducer that allows for increased reflection of the ultrasound energy can cause heating of the skin and possible skin burn. Examples are air bubbles present in the skin folds and around the hair, oil between the skin and the transducer.

In all the cases of first and second degree skin burns, all patients had hair in the sonication pathway. And also, early in the pivotal trial the patient moved and decoupled from the acoustic gel, resulting in a first degree skin burn.

The steps in the training manual to reduce the risk of skin burns want the patient to shave the hair from the lower abdomen to two centimeters below the pubic synthesis. The abdominal is cleaned with alcohol to remove oil on the skin, and patient movement is limited with a table strap. And lastly, the MR planning images are examined for air bubbles at the skin-gel interface and for skin folds.

These steps were re-emphasized to the investigators during the pivotal trial and prior to the continued access study. Preliminary results on 54 patients treated in the continued access study suggests that the mitigations and retraining have reduced the incidents of skin burns as no cases of skin burns have been reported.

I'd like now to focus on the nerve injuries. These injuries have been the subject of extensive review by FDA and InSightec. In all, there were five cases of sonication-related nerve injuries during the pivotal trial, and the patient's symptoms lasted anywhere from two days to twelve months.

In addition to the five nerve injury cases, there were three cases of what we're calling nerve stimulation. These cases differ from the nerve injury cases in that the leg pain did not extend beyond the day of treatment, but we think that in the continuum of unintended heating of the nerve by unfocused ultrasound, these cases represent the mild form of heating of the sacral and sciatic nerve in the far field of treatment. This is a point that I'll get back to in subsequent slides.

Now getting back to the five cases of nerve injuries, InSightec analyzed these cases and found that common to all five cases were the following. Lower extremity pain was acutely felt by the patient during the treatment. The distribution of pain is consistent with either sacral of sciatic nerve injury. There was rapid onset of pain during the last three to five seconds of sonication, and the sacral nerve or sciatic nerve bundle was identified in the far field of the ultrasound beam.

There appears to be varying degrees of peripheral nerve injuries related to sonication. A mild effect is nerve stimulation resulting in leg pain which resolves the day of treatment. The next degree of effect is nerve injury resulting in the interruption of nerve function without anatomic discontinuity axon. This injury can take days to weeks to recover.

The most severe form of sonication-related nerve injury we have seen in the pivotal trial resulted in the interruption of the axon, requiring regrowth of the axon. This injury can take months to recover and considered permanent if symptoms persist for greater than two years.

This worst case was that of Patient 919, which occurred near the end of the pivotal trial. The patient complained of leg pain at the completion of the sonication treatment, and developed left lower extremity weakness. She had difficulty walking and climbing stairs. She had numbness and tingling from her left calf to the dorsum of her left foot.

She was evaluated by a neurologist at six months, and physical examination revealed that nerve injury consistent with neuropraxia had resolved. However, minor deficits present at six months due to axonal loss would recover over a much longer time point as the axon has to regenerate from the pelvis all the way to the target muscle.

Evaluation at 11 months by the same neurologist showed that the patient had almost fully recovered, except that she was unable to flex her left toe. She had otherwise returned to her baseline level of activity.

Now because of the symptomatology and because of the location of the sacral nerve bundle in the far field of the beam, it's believed that this patient sustained injury to the sacral nerve bundle located in close proximity to the sacrum. This axial MR image of the pelvis is to show the proximity of the anatomic structures of concern. This is not an image taken from an actual treatment of Patient 919.

To orient the audience, the patient is laying face down, and at the bottom of the screen is the abdomen, and at the top of the screen is the patient's back. The structures of particular interest to us are the sacrum pointed to in a dark blue arrow, and the sacral nerve, 4 centimeters away from the treatment volume pointed to by the red arrow, and the treatment volume represented by the rectangle, pointed to by the orange arrow.

As was mentioned by Dr. Zaremba, since MR thermography cannot provide temperature measurements at the bone or at the interface between the nerve and the bone, the company has provided temperature modeling to help explain how nerve injury could have occurred in Patient 919. This temperature graph is slightly different format from that that was presented earlier by Bruce Herman. This graph shows the temperature as a function of distance from the transducer. The temperature graph depicted assumes that the angle of sonication is perpendicular to the sacrum, and the baseline temperature is at 37 degrees Celsius.

When the focused ultrasound causes temperature to rise to 85 degrees Celsius at the treatment focus in the fibroid, unfocused energy in the far field causes the temperature to rise at the nerve to 42 degrees Celsius, and 55 degrees Celsius at the sacrum.

Following the caveats previously mentioned by Bruce Herman, if the peak temperature at the treatment focus is higher than 85 degrees Celsius, the temperature at the nerve and bone will be higher proportionately. Conversely, if the incidence angle is turned away from the perpendicular to the sacrum, the amount of absorbed energy to the bone will be less, decreasing the rise in temperature at the bone. And also, the temperature at the nerve will increase if the nerve is closer to the bone.

The previous graph had assumed that the baseline temperature was at 37 degrees Celsius throughout the beam path. Now this temperature modeling is of the nerve tissue 4 centimeters from the treatment focus. The red line shows that if the nerve is in close proximity to the bone, the temperature will be at 39 degrees Celsius at 90 seconds after completion of the treatment.

The blue line, on the other hand, shows that if the bone is in close proximity to the nerve, the temperature will be 42 degrees Celsius 90 seconds after the completion of the treatment.

Now if the length of cooling time is not extended beyond the nominal 90 seconds, the cumulative effect of this small difference in temperature can be significant after a series of sonications.

Bruce had previously shown this temperature graph of the nerve in very close proximity to the bone. It illustrates the concern for adequate cooling time between sonications that I had mentioned in the previous slide. From a practical clinical standpoint, as the user becomes more efficient at targeting and sonicating a focus in a fibroid, the nominal 90 seconds of cooling time becomes insufficient to allow the sacral and sciatic nerve to return to baseline temperature before subsequent treatment is initiated; thus, increasing the risk for nerve heating and injury.

From the temperature modeling, the possible steps to prevent heating of the sacral and sciatic nerve are to alter the incidence angle of the beam to decrease the amount of absorbed energy at the bone, establish a minimum distance from the treatment focus to the sacrum, lower the peak temperature at the treatment focus, and possibly increase the cooling time between sonications to allow the bone and nerve temperature to return to baseline before subsequent treatment.

Of these, the company has implemented in the training module, a change in incidence angle away from perpendicular when the sacrum is 4 centimeters away from the treatment focus, and to maintain a minimum distance of 4 centimeters when the sacrum is in the far field of the beam.

Now working with the company, we recently compiled the incidence angle and distance data to see if the current mitigations if implemented in the treatment of the five patients with the nerve injury would have prevented the nerve injury. From this chart we can see that the most severe case could have been prevented, but clearly, two cases meet the requirement greater than 4 centimeter distance, and greater than 30 degrees change in the incidence angle.

The question is, are additional steps such as lowering the peak temperature at the treatment focus and increasing the cooling time warranted to decrease the risk of nerve injury?

Now we can also look at the preliminary results of the continued access study to see if the mitigations are working, and so far there has been one case of a patient with symptoms consistent with nerve stimulation, and one case of nerve injury that resolved two days post treatment. It's reported that this patient experienced warmth down the right leg during two to three sonications, and that one day post treatment she felt her right foot hitting the ground harder than the left, and she had stumbled once.

Now in conclusion, it appears that the skin burns have been limited by additional training, but nerve injuries have not been eliminated by currently implemented mitigation methods. And while we believe that the risk of nerve injury will not be completely eliminated, are additional mitigations warranted?

The attachment to the discussion questions provides a listing of the mitigations implemented in the pivotal and continued access studies. In the discussions of Question 5, please also comment on whether or not additional mitigations are warranted to prevent unintended heating of the sacral and sciatic nerves. This completes the FDA presentation, and I turn it back to you, Dr. Noller.

DR. NOLLER: Thank you very much. Once again, thank you for staying within the time allowed. We have a few minutes now before lunch time, and during this short period of time, what I would like to do is to ask the panel if they have any questions that they would like to perhaps pose to either the FDA or the sponsor this afternoon. We won't discuss them now, but if we present the questions now, it will give them a chance to think them over, and develop some answers.

I'm going to take the Chair's prerogative and ask my questions first so I get them all in. First of all, I'm curious - and this is a question for the sponsor - I wonder if you have an explanation for the variability between the dose volume and the non-perfused tissue volume. You stated it was not blood-flow dependent. From the charts we saw that the FDA presented it wasn't consistent either. If it were always 1.5 times the volume or 2.3 times the volume, you could make calculation, but it varied all over the place. And if you could try to explain that to me.

Secondly, in the material we received, and this goes along with safety and education, I didn't see any teaching about conscious sedation. You might have thought that's not necessary because hospitals have rules, but this is an out-patient procedure that could be done in a free-standing place where that wouldn't be required, so are you going to include that?

And last, is there any sort of lockout feature that prevents providing pulses closer together than every 90 seconds as you now have them set up? If somebody is in a hurry, they have to get to their golf game, could they do pulses every 15 seconds and cause damage, or do you have a lockout feature. Yes, sir. Dr. D'Agostino.

DR. D'AGOSTINO: I have a few questions which may have been covered but I missed them. The first one is the symptom severity scale. We talked about validation. Is there a literature that I've not been able to find that talks about a change of 10 points as being some high level of clinically meaningful change?

My second question is that I'm struck by the control group, that I would have thought the adverse events that the hysterectomy group was going to be observing would be somewhat different than what this new treatment is. And you may have said it, but what was the actual logic? I heard a lot of negatives on why you had to get a control group, and you ended up picking hysterectomy, but I don't hear any positives on how you could really make these sort of safety comparisons, especially on the issues that are relevant to the ExAblate.

And then my third question is that the symptom severity scale in the baseline versus the screening seemed to have on the data that was presented, that you almost had like 20 percent or so of the subjects changing by a score greater than 10. And I'm concerned when you get to the year, you have about 40 subjects who are positive, meaning greater than 10; where if you just took repeated measures you may have had something like a change of 10 out of those 20 just by chance alone. And so I'm concerned about how I'm supposed to interpret the 12 month results given the variability in the scale, and also the fact that you only designed a six month follow-up, and how is the panel supposed to respond to that? Thank you.

DR. NOLLER: We'll go to Dr. Brown and then Dr. Crum, then Dr. Miller.

DR. BROWN: My questions are really all for the sponsor. I have a question about and a concern - we talked a lot about the risk to the sacral plexus and sacral nerves. And from my knowledge of anatomy and review of the materials, it seems to me that the colon and rectum are going to often be included, if not almost always included in the post focus point beam. And I wanted to get a little more detailed information about exposure of the colonic mucosa serosa to the post focus beam energy and the effects that may be have seen in the patients in terms of GI symptoms. And also, there was some discussion about air stopping the beam. Was there any consideration given to patients having an empty rectum at the time, and is that being looked at in the ongoing study?

Second question is what provisions were taken and will be recommended, particularly in the training, to ensure that the abnormalities were being looked at in MRI are actually fibroids and not some other entities such as a sarcoma or adenomyosis?

Third question is, who are the intended potential practitioners? Is this intended to be used only by radiologists? Is it intended to be used by practicing OB-GYN physicians. And again, I didn't really glean that from any of the materials.

Another question for me is one of my more important questions. What are the implications of the fact that only 11 percent of your treatment group in the pivotal study were African American women, only 1 percent Latino, 3 percent Asian, no Native American or Native Hawaiian women. What are the implications of this to the generalizability of your results in the population in the United States? As we all know, the group who would probably most benefit from this treatment and have the highest incidence of symptomatic fibroids are African American women, so could you comment on that? That's about it.

DR. NOLLER: Dr. Crum.

DR. CRUM: This is for the sponsor. In your panel package, a statement was made, "The ability to predict the ablated tissue volume as a result of a given sonication is the central factor upon which the entire treatment plan is based", so this issue of predicted thermal dose area versus non-profuse volume I think is the central issue there.

And it seems to me that the fault is in the thermal dose prediction in the model, and I would like to ask because I couldn't determine it from our package, in the thermal model do you consider temperature dependent attenuation, because the attenuation can go up by a factor of 50 to 75 percent during a temperature elevation, frequency dependent attenuation, because if you have a water path in front, you get non-linear effects, which means you get higher frequencies. And those are -- those attenuations are frequency dependent.

No perfusion I could see in the model. No long linear effects, as I mentioned earlier, and no cavitation. That's a difficult issue, but cavitation-related heating is, of course, in the recent literature a very important factor, so I'd like the sponsor to address that.

The second thing following on the point of cavitation is there is some statements, page 35 I think, that says that the threshold for cavitation is approximately 1,300 watts per centimeter squared, D-rated. The intensity would be on the order of 800, so that's significantly below the threshold for cavitation, but on the other hand, that data from Hynynen was based upon pulsed cavitation, pulsed acoustic protocol. And it also was done at a fixed temperature. The temperature at 85 degrees, the threshold for cavitation is significantly less, of course, than 1,300 watts per centimeter squared.

I know you have a cavitation detection system that was never mentioned, and I'd like to see how that works, and if it works. Thank you.

DR. NOLLER: Dr. Miller.

DR. MILLER: Yes. My questions are also directed to the sponsor. I'm interested to know in the primary pivotal study why a non-randomized design was chosen. And I again I may have missed it, but I don't see enough address of the differential in the study populations since it wasn't randomized. And the specific issues that I'm interested in are these populations differed, as we've already heard, by race. I don't find any report of the number of fibroids in the TAH group, or the volume assessments of those fibroids, and how they compared.

Also, in terms of the calculation of disability, there's a lot of analysis relative to the differential in calculated disability, but if you consider that 33 percent of the focused ultrasound group needed to be retreated, some of whom were then treated by hysterectomy and other modalities, there doesn't seem to be any aggregate calculated disability that would include complete treatment, particularly when you're looking at outcomes over a six and twelve month period.

In terms of the health-related quality of life scale, and this again gets back to the differential in the populations, if I read this right, there was a significant prevalence of underlying depression in the TAH group, which wasn't reflected in the focused ultrasound group. And there were some other differential characteristics, like anemia and hypertension, what medications were they taking for their hypertension? Again, these all speak to the fact that these populations were very different. And since you're basing your efficacy on this 10 point scale, what analysis can be deployed to understand the comparison?

And the final thing that I want to ask you about is what post hoc analysis was done to better understand the treatment failures in the focused ultrasound group? Clearly, it's a significant population. Obviously, I would think that you'd want this to be a modality that would be effective for the long term. Can we have any better understanding of what patient population is this better designed for, and maybe that would inform the exclusion criteria in the future. Thank you.

DR. NOLLER: We have about eight more minutes and five people. Dr. Hillard, you were first, then Dr. Brill and Dr. Solomon, Dr. Diamond.

DR. HILLARD: Questions for the sponsor, questions about the patient death. Was an autopsy performed on that patient? What were the findings, particularly the findings in the pelvis for this patient? And given that she did clearly have additional risk factors, are there any screening issues that could be recommended. If you had known she had Factor V Leiden, could or should this have been an exclusion for treatment?

In follow-up of the questions about who is the intended practitioner, if this is a radiologist, what recommendations would be given for communication between the gynecologist and the radiologist in terms of both follow-up, and also in terms of immediate potential for complications, the potential for acute bowel injury or intra pelvic hemorrhage, so these need to be addressed in the planned training and recommendations.

DR. NOLLER: Dr. Brill.

DR. BRILL: Yes, I have a number of questions. First query regarding inclusion or exclusion and the follow-up. For what reason FSH was not followed in the patients after therapy? And I wondered if there's any stratified data regarding the effect of this and oral contraceptives which appears to be acceptable in the protocol in the patients after treatment.

In regards to the myoma treatment itself, if I'm reading the materials correctly there were a number of myomas of 2.3 per patient, and a mean number of treated 1.3. So the question is what method was used to choose which fibroids were to be treated, number one. Number two, what was used to rule out the fact that they may be degenerated already, and why were non-perfusion volumes not applied before the institution of the protocol?

And third, Dr. Stewart, you mentioned that those amenable to hysteroscopic or laparascopic myomectomy would not necessarily have been treated, and I'd like to know more details about that statement regarding the pivotal trial.

And last, regarding the non-perfused volume - we heard a number of references to the point that in fact the NPV appears to under-estimate the histology. Well, if that's the case, if we take the statistics that were presented with a non-perfused volume of average of 68.7 Ccs and a percent of the myomas treated 23.6 percent, then how at six months do we have a 14 percent shrinkage, and thereafter, a 9.4 percent shrinkage from the intent-to-treat, if indeed the area was greater than treated versus less.

DR. NOLLER: Dr. Solomon.

DR. SOLOMON: In the material presented to us, the test arm inclusion criteria include MR accessible fibroids, but there isn't a discussion as to how many patients were rejected because of interrupted -- intestines in the pathway of the beam or calcifications in the fibroid, so that there are a number of patients that were obviously excluded, and we don't have a good sense of that in the materials.

Secondly, the beam pathway can be affected by interactions or interfaces between different tissues. And the beam then can be in a different -- the focus can then be in a different place from where the predicted focus would be, and that's part of the calibration procedure early in the program here. The question is how often is the sub-lethal dose different from the actual focus, and how far do you have to move it, because that may be something that comes up in other parts as you move the focus around, that you could be endangering other tissue.

And the third question is, we have in here the use of MR thermometry in order to mitigate the risk of unnecessary heating of critical structures, but it appears that in the case of the skin and the nerves that MR thermometry safeguard was unable to succeed, and so maybe further discussion there would be helpful. Thank you.

DR. NOLLER: Dr. Diamond.

DR. DIAMOND: I had a number of questions actually about the logistics of how the study was conducted. First of all, you said that the decision was made to use a 3-2 ratio for randomizing patients or assigning patients in the treated arm or in the control arm, but why was that -- what were the demographics that were utilized to come up with that power calculation, and the idea to use that ratio?

The evaluation of the perfused area, like Dr. Brill, I wondered was that done before the study. It's my understanding it was. What percent was not perfused at that point, or were they all totally perfused if they were going to be included in the study. Was the -- obviously, the practitioners, the radiologists were making that assumption in the decision at the time of the study, but was that the actual data that was utilized for the calculations that we see here, or was the ultimate data that we've seen here generated from a central review of perfused and non-perfused areas. If it was not, how was that standardized between different investigators. And did the individuals doing it, if it was central site, did they know whether it was initially procedure, immediately post procedure, or at later time points, because without a control group which has those sorts of measurements, I think there's a potential for bias knowing that it's potentially treated; and, therefore, subconsciously thinking maybe it should be either larger or smaller, whichever that could potentially go.

You've given us the change in fibroid volume and perfusion volume. You didn't give us the uterine volume changes though, and I'd be very interested to know that, typically since you're treating just 1.2 fibroids in these patients.

It's also very curious to me that with the small percent change that you saw in the fibroid volume, that you saw the benefit that you did see. That's one of the reasons why I want to have the total uterine volume changes, but do you have any idea why such a small change in what I'm assuming will be total uterine volume would have the beneficial effects that you did identify?

Other logistic issues, you mentioned the thermography is measuring temperature change, and have you done the standardized temperature of the patient or temperature of the room? Have any studies been done in other animals that have uteruses like humans, like primates, with thermistors to see whether the ultrasound treatments -- what sort of temperature changes are seen there, and how well that is picked up with all the modeling that's been shared with us, whether that correlates or does not.

It appears to me that, at least as I've read the documents, that some of the patients that were in initially treated ended up being found out that have adenomyomas and, therefore, were excluded from the analysis, although they had already been treated. Since that's likely to happen in clinical practice, as well, the practitioners were not able to differentiate initially, I'm not sure that's the right thing to do. Similarly, I'm not sure that it's necessarily appropriate to exclude those patients who are outside the window when they had their follow-ups, and would like to see data as to how that might affect uterine volume changes in those individuals.

The question came up is if you have multiple fibroids, which one was treated. I'd also like to know where within the fibroid was treated, and how that was decided, and whether that has any impact? Perhaps was it closer to the uterine surface, the middle, or what location.

And then the question --

DR. NOLLER: Last question.

DR. SOLOMON: Last question. Another way to assess the effect of the sonication might be functional MRI looking at blood flow changes, and has that been done?

DR. NOLLER: Thank you. One quick question.

MR. WEEKS: If I may, three really brief questions.

DR. NOLLER: Ten seconds each.

MR. WEEKS: First, the difference between the thermal dose area and the non-prefused area, have you looked at the relationship between that variance and BMI?

As far as your questionnaires for symptoms as a tool, two questions. Was the timing of the questionnaire, the taking of the information any way related to where the patient was in her menstrual cycle? And third, has there been any data on the particular tool with regard to placebo effect? So for example, has the same tool been used in medical studies where there's been a prospective randomized placebo control trial, and what is the magnitude of placebo effect? That's it.

DR. NOLLER: All right. Obviously, we don't expect you to respond to all of those. We'd be here for all week. And we will not be asking you to respond to all of them, but those are the sorts of questions we have that may come up in our discussions of our nine questions. But if you could prepare short one or two sentence answers to some of those, and some of them won't be asked, I'm sure, because the questions won't directly affect that.

We will meet back here at exactly 1:00. For the panel members, there's a place in the back of the restaurant for panel members only, and I have a message here I don't understand. Why don't we do those as we come back in our open panel discussion, doing that, because I think a lot of people have more questions. One o'clock, it's now 12:01.

(Whereupon, the proceedings in the above-entitled matter went off the record at 12:01:35 p.m. and went back on the record at 12:59:47 p.m.)

DR. NOLLER: During intermission we worked out a couple of things here. I know there are other panel members that had questions. We'll try to get to everything today. We certainly don't want to cut discussion short, but I think what we'll do for the next little bit, we have now panel discussion from 1 to 2:45. We're going to ask the sponsor to respond to the questions that they had in general categories. And they said they can do that in about 10 minutes. At that point, we will start our discussion of the nine questions. And the work we have to do this afternoon is to develop some answers to those nine questions, and develop an overall opinion concerning the approvability.

The open public hearing from 3 to 3:30, we know that there will be at least one person to speak then, so I ask the panel to watch the time and the length of comments, questions, et cetera, but we will go until we're done. We hope that go until we're done is 4:30. Is the sponsor prepared to respond?

DR. STEWART: Thank you, Dr. Noller and panel. What we tried to do is get some of the areas of maximum overlap in terms of questions. And while we'll be here to answer all questions. Our beepers and cell phones are locked away in the suitcases, so if we missed your question on the first round, we'll clearly go back to it. But it appeared to be a cluster of questions around the primary efficacy endpoint, and why the symptom severity score was an appropriate measure, was the 10 points the appropriate measure, was there too much variability inherent in this measure. And I think we chose this as a primary efficacy endpoint because symptomatology is really the primary complaint for women with fibroids. And that this is significantly impairing their lives.

We chose the only fibroid-specific validated measure. Again, it's a shame that the field is so far behind in measuring disease impact for women, and it wasn't until the last several years that there was, indeed, a validated study. And so the symptom severity score of the UFS-QOL is really the appropriate measure for this disease.

Several people raised the issue of why 10 points, and if we can go back to one of the slides that was in my presentation this morning, I know we had a lot of different things over the morning, but the 10 points was defined at the outset of the study for two very different reasons. First of all, we believed that it meant that there was clinically significant improvement, that if we can get the graph up eventually, if you'll recall, when they were validating this questionnaire, the women with fibroids had mean scores about 40, and women without fibroids in the normal population had mean scores in the 20s. So 20 points separated significantly effected individuals from non-effected individuals.

And just like with other treatments for fibroids, such as GNRH Agonist, if you get a 50 percent volume reduction or a 50 percent reduction in bleeding, this is generally clinically significant. So for that, reason we chose 10 points.

Here we go. Do we have the pointer again? So it's the two bars on the left, the symptomatology and the fibroid bars are in blue with a mean of 44. And the normal women were at 23. And, in fact, what we found from our data was not that 10 points separated our groups, but a mean of 23.8 points separated our group. So again, if we look at the differences that would bring the fibroid patients really down into the normal range.

So if we had set our criteria for success instead of at 10 points where we got 70 percent of patients to respond, and we had hypothesized that there would be 50 percent of patients, even if we move that up to 15, we had 70 -- we would have 50 percent of our patients having improvement. And we found, again, a 40 percent reduction in symptoms which again from our previous experience with drugs such as GNRH Agonist or Mifepristone generally does translate into symptomatic improvement.

We also chose this cut-off for several methodologic reasons, and if we can go back to the word slide in terms of statistical methodology that 10 points was very close to the standard deviation of the population very near to the standard there or the mean, and it correlated with a moderate effect size.

The other issue that had been raised by several individuals was that there was a variation between the screening that we obtained on this questionnaire at the screening visit, versus the baseline at the treatment visit. And we went back to look at those issues to assess what the differences were.

It turns out that the treatment day assessment of symptom severity and the follow-up ones are very consistent. It was really the screening day that showed variation. And in trying to understand that difference we looked at several things. We looked at difference to menstrual period and the menstrual cycle. Cyclicity didn't seem to make any difference.

What we found out was that there were some centers that were not administering it in the standard format. As you may know, quality of life questionnaires really do depend on you using the same instrument in the same way. And some of the recruitment centers, especially ones that had patients coming in from long distances would sometimes use fax copies or phone interviews to try to assess symptomatology.

So we then were able to look at that. We also found that there really wasn't any difference in which measure we looked at. Both the median and the mode of the differences were zero, and if we assessed values between screening and six months versus baseline and six months, we got the same difference. So I think although it is a concern, it doesn't affect the results, and this has proved to be a dynamic measure.

There was also question about whether this could represent a placebo effect. And although any self-reporting measure is vulnerable to a placebo effect, the first thing is that we do have clear documentation that we had an effect. We have the radiographic imaging. We see the differences in blood flow, and everyone did, indeed, have an MR image prior to treatment that showed that there was enhancement. That was one of the inclusion criteria.

We also know that size is not the only criteria for efficacy. The questions were raised about why we didn't get more size reduction, and I think the UAE experience has told us that size reduction doesn't necessarily correlate with symptom reduction. And there are many changes that may be going on in the consistency or the density, having a lead ball sitting on your bladder may be very different from having a cotton ball sitting on your bladder and the size reduction doesn't have to factor in.

Finally, although there may have been some placebo effect at three months, that at six months I think we were seeing a real effect. We got some patients who would maybe see symptomatology relief at three months, and by six months it was pretty clear to any of us that spoke to the patients that they clearly knew they were using less tampons or getting up less at night to urinate, or they weren't, and I think that goes in general timing with the known information about placebo effects.

And finally, we had multiple parameters that were consistent. We didn't just rely on the symptom severity score. We had SF-36 monitoring to prove that we were getting concordance in our study sample. We also had health-related quality of life and overall treatment effect, as well. Is that my 10 minutes?

DR. NOLLER: It is.

DR. STEWART: Okay. Is there any questions?

DR. NOLLER: Thank you.

DR. STEWART: Okay.

MR. NEWMAN: If there's any of the questions that were asked earlier that need to be covered before the deliberation, we'd be glad to cover those.

DR. NOLLER: We may ask you back up to the podium as we go along here, you and/or the FDA. Okay. That was 10 minutes. Is the company okay with that? That's what you asked for, that's what you got. Are you --

MR. NEWMAN: We said we believed in 10 minutes we could cover the main topics that had been covered by several people. Of course, that left many other questions unanswered, the physics questions and some of the other more specific things, and we'd like to cover those before we get to the deliberation and vote.

DR. NOLLER: Okay. That sounds good. What I'm going to do now for the panel is to read you three definitions. The definitions of safety, effectiveness, and valid scientific evidence. These are the measures that we are supposed to use in making our decisions today.

Safety, the definition reads: "There is a reasonable assurance that a device is safe when it can be determined based upon valid scientific evidence that the probable benefits to health from use of the device for its intended uses and conditions of use, when accompanied by adequate directions and warnings against unsafe use outweigh any probable risk."

Definition of effectiveness is: "There is reasonable assurance that a device is effective when it can be determined based upon valid scientific evidence that in a significant portion of the target population, the use of the device for its intended uses and conditions of use, when accompanied by adequate directions for use and warnings against unsafe use will provide clinically significant results."

And then finally, the definition for valid scientific evidence: "Valid scientific evidence is evidence from well-controlled investigations, partially controlled studies, studies and objective trials without matched controls, well-documented case histories conducted by qualified experts, and reports of significant human experience with a marketed device from which it can fairly and responsibly be concluded by qualified experts that there is reasonable assurance of the safety and effectiveness of the device under its conditions of use. Isolated case reports, random experience, reports lacking sufficient details to permit scientific evaluation, and unsubstantiated opinions are not regarded as valid scientific evidence to show safety or effectiveness."

MS. BROGDON: Dr. Noller, I'm sorry to interrupt.

DR. NOLLER: Yes, ma'am.

MS. BROGDON: I just have a procedural question. The panel laid out a number of questions for which you'd like answers from the firm. I'm just not clear on when you intend the firm to be able to answer those questions.

DR. NOLLER: I think those that were not answered will come up as we go through these questions, and we will give the sponsor time as we go through the questions.

MS. BROGDON: Okay. Thank you.

DR. NOLLER: And certainly before we do our voting. The first group of discussion questions are six, and they deal with safety and effectiveness. And the first is the primary effectiveness endpoint for the pivotal study is the symptom severity scale derived from the uterine fibroid symptom and health-related quality of life questionnaire. Success was defined as a 10 point improvement in the symptom severity scale of the UFS-QOL instrument in at least 50 percent of ExAblate patients at six months. Is the 10 point improvement at six months a clinically meaningful measure of success?

As we go through these, I will ask our primary reviewers, Dr. Diamond and Dr. Roberts, to start the discussion of each one of these points. Dr. Diamond.

DR. DIAMOND: At this point, I guess I remain unconvinced that a 10 point drop here was a clinically significant difference. It seems like this group of patients is a very select group of patients with uterine fibroids. We were told that they were not amenable to hysteroscopic treatment. They were not amenable to laparascopic treatment because those patients would have been treated in those fashions.

Furthermore, the average hematocrit of these patients I think was about 37 to start with, and so there are a group of patients who have fibroids but are not overly or imminently symptomatic, at least as the data has been presented to us. If you had included some patients who have greater symptoms, greater amount of bleeding, then the original score in the fibroid group might have been more than 40, and the differences going back to a normal patient population may have been 30 or 40, which would have, as the company has presented to us, would have influenced what they were seeing as a clinically significant difference because they were looking at standard deviation or standard error, or half the distance from where the treatment group was to the control patients in the validation study which described the use of this instrument.

And I guess the last component I'd want to make about that comment is, as I read the manuscript which was presented to us in our packet, that manuscript describes a difference between patients with fibroids and patients without fibroids, as to what you would expect to see on the scores, and that data was presented to us. But that manuscript does not talk about at all any sort of changes, or what is a clinically significant change for that instrument. So while it's been validated, differences for patients with fibroids as opposed to without, I'm not convinced that it's been validated for changes of symptoms for patients who are having treatments for fibroids.

DR. NOLLER: I'll say just as an side, I understand that at the open public hearing we will be getting some more information about the instrument that was used from a public member who -- from a person in the audience who will speak. Would you like to respond, sir?

MR. NEWMAN: Chairman Noller, can we respond to that question?

DR. NOLLER: Yes.

MR. NEWMAN: That was one of the questions, we didn't cover that in our 10 minutes.

DR. STEWART: I think that it is clear that the patients that we saw were significantly symptomatic, that if you look at again our symptom severity score, we were well in excess of what was defined as a mean level of symptoms for women in the validation study. We were looking at a mean of 40, and our patients in the MRI guided focused ultrasound group had a mean of 61, which was very similar to the mean of 69, I believe, in our hysterectomy group. So although they may not have had significant anemia, they did, indeed, have significant fibroid symptoms.

They also had a significant uterine volume, that the mean uterine volume was 600 cubic centimeters, and with the standard deviation there were patients in this population that had well over 1,000 cubic centimeters, so I would agree that they didn't overlap the patients for whom we would perform a hysteroscopic myomectomy or a laparascopic myomectomy, but they clearly were the patients who would currently undergo an abdominal myomectomy, a hysterectomy, or in many institutions a uterine artery embolization.

DR. NOLLER: Thank you. Dr. Roberts.

DR. ROBERTS: Well I had some of the same concerns because when I -- to some degree I think have been answered, but when I went through the paper looking at the quality of life measurements, and I went through other papers that talked about looking at quality of life, I couldn't find anywhere that 10 points meant anything, so you've somewhat answered my question in terms of whether or not that's a realistic goal. I don't think we really have a good answer for it.

I was impressed, though, when I looked at that, that in comparison to the group where there was in the validation study where the patients with uterine fibroids certainly had a -- the patients in this group had a much more of a severity index than those patients, so they obviously were quite symptomatic.

The one thing that I was little bit - a little bit off the subject of this - but one of the things I was a little bit confused about in terms of the study was, I didn't find anything anywhere that indicated what were the primary symptoms of these patients. In other words, were most of them coming in with bleeding, were most of them coming in with both symptoms? When you look at the anemia levels, unless they're doing pretty well on keeping up on their iron levels or something, and being able to keep up with their blood less, you would have to say that maybe most of them were bulk symptoms. But I think that's very important because when you look at some of the patients that then go on to have hysterectomies or drop out of the study, you get the impression that many of these are for bleeding problems. And maybe I missed it, but I just couldn't find it in there to indicate what it was that these patients were coming in with.

DR. NOLLER: Let's hold off just a second, get other comments, and then we'll ask you. Any other panel discussion on this number one? Yes, Dr. Hillard.

DR. HILLARD: Just in thinking about whether 10 points is clinically meaningful or not, one could just mathematically come up with a situation in a patient who is maximally symptomatic, so answering five at the far extreme of the scale for all of the symptoms, and could with only half of those symptoms drop down to having symptoms a great deal. And that would be a drop of over 10 points, so if you ask me if that's a clinically significant improvement, she still has symptoms a great deal of the time for many of her symptoms, so I think that mathematically I have -- in theory that would not qualify as a success. And so, therefore, I wonder about having some sort of an absolute value in addition to a magnitude of decline.

DR. NOLLER: Dr. D'Agostino.

DR. D'AGOSTINO: I was just going to say in terms of when I'm involved with these type of scales, the validation with the change, you'd like to see the group move to some other clinically relevant group. And they're moving to 40, which is sort of what the original comparison group was with the normal, so it's hard to figure out I think what the 10 means, and even where they moved into.

I do also think that if they got very extreme individuals, it tends to be on the scales, the ones that are extreme tend to change the most. And I don't remember the way the comment is, but I don't know what's driving the scale, and what the 60 to 40 actually means in terms of the clinical symptoms, and is that really clinically exciting? And I think we're missing that by just putting everything into a number like 10.

DR. NOLLER: Dr. Brown.

DR. BROWN: Another question that I hadn't asked was, which I didn't find, was what were the percentage of the patients that hysterectomies that had a 10 point change? Because to me, that would give me some perspective to gauge again, so we say that 100 percent of the patients that had hysterectomies had a 10 point change, that would mean one thing. If you said 30 percent of them had a 10 point change, that would mean another thing, so I thought that bit of information would be helpful, and I wasn't able to find that anywhere. I don't know if that could be addressed.

MS. MOONEY: If I understood the 10 points correctly, I think in looking at the validation or the comparison between fibroid patients and normal patients, the scales we looked at earlier, it was a 40 to 20 drop, so the 10 points there represented a 50 percent change. And I think, if I'm reading this correctly, from the six month data, it looks like baseline for the intent to treat patients was 61 as was just mentioned, and dropped to 34. So I think in terms of that 50 percent improvement for this patient cohort, it seems fairly similar to that 40 to 20. So I think the 10 points was not so much an absolute number. It was chosen, if I'm understanding correctly, but it represented a 50 percent change from fibroid patient to normal patient.

DR. NOLLER: Dr. Janik.

DR. JANIK: I think six months is very short. We need to really look out a little bit farther to see if there's a risk gain that's beneficial. To take any risk for a six month temporary improvement is similar to medical management, so we have to go beyond that, and also correlate with how many people go on to have hysterectomy. It's too short.

DR. NOLLER: Yes, Dr. Solomon.

DR. SOLOMON: One of the things I'm having trouble with on the scale is the meaning of all of this in absence of what I think the ideal control would be, which would be to separate a placebo effect. In this study, we wave all these fancy machines and we tell the patient things went great, and they go home, fill out a survey. And then you say well, how did you do, and they want to please. And we don't have a control here where maybe -- in any study I've seen, this is the best one because you can do a sham operation without having to put the patient through a big incision. There's no real downside than taking up some time. They can sit on the machine, the lights flicker and they say hey, you got the treatment, and then they get the survey later. I think something like that would give meaning to a 10 point change alone.

DR. NOLLER: Dr. D'Agostino.

DR. D'AGOSTINO: I was going to follow up on the 12-month follow-up. They ran into trouble because there weren't evidently designing a 12-month study, but the 12 months certainly makes the comparison or the numbers look a lot less exciting. They're down to less than their 50 percent changing over by a 10 point scale. And they also were in a situation where they have a lot of individuals moving to another procedure, and getting the hysterectomy, and I did ask that as one of the questions. And I really, if possible, to try to get a response from the sponsor on that, because I don't see how we -- personally how I move out of the dilemma that I'm very unimpressed by the 12 month data.

DR. NOLLER: I'm going to call on Dr. Miller, and then I'm going to ask the sponsor to respond to some of these items that have come up. Dr. Miller.

DR. MILLER: Well, interpreting the success or the meaning of the value of that 10 point drop, I mean I think we have to understand it in terms of the real attrition of this population. So if 12 months is a meaningful endpoint, at that point we only are dealing with 44 patients that we have data for. We've lost over 60 percent of our sample, and I think that makes the question of bias, and who stayed in, and whether people who actually felt better about their treatment; in other words, were maybe less symptomatic to begin with, were more satisfied to begin with, and were still in the study. So that 10 point as a reference point is material in terms of who stays in the study when you have an attrition rate that's that high.

DR. NOLLER: In fairness to the sponsor, too, let me remind people what we said earlier, that originally there was some misunderstanding, I guess, in the length of follow-up, so they didn't recruit these people to be followed for a year, so they had to go back and find some of them, and so there was some additional loss that normally we wouldn't expect.

Everyone who comes to the microphone from now on, even though you've spoken before, please give your name as we're recording this, so they have the name. So does the sponsor want to respond to some of these issues?

DR. STEWART: I'm Ebbie Stewart, and it sounds like there's still concern about the 10 point threshold. That clearly at the outset of the study, it was a hypothesis that this was an important difference. I think there are good methodologic and clinical reasons to suppose that this is an important endpoint, but I think, in fact, we saw substantially better improvement than that. We found 24 points on average, and many patients improved 30 or 40 points on this scale.

I think we also got objective endpoints in terms of by a reduction in terms of comparability to the treated endpoint, and outcomes on SF-36. I think also some of the concerns addressed the control group, and I would love to be able to tell you we could perform a randomized study. I think at the outset of this, there wasn't any way that we could feasibly perform a randomized study that we didn't have a minimally invasive treatment that had FDA approval. We had, ourselves, the experience of randomized, or trying to get people to trade-off between this minimally invasive procedure and an open procedure, and had seen what had happened with the attempts to do the same thing for uterine artery embolization. So we, therefore, chose the most comparable group we could find, and I think that the women in our study were indeed women who would have qualified for hysterectomy in any institution.

We weren't able to blind people to this treatment modality, and I think that we have established that we did get significant efficacy with these patients, that there are very significant clinical improvements that we have seen.

There's other investigators here that may be able to give their input since you've heard a lot from me. I'll introduce you to one of our other investigators.

DR. NOLLER: Please limit to about a minute.

DR. GOSTOUT: Okay. In one minute. I'm Bobbie Gostout, and I'm from Mayo Clinic, and I'm a consultant for InSightec, but I do operate under the guidelines of Mayo Foundation, and under the guidelines of IRB at Mayo Foundation. And I should state just to be clear that my travel and accommodations here were provided by InSightec.

A couple of things to just briefly wrap up. Some people are saying well, maybe if the patient's initial symptom scores were higher, we could say more about what this means, and I'd just like to point out that I believe we really are presenting to you a spectrum of patients with a range of symptoms of human fibroids that require treatment, which I think makes it, if anything, the best study that we could be presenting to you, rather than saying we only took patients that were at the maximum on every symptom possibility.

I'm hearing questions saying well, what is the clinical significance of this reduction on this symptom severity scale? I think it's important to consider a couple of slides that we presented before when we looked at patient satisfaction, and documented that at six months the patient's satisfaction in terms of saying that it was effective is, I believe, 72 percent. And at 12 months, we're looking at 79 percent. So the patients are telling us that the difference that they're measuring on this objective score means something to them. And in my mind, that validates the clinical significance of this. They're telling us that -- we asked them to put it in numbers and make it objective, but they're telling us saying I call this treatment effective. And, in fact, American College of OB-GYN's recommendations are that you treat uterine fibroids when they're a bother to the patient. And the patients are telling us that we effected the change that she came requesting.

If we go to a randomized clinical trial, I would tell you that at least one-third of the patients that I see that have entered into this trial, it would probably be an ethical question whether or not I could do a sham procedure for her, and in fact do nothing for her symptoms, because really, in fact, we are dealing with a number of patients that had the severe type of bleeding that makes me concerned about fainting episodes, that makes me concerned about them driving or even caring for their children when they're having their period, so we have a significant number of patients that were highly symptomatic, and I would be concerned about just randomizing them to no effective treatment.

DR. NOLLER: Thank you. I'm going to go on to the second question, because it also deals with effectiveness. The intent-to-treat success rate at six months was 70.9 percent as indicated in the table below, and you all have that table.

The ITT success rate at 12 months was 40.4 percent. The success rate dropped in part due to patient lost to follow-up between 6 and 12 months. By 12 months, approximately 20 percent of the ExAblate subjects had undergone alternative treatment for their fibroids. Secondary endpoints included fibroid volume changes at 6 months. On average, the treated fibroid volumes decreased by 16 percent. The question: Did the patient reported outcome data from the quality of life instrument at 6 and 12 months, when coupled with the clinical result of actual volume reduction of reduced fibroids support the effectiveness of the ExAblate for the treatment of uterine fibroids? Panel discussion. Yes, Dr. Diamond.

DR. DIAMOND: In order to put into perspective for me the 16 percent drop in fibroid volume, it would be very helpful for me to know some of the things I asked for before the break, which was when happened to total uterine volume before and after? So was the 16 percent consistent with what happened to total uterine volume going down, did it go in the opposite direction of total uterine volume? How were the readings done? Was it done by a blind reviewer centrally, or controlled for potential bias as to when it was being done; again, some of the logistics of how the study was actually conducted to get a better feeling for whether the 16 percent is -- how real that data is.

DR. NOLLER: Dr. Wood.

DR. WOOD: I'd like to make a point of clarification on the 16 percent, and it relates to thermal ablation therapies are often staged in cancer. And in this case, followed with volumes, and it points to the -- it's maybe not representing what you think it is. And if you think about a tumor, or in this case a fibroid, changing characteristic, becoming soft, that's not represented volume, and overall your thermal lesion, your effective devascularized coagulative necrosis area is also represented in this fibroid volume, as I understand it. So it's not necessarily a very pertinent measure.

DR. DIAMOND: Well, it's what we're presented with, and it's one of the markers that the sponsor has put forward as a marker of efficacy.

DR. WOOD: I understand that. I just want to clarify for the panel that this number is not necessarily indicative of -- you're not talking about a fibroid that is just shrinking 16 percent. The characteristics are shrinking and changing potentially, and they're not representative of that number.

DR. DIAMOND: That's true, but I don't know of any information that suggests that the characteristics of the fibroid, whether it spongy or whether it's hard, where that's been documented to show that that was associated with different symptoms by the patients. I think that's relevant.

DR. WOOD: Speaking from cancer therapies, we can treat a tumor, have it remain the same size. We can partly treat a tumor and get symptomatic relief that's long-lasting and not have he volumes change whatsoever in the measurement, so that's why the cyst criteria don't really apply to thermal ablative therapies in cancer, for example. I know this is off the subject, but the same sort of paradigm here may not fit. It's just a point of clarification.

DR. NOLLER: Other comments from the panel. Dr. Roberts.

DR. ROBERTS: Well, the only thing that I would say is that certainly -- my experience has with uterine artery embolization that even though you may not get a huge decrease in the size of the fibroid, the patients will tell you that there's something different about the fibroid, even if it's more or less the same size. But I do think it's -- I think that there is some confusion, certainly in my mind, and maybe the sponsor would be willing to look at this, or to give us some information about that; and that is, how does the change in the fibroid volume -- do you have any documentation about the uterine volume at the same time?

DR. NOLLER: Let me ask the sponsor here to answer that. Do you have data on the total uterine volume? And then the other question was were the volumes and results blinded as to whether it was treatment or not?

DR. TEMPANY: Yes. I'm happy to respond to that. I'm Clare Tempany. The uterine volume was measured at baseline, but it was never measured again after that in the follow-up examinations.

DR. DIAMOND: Don't you have that? I mean, you've got the images.

DR. TEMPANY: Yes, but I don't have that data to present to you now.

DR. NOLLER: Let's let them finish the --

DR. TEMPANY: We have the total fibroid volume, and we have those numbers that we've shown you already. And to go back to your other questions, Dr. Diamond, about the measurements, and whether they were done by single readers at single sites, or whether they were done at a core lab. They were, in fact, all done at the core lab with standardized interpretation and measurements ahead of time by a single person.

DR. BAILEY: And was that reviewer blinded to whether it was a pre, or a six month, or a twelve month evaluation?

DR. TEMPANY: No, I believe they knew which examination it was. They knew it was baseline, they knew it was six months. And then the other questions you had asked, just to clarify those to make sure that measurement of the perfusion areas, everybody had a totally perfused fibroid to enter study, so all of the non-perfused --

DR. NOLLER: Yes, you said that before. You said that in the 10 minutes.

DR. TEMPANY: Okay. Sorry. And just to concur with what Dr. Wood was saying also, I mean, I think a lot of us now believe in imaging. Certainly, that size and volume are very imperfect measures of any treatment effect. Certainly, in the cancer world, it's not being regarded any more as really being the most accurate measure of effective drugs, and this is where I think we're going to learn a lot more. And somebody asked about FMRI and perfusion imaging, and these are certainly neutrals that we're definitely going to apply. They're certainly not standardized today, and certainly not something we could have used in this trial. But to look at the consistency of the fibroid, and its perfusion, those are indices that I think we're going to learn an awful lot more about softening and reduction in pressure in the capsule which we think is probably occurring in this treatment.

MR. NEWMAN: I'd just like to add a little bit more to that. This is Rob Newman. Your question about was the reader blinded. The treatment effect is very obvious, and it would be -- you see a very large -- there's just an example from an image we showed you before. When you get a very large non-perfused volume in the middle of a well-perfused fibroid, it's going to be very easy to tell without any dates on the image to tell which is the pre-treatment and which is the post-treatment images.

DR. NOLLER: Other panel discussion regarding question 2.

DR. MILLER: It would have been very worthwhile to have the data stratified regarding the location of the fibroids. I mean, even if we acknowledge that size may or may not have a significance, I think there's reason to believe that location does affect at least the symptom severity scale, assuming that most of it represents menstrual bleeding aberration, and some of it represents pressure aberration; that indeed those myomata that were deeper set or impact the cavity, or in fact may be within a cavity in part, may or may not relate to the reduction of your scale. Do you have any information about that?

DR. STEWART: This is Elizabeth Stewart. We do have some information about location influencing symptoms. That was one of the things we looked at in our logistic regression model, and we found that fibroid location did not effect outcome.

With regards to the question about whether we primarily got improvement in bleeding symptoms or bulk symptoms, that the way that the symptom severity score is designed, it actually has questions that cover bulk. It isn't separately validated to be able to say we improved on bleeding versus we improved on bulk. However, we did look at the data because some of the questions clearly relate to more bleeding-related questions, and some related to bulk. And there didn't seem to be a difference. We got benefit in both.

DR. NOLLER: Yes, Dr. Roberts.

DR. ROBERTS: One of the questions that I had that was not clear to me from this was oftentimes it seems as if, if you have a, for example, a submucosal fibroid and you've got someone with bleeding, it's probably that submucosal fibroid that's causing it. Even if they've got other fibroids that are subserosal, they're not causing the bleeding. It's the submucosal one that is. My question is, if you had someone, let's say, that was bleeding, and had multiple fibroids, could you target the fibroid that you felt was causing the symptoms, whether it was bulk symptoms or whether it was bleeding symptoms?

DR. STEWART: Absolutely.

DR. ROBERTS: You could.

DR. STEWART: Absolutely, and in fact --

DR. ROBERTS: And you're saying that it didn't make any difference, you could treat the one that you would say was probably causing the symptoms, and it didn't make any difference, didn't help at all?

DR. STEWART: No. What I'm saying is that we saw patients who had benefits in bleeding, and we had patients who had benefit in bulk symptoms. But we did absolutely tailor the treatment to the patient symptomatology.

Now clearly in some patients where there's only a single fibroid, there's no decision making to be done. It's intuitively obvious if there's one fibroid and you treat it. However, the second level of assessment in terms of determining which fibroid to treat in women that have multiple fibroids, is are there any that are unsafe to treat. And so going back to the issues about is it too close to the spine, does it meet our treatment parameters? We have to assess that. But barring an exclusion for that, then we did have the patient symptomatology at baseline and could then choose to have -- for example, if the patient had bladder frequency, and she had one fibroid that was clearly in the lower uterine segment, and another that was up on the fundus and fairly subserosal, we could choose to treat the one nearest the bladder as a primary goal. And so we did tailor our treatment to the patients presenting symptomatology.

DR. NOLLER: Thank you, Dr. Stewart.

DR. ROBERTS: Can I just --

DR. NOLLER: Yes.

DR. ROBERTS: Just a follow-up. And were you successful in all of these patients in treating the fibroid that you felt was the most likely one to cause their symptoms, or were there patients where the fibroid that you would have wanted to treat, you couldn't treat because of its positioning?

DR. STEWART: I don't believe there were any cases where we couldn't. I can defer this to Clare. I think in almost all cases we were able to target the primary fibroid that we thought was symptomatically most important. And in some cases, we were able to target more than one in terms of the limitations that we sometimes went into treatment saying we'd like ideally to treat two fibroids. And in some of those cases, we could treat two, and in others we could only treat one. And I don't know whether you would like to amplify.

DR. TEMPANY: Yes. This is Clare Tempany again. No, I mean, absolutely. It's very rare that we weren't able to treat a fibroid that we had identified on baseline imaging. This one case that I remember that the bowel literally had fallen all the way down between the fibroids and the anterior abdominal wall so we didn't treat her, but very rare.

DR. NOLLER: Dr. Weeks.

DR. WEEKS: Jonathan Weeks. In your evaluable patients in the pivotal study, as I recall, you had 20 some patients that went on to receive other treatment, hysterectomy or other treatments. I think perhaps there was a link there that ties into what Dr. Diamond was getting at. If many of these patients had multiple fibroids, and over the course of six months or a year you may have effectively treated one or two of the fibroids, but one would expect that there would be growth in some of the other fibroids. I think it would interesting to look at those images and get volume information on the entire uterus for that reason.

And I also wondered if you got pathological information from those treatment failures so that you could look at the extent of thermal injury and the original fibroid sizes.

DR. STEWART: Ebbie Stewart again. We have obtained tissue in as many cases as we can from patients who underwent surgery, and we have at the minimum gotten operative notes and pathology reports on everybody that has gone on to have additional surgery.

We haven't been able to find any evidence that there was more extensive thermal damage than was recorded at the time of the treatment, and there was no case where there were found to be significant pelvic adhesions or injury that suggested that there was damage to the serosal surface. And clearly, we want to take those specimens and study them further from here on out.

I think also the fact that women went on to a different treatment, we have classified them as failures, and I think that's the right thing to do for them. But many of them actually did have some significant decrease in their symptoms from treatment, but then when they had a recurrence of their symptoms felt that they wanted to go on to more definitive treatment.

I think the constraints that we had in this protocol really maximized safety, to make certain that we didn't get to the serosal surface, the endometrial surface. We didn't ablate large volumes of tissue, but I think that did limit our efficacy. And as we move forward, I think to get maintenance and longstanding relief, we may either need to be more aggressive in our treatment, or to choose different treatment candidates. I think that the efficacy that we saw was very impressive given the constraints that we worked under for the protocol.

DR. NOLLER: Thank you. I want to remind the panel that this isn't a question and answer between the sponsor and us. We are discussing among ourselves. There just happen to be 60 people out there that are listening in on our private discussion here, so let's try to discuss and help each other out with answers. For example, as we had going there, and then we'll turn to the sponsor if there are things we really don't know that we need to know, but let's try to discuss some of these among ourselves.

Were there other things about Question 2 that you want to discuss before we move on to the next one?

DR. BRILL: Well, I would just like to add that just as Dr. Stewart has said, that our ability to evaluate these patients is relatively primitive because of our lack of instruments. I think we're even more primitive in our ability to explain why women do or do not bleed with uterine fibroids, and even beyond that, it's incredibly presumptuous to assume that a patient with abnormal bleeding and fibroids necessarily is bleeding from those fibroids, and doesn't have a coincident coagulopathy, and doesn't have some sort of coincident dysfunction at the endometrial level which we haven't been able to identify.

With that said, what I'm hearing in part is when you have a technique that is a selective myomectomy technique, and unto that I think there's some presumption, because clearly I do not which fibroid to selectively destroy having multiple fibroids. We do know from the uterine artery embolization data that the response in the context of diminished abnormal uterine bleeding is not necessarily related to the location of the myoma whatsoever, so I think that we're all sort of treading water here, and walking on thin ice when it comes to really knowing what we are or aren't doing, and deciding in a defining way which fibroid we're going to treat.

DR. NOLLER: Yes.

DR. ASCHER: I have, maybe it's a naive question and I apologize. Does not knowing the ultimate durability really change whether something is effective even in the short term? And maybe someone in the panel or elsewhere can address it. I'm trying to get a handle on effectiveness.

DR. NOLLER: Things can be effective for an hour, a day, a week, a month, a year, for life.

DR. ASCHER: Okay. That will certainly impact on some of the discussion.

DR. NOLLER: I don't think we are going to consider approving something that works for an hour or a day, but 10 days - where does it end?

DR. ASCHER: Something like GnRH analog as a temporizing method. I mean, there is some precedent for temporizing.

DR. NOLLER: Particularly in the peri-menopausal. Did you have a comment, Dr. Wood?

