Clinical Review Template [032201]

The efficacy review is based primarily on two multicenter trials of RSR13 entitled:

(1) RT-009: A Phase 3, randomized, Open-Label, Comparative Study of Standard Whole Brain Radiation Therapy with Supplemental Oxygen, With or Without RSR13, in Patients With Brain Metastases

(2) RT-008: A Phase 2 Study To Evaluate the Efficacy and Safety of RSR13 Administered to Patients Receiving Standard Cranial Radiation Therapy for Brain Metastases

Below, the protocols for each of these clinical trials is reviewed independently.

RT-009:

A PHASE 3, RANDOMIZED, OPEN-LABEL, COMPARATIVE STUDY OF STANDARD WHOLE BRAIN RADIATION THERAPY WITH SUPPLEMENTAL OXYGEN, WITH OR WITHOUT RSR13, IN PATIENTS WITH BRAIN METASTASES

PROTOCOL REVIEW

Table 2. Protocol Milestones (Derived from Sponsor’s Table 9.15, Final Study Report)

Milestone	Date	Comments
First patient enrolled	2/16/2000	N/A
Amendment #1	3/2/2000	Stated MRI preferred over CT. PET added as an option for staging. Dosing adjustment Guideline was changed to include the instruction “if SpO₂ while breathing room air on any RT day < 90%, RSR13 was to be omitted.” Physician judgