Microbiology Devices Panel Meeting Summary for December 8, 2000
On December 8, 2000, the Microbiology Devices Panel met to discuss the appropriate types of information necessary to determine the effectiveness of in vitro diagnostic devices that detect human papilloma virus (HPV) DNA when these devices will be used in conjunction with Pap smear in women 30 years or older to increase the effectiveness of Pap smear screening for cervical cancer. Additionally, the panel discussed and made recommendations on issues concerning the use of HPV devices without Pap smear to determine a woman's risk of cervical cancer. This was a General Issues meeting and not for a specific device application.
Two manufacturers, Digene Corporation and Impact Diagnostics, provided background information on several clinical trials that have been conducted to evaluate the performance of HPV testing for detecting cervical cancer and cervical dysplasia, (the majority outside the U.S.). There were nine presentations during the Open Public Hearing session, from the American Social Health Association, the Society of Women’s Health Research, the American Medical Women’s Association, and the Association of Reproductive Health Professionals. All reported receiving grants from Digene Corp. Additionally, clinical investigators associated with Digene, from Kaiser Permanente - Sacramento, CA, Cleveland Clinic Foundation, Columbia University, and Dedham Medical Associates, Wellesly, MA provided additional information on clinical trials that they conducted. A letter was received from the Medical Director at the University of Nebraska Medical Center, who consults for Cytec Corp., and portions read into the record.
Penny Hitchcock, Chief, Sexually Transmitted Diseases, NAID, NIH, provided relevant information on the progression of HPV infection and cervical disease, along with questions raised by currently published research. The FDA followed with a review of current published clinical trial data and discussed problems with interpreting current data in literature, and problems with using simplistic statistical models. The FDA statisticians also suggested good study design principles and models that could be used to analyze literature studies.
The panel discussed the safety and effectiveness of using HPV devices as a screen for cervical cancer. It was suggested that before HPV testing can be used with Pap smears for routine screening for cervical cancer, FDA should answer questions such as the following. Would HPV testing allow cancer precursors to be treated at an earlier stage? Could HPV testing detect adenocarcinoma as well as squamous cell carcinoma? Can it be used as a Post Treatment test of cure? The panel agreed that in order to evaluate the safety and effectiveness of the device the FDA should consider whether the test would allow for minimally invasive treatment at an earlier stage, minimize patient discomfort and anxiety, and provide an adequate and accurate specimen for analysis. The device effectiveness could be measured by reduction in rate of cancer and reduction in morbidity or a reduction in CIN III with time. The panel also cautioned that the HPV test was inadequate for screening for adenocarcinoma and squamous cell carcinoma so no reference should be made to HPV infection as a precursor for all cancer types. The panel agreed that a negative HPV test at any timepoint could not exclude HPV infection or cervical cancer in the future.
The Panel provided recommendations on eight questions presented by the FDA. As to the appropriateness of the indications for use of the HPV device, the panel suggested that the available data did not demonstrate an appropriate sensitivity to be used as a general screen. They recommended that the strength of HPV testing might be in the negative predictive value, when interpreting results in conjunction with the Pap test. They indicated that if the HPV test was used in combination with Pap tests to decrease cervical cancer mortality, then there was use for the test as a surrogate endpoint. It was suggested that studies should include patients with CIN II/III, patients with representative demographics, should be multi-center, and should include a minimum of three thousand patients. The data should be accurate and should be designed to minimize bias. In response to FDA’s question regarding appropriate endpoints, the Panel suggested that, results of Pap test readings, colposcopy and biopsy be used as appropriate study endpoints. The Panel agreed that use of foreign data might be problematic if screening practices differed. For example, if the screening intervals vary, the performance of the HPV test in an unscreened population was likely to be non-equivalent to performance in a highly screened population. Other populations may be used as a model. Because there were no actual data to evaluate, the panel could not provide recommendations on how to adjust assay cut-off to maximize sensitivity without compromising specificity. However, the panel statistician recommended that the data was evaluable, and that the data should be analyzed using ROC curves. The panel recommended that published studies could be used as a guide for future studies. Because each study was designed to answer certain questions, it may not be appropriate to infer results from one study to another. The raw data from these studies could be analyzed using a Bayesian Meta analysis model if proper co-variants were identified. During the second Open Public Hearing session, Mark Schiffman from the NIH, a clinical investigator of HPV studies, commented that there were some ongoing U.S. studies that have significant data on the performance of HPV testing and asked the FDA to keep an open mind. Dr. Gutman asked the panel if they took into consideration the least burdensome policy, would they accept published data as supporting evidence, particularly if this data allowed access to raw data results. The panel did not discount the possibility of using published data, but did indicate caveats in the interpretation and use of this data.
Consideration of self-collection and alternative specimen sources were issues deferred for future discussion.
Contact: Freddie Poole, Executive Secretary, (301) 594-2096, x111
Transcripts may be purchased (written requests only) from:
507 C St. NE
Washington, D.C. 20002
(202) 546-6666/(202) 546-1502 (fax)
Food & Drug Administration
Freedom of Information Staff
5600 Fishers Lane, HFI-35
Rockville, MD 20852|
(301) 827-6500 (voice)/(301) 443-1726 (fax)
Information on panel meetings can be found at "Microbiology Devices Panel".