DR. WOOD: I was just going to reiterate what I said before, that I think the imaging out parameters as a surrogate marker of efficacy are unreliable, and we should rely on them more for safety issues here. And the bottom line is we don't have a perfect questionnaire mechanism, quality of life mechanism, but they used the best one that's out there.

DR. NOLLER: Other comments before we move on?

DR. BROWN: So my answer to 2 would be that the volume does not seem to contribute at all, because it doesn't seem -- to me, the volume changes were not impressive. And secondly, we don't know that that means anything, but there were consistent changes in the questionnaire that although the amount of effect was not as great at 12 months, there still was some persistent effect.

DR. WOOD: As an aside, the time course and volume changes are similar to palliative thermal ablation with needle-based techniques for soft tissue tumors and cancer, and the periation can happen. Again, maybe if -- I don't know the mechanism, nobody does, maybe decreased interstitial pressures within the tumor can have an effect without actually getting rid of the whole cancer, just as an aside again.

DR. NOLLER: Should we move on to the third question? Here I think we will need the sponsor to respond to some of the physics questions we asked before. The third question is, has the sponsor demonstrated the MR thermal mapping provides adequate intraoperative feedback during the treatment regimen, sufficient to ensure safe and reliable dosing to the intended fibroid tissue? Drs. Diamond and Roberts, any comments about that?

DR. MILLER: Dr. Noller, where are you reading these questions?

DR. NOLLER: It's in your packet. The left-hand pocket. You may have received a previous version. It's a couple of weeks old. It looks like this. Okay. Yes, Dr. Crum.

DR. CRUM: This follows up some of my earlier statements, and I'd sort of like to make a comment because we're sort of discussing this amongst ourselves rather than asking the sponsor and so forth.

DR. NOLLER: But we will ask them.

DR. CRUM: So it seems to me that the monitor, MR temperature monitor assumes that you get a temperature elevation. It's not an absolute measure, it's temperature elevation. And it also makes the assumption that after the treatment, we start in a sense de novo again at exactly the same temperature that you started last time; that is, at 37. And when you destroy the vascularity of that region, you also shut down the perfusion. And so the perfusion is not a major factor in carrying the heat away, but nonetheless, if it's adjacent to a treated region, the heat has to go somewhere. So to assume that you're going to come back to the initial conditions ab initio in each case, I think is a questionable assumption.

The temperature increase that does occur in this model assumes some very simplistic conditions, no non-linearity, no temperature dependent attenuation, so forth and so on, so it's not surprising to me that the prediction of the thermal dose differs from the non-perfused volume. This is complicated stuff. I mean, doing modeling inside biological tissue and these sorts of conditions are difficult to do.

But the endpoint of this is that after you do this treatment, you can go back in and get a non-perfused volume that's correlated pretty closely; that is a factor of like 2 or some number different from the computed thermal dose. So even though there might not be a one-to-one prediction, there is a correlation. And similarly, empirically one can get the non-perfused volume, which I think is certainly what induces the biological effect and the successful treatment, so I think that there are some inefficiencies and inaccuracies perhaps in the thermal model, but to answer the question, does this provide a reliable doing, with some modification, yes. I think that that's true, and I would invite you to comment on it.

DR. NOLLER: Dr. Diamond.

DR. DIAMOND: I have a question. Do you understand from the way the machine works whether a second sonication would be adjacent to the first one, or would it be some place far distant, because if it's next to it, you're going to have a potentially greater thermal effect than if it's at a big distance on the other side of the fibroid.

DR. CRUM: I don't know how they do that. I know that in other situations involving HIFU applications, focused ultrasound, people tend to go here, then there, and then back and forth, rather than do the adjacent one, so it's a question now for the sponsor to answer; do you do adjacent ones, or do you do --

DR. NOLLER: Let's hold up just -- did you have something to say, Dr. Wood, and then we'll go to the sponsor.

DR. WOOD: I was just going to say there's a spiral mechanism out there, as well, which takes advantage of that very effect and changes the modeling completely, and the effect. And one more question - maybe you guys could answer while we're here - adjusting the calibration in the beginning offset the test zone. That test zone, is it just done once? And if not, do we have any information on how tissue tissue-to-tissue interfaces in interactions between your transducer and the fibroid can affect future inhomogeneities in your treatment zone? And would it be more effective to have more than one test zone done?

DR. NOLLER: Would you address those questions, please?

DR. TEMPANY: Yes, I'd be happy to. This is Clare Tempany again. To answer your question about how we do the sonications, you can do it either way. You can either go one, two, three, four, five and across. More commonly, we will go from one part to the other. There is a time penalty with that because you're moving the transducer. But certainly, we receive the return to temperature baseline on the graph, so you'll see it return to baseline temperature. If it doesn't return to baseline temperature in the 90 seconds, clearly the easy thing to do is to move to a completely different location and go to the opposite side of the fibroid, and then take that even two or three sonications later before you get back there.

And yes, we're continually verifying the location of the sonication because you have the coordinates. So even if there is beam attenuation, you will see whether it's off, and everything is registered both in short axis and long axis. We can choose to do it in sagittal or coronal plane. Does that answer?

DR. NOLLER: Thank you. Other discussion on this topic?

DR. BRILL: I'd like to just to ask the technical expertise of the panel, is there any concern that there's monitoring of the differential between core temperature and realized temperature versus absolute instantaneous temperature monitoring as it was presented this morning? I believe it was stated that it was a differential between core temperature, was it not?

DR. NOLLER: Dr. Crum.

DR. CRUM: I think the question was appropriately answered. If you have a monitor that it drops back down to the background, they want it to drop back to the initial condition. And I presume that the initial condition was 37, so if it drops back to that initial condition and they do not fire the second shot until you've dropped back to the initial condition, that you're back ab initio again, you're back where you started. So I don't see that the core temperature is going to come up very much, maybe 1 degree or so, but I don't see that being a problem myself.

DR. DIAMOND: I'd like to ask also again, the FDA presentation, they had two physics presentations, each of which seemed to be suggesting that there are lots of potential where measurements and assumptions could lead to big temperature elevations greater than might have been expected and leading to tissue injury, but yet both ended up concluding that the likelihood of injury was actually rare, and having given us that presentation. What is your interpretation of that from what was presented? Is that going to be an exceedingly rare occurrence, or are the scenarios that we're describing as potential concerns, are they things that are perhaps more frequent?

DR. CRUM: I mean, I have lots of concerns, but for example, there was no treatment of what happens if there is bowel gas. I didn't hear a discussion of that. I'm sure they have thought about that, but if you have bowel gas, then you have other effects that are non-thermal, cavitation. And the literature is full of descriptions if you have bowel gas and you have either Lithotriptor or other kinds of even diagnostic ultrasound pulses, you can get damage to the tissue in the bowel from the result that you're trying to drive a pressure release interface. That was not discussed. Maybe the sponsors could discuss that.

The other thing is that non-linear effects and changes in the various tissue layers, as Dr. Wood was mentioning, will cause focusing in different regions that you would anticipate. And, of course, they argue that they do the sublethal measurement each time, and then they correct. If they do that every time, then that corrects for that error. And so I think that the FDA's analysis is that there is some areas that should be watched, and I presume that those directions will be given to the sponsor, and they will address those.

MR. WEEKS: This is a question, and I have no idea what the answer is; but the coupling gel itself, is there any reason to believe that there can be a variation in say that quality assurance and the density of that material, and as it's mass produced or whatever, can there be enough variation that that changes the thermal injury or damage in any way?

DR. CRUM: What happens with that is that -- and I think that's been addressed also, is if it's not adequately coupled to the patient, and if you move around and you introduce a bubble or a void of some sort, then you have a potential problem because then you're going to get a skin burn or something.

MR. WEEKS: I'm concerned with like the density or say the density of the material itself, so assuming it's got good skin contact --

DR. CRUM: It's typically acoustically transparent, so there's very little attenuation in that material, and so there should be no effect on that material from the ultrasound because it's essentially acoustically transparent. When we have an interface with gas, then you can have a problem, but not if you have adequate coupling.

DR. NOLLER: Dr. Samulski, did you have --

DR. SAMULSKI: I agree with this off-the shelf kind of interface that you use when you use ultrasound in a clinic for diagnostics, and this application as well. I think the sponsors must have tons of data testing animals and stuff like that, and could give you an idea of what the outcome was.

DR. NOLLER: Why don't we stop at this point and see if the sponsors have any comments regarding this discussion that's been going on for a few minutes.

DR. VORTMAN: Kobi Vortman, employee from InSightec, paid by InSightec. I'll try to address first the cavitation. Cavitation was a major issue in the design of the system. We had a real time integrated detector that continuously during the sonication is receiving the signal and looking for any clue for cavitation, which right now is between half of the frequency and then several minutes. In our training or in our instructions for use, the spectrum appears. You can see it here.

On the left side you see a different spectrum without the cavitation. Okay. And on the right side here, you could see what happens when the detector detects cavitation. You'll see a wide spectrum white noise, and the user is instructed to immediately switch off the system in this case, and move to some other place.

And as Dr. Crum said before, cavitation pressure could change as a result of temperature and so on, so this is a major tool. I would add to this that in addition to the cavitation at that point, we have a reflection detection continuously. So if the signal received by the system is detecting any air bubble or air surface, the system will detect it, and again the user is instructed to switch off. Okay. So that's, I believe, in response to -- here you could see the reflection detector. And what we do, in addition to measuring reflection, we are measuring it on the scale, AP coordinate, so you will be able to allocate it to the area which generated the reflection. So if it be internally, the bandwidth would be 76 or 102. At the interface between the transducer and the skin line, it would be 50, so you would be able to allocate the reflection to whatever place.

DR. CRUM: May I just ask you if that's a passive cavitation detector or an active cavitation detector?

DR. VORTMAN: It's a passive.

DR. CRUM: Passive, and what's the bandwidth?

DR. VORTMAN: The bandwidth is between 20 kilohertz and about three and a half megahertz, while we're working at 1.1.

DR. CRUM: So you're just looking at the subharmonic, are you not?

DR. VORTMAN: Correct.

DR. CRUM: And you realize the subharmonic means stable cavitation, which means that if you had this kind of intensity, you would get inertial cavitation rather than stable cavitation.

DR. VORTMAN: Of course.

DR. CRUM: Why wouldn't you be looking in a bandwidth up around 10 megahertz or something like that?

DR. VORTMAN: Okay. What we have said before, we are using continuously thermometry of the whole field of view, so if you will have any non-linear effect, like CW cavitation, we'll be able to see non-linear effects through heating, and detect it immediately, so we are using both. We are using both passive detector and thermometry to monitor those non-linear second and certain harmonic effects.

DR. CRUM: Do you ever think that maybe the reason the thermal model is not working is that you're getting submicroscopic cavitation, and that's coupling through the bubbles through increased heating?

DR. VORTMAN: This is an issue that we addressed extensively. The predictor in our system, both in our accumulated experience, is over-predicting the dose. Okay. So we didn't see effects like this at the levels that we are working. However, it could have been, and as I've said, it will have something like this. This should translate to either moving down, shifting down the focus immediately, if the focus were generating, or at some point in the way overheating. So we are training our user to watch for this.

DR. CRUM: So when you did histology, did you ever see cases where you had non-cigar-shaped lesion in some areas that were not treated because of the non-symmetric shape of the lesion indicative of some kind of inhomogenetic effect?

DR. VORTMAN: WE've seen it in animal work. Since we've driven the system into a very high intensity and we have seen in this case.

DR. CRUM: What about the hysterectomies?

DR. VORTMAN: Didn't see it there. We've seen -- scoped minimum number of cavitations that were switched off.

DR. NOLLER: While you're there, let me just ask you one question. You said that the operator is instructed to turn off the machine, but there's no automatic shutdown, or is there?

DR. VORTMAN: Correct.

DR. NOLLER: Thank you. Okay. Yes. Another point?

MR. NEWMAN: If I could just finish answering the question too about the acoustic gel. The acoustic gel, we buy it from Parker Laboratories. It's a general manufacturer of it, and it's a one-time use. Parker pours it for us in a specific shape. It has an expiration date. We use it for a single patient treatment, and then it's discarded. So we've done the biocompatibility testing and all that kind of stuff. Parker has done that already. And if there was an issue with a problem with the density or the acoustic properties of it, you'd see it through the reflection monitoring when you did the testing of the treatment before the patient is put on the table. We set the system up and we put a phantom patient, if you will, on top of a tissue specific phantom on the acoustic coupling gel, and do a quick in-room check before the patient is put on the table, that we would detect a problem like you had suggested.

DR. NOLLER: Thank you. Let's move on to question 4. A number of adverse effects specific to ultrasound treatment occurred during the clinical trial, including nerve injury, leg pain, and skin injury. Question: Do the results from the thermal modeling and our understanding of the underlying physics provide sufficient information to understand the etiology of the injuries that occurred in the study?

I'll respond. I think the presentations did a very good job of explaining how the people got the injuries, I thought. It's a different question whether or not there are adequate ways to prevent them, but I thought the explanation made sense. Any comments? Dr. Diamond.

DR. DIAMOND: I thought for the skin injuries, that was well-explained, and I don't have any questions. For the nerve injuries, some data was presented to us about incidence angle and estimated distance form the focus to the sacrum. Other things I guess I would have liked to have seen what was what known about the number of pulses, sonication pulses that went to the area where the nerves were in those patients. What happened to the temperature thresholds, and we're shown example of how if it's a little bit too high, the next sonication they lowered the energy. What do we know about the patient actually got injured? Did they have higher energy that was transmitted?

DR. NOLLER: It would be particularly interesting in that one page they had months to resolve that.

DR. DIAMOND: Yes, the 9019, yes.

DR. NOLLER: Other questions? Yes.

DR. CRUM: I'd like to ask either Dr. Herman or the InSightec people if you've examined the side lobes, because when you show these pictures, you assume a nice conical shape input for the focus wave, a nice conical on the outside. What happens, of course, with acoustic propagation is you get what we call side lobes, which means that it doesn't follow that nice conical thing. So if you had side lobes, you would get hot spots outside that area. If you had a hot spot on the nerve in which you didn't think you were illuminating you could get some damage. So I don't know whether Dr. Herman has modeled that or whether InSightec people have modeled that, but I'd just like to ask that question.

DR. NOLLER: Do you have a response? Please.

DR. VORTMAN: Again, Kobi Vortman. Yes, we have extensively modeled side lobes. The design of the transducer is such that normally you don't have any side lobes. The distance between elements and the number of elements is such that you don't have any deviation from a closely spherical surface. In addition to this, we have in system tissue aberration correction that is used -- I'm now probably getting into too much detail but I'll say the following. We are using the MR image in segmenting between muscle and fat, and this could be downloaded to the bin former to correct for tissue aberrations, speed of sound and so on, so the bin former takes this in and refocuses the focus.

The third point is you have real-time feedback on the focal shape in any second hot spots from the thermal image. So that is the reason we've addressed it.

DR. NOLLER: What about the answer to Dr. Diamond's question? I'd like both the sponsor and FDA to respond to that, if possible, about particularly looking at those patients that had injuries, if you learned anything from those.

DR. ROBERTS: Could I also just ask maybe a part of that, is that you've set up a 4 centimeter distance. My question is did any of the patients who have nerve injury, in fact, be within what you would say now is within 4 centimeters, or were there ones where everyone outside of that 4 centimeter area and still had a problem? So I guess my question is, do we really know that 4 centimeters really means anything, or is that a theoretical construct?

DR. VORTMAN: Again, Kobi Vortman. What is done, we've looked at energy intensity. The factor that generates the heating is energy intensive heat, and it's a multi-parametric problem. It's not only what is the energy intensity, it is what are the incidence angle of the beam impinging on the bone? If this will be above 38 degrees, you will have no heating at all. So we assumed that the combination of 4 centimeters that is keeping the energy intensity at a reasonable level, and non-perpendicular bin would build this into a safe environment, or envelope.

Have we seen cases in which even more than 4 centimeters and we had some leg pain? The answer is yes. And many of those cases that I've looked at, the bin was perpendicular, so you need to use both. One wouldn't do the job.

DR. NOLLER: Did FDA have any comments about that?

MR. HERMAN: I think we would agree with that.

DR. NOLLER: Give your name, please.

MR. HERMAN: My name is Bruce Herman. We would agree with that assessment with the caveat that while searching for non-normality and the 4 centimeters are definitely steps that would markedly mitigate the possibility of adverse effects to the sacral nerve or something, it wouldn't -- considering what I think I know about the possible variation of physical structure, wouldn't absolutely rule out the possibility. But again, go a long way to minimize the possibility of those effects.

DR. NOLLER: That helps us too with the next question.

MR. HERMAN: It's not an absolute demarcation as you mentioned, below 4 centimeters you can get damage even with normality and beyond. You can't. But putting both together gives, I think, a broad range of physiology and tissue characteristics. Looking at it with a fairly conservative eye, a reasonably good possibility of having only very rare occurrences of, let's say, sacral nerve damage.

DR. NOLLER: Thank you.

DR. DIAMOND: Dr. Noller.

DR. NOLLER: Yes.

DR. DIAMOND: I find both those answers actually sort of unsettling, because the data that was presented to us by Dr. Del Mundo, three out of the five patients who actually had nerve injury had incidence angles of 30 degrees or more, and distances of 4 centimeters or greater. I'm uncomfortable, which is why I was asking are there other parameters, such as thermography or other ways that you've looked at those patients where you could shed more light on that information to us, and give us a greater degree of comfort for the future above and beyond just these two criteria.

DR. STEWART: Ebbie Stewart again. I think the clinical input is also important here. This is where the patient having conscious sedation and being able to respond to discomfort, that the patients who do have heating of their nerve typically do feel sacral pain, buttock pain, and are able to terminate the sonication. In that case, there can be reassessment of the treatment plan, and moving of sonication points. I think the index case where there was a significant nerve injury, first of all, caused us to focus in more carefully on what was going on in the far field. And I think a lot of our mitigation steps have brought that to bear. But I think the other thing it's made us very cognizant of is the fact that interacting with the patient and responding to discomfort that she presents is important, as well, because oftentimes when the patient is having right buttock pain and you're sonicating near the right serosal surface of the fibroid, that's an indication that you are potentially getting some nerve heating, and that you should move before anything more happens. That I think nerve injury certainly can take place at the time of hysterectomy, as well. There are well reported cases where the retractor or the positioning caused this to bear. But we do have feedback, and now we have over 600 cases where we haven't seen a significant nerve energy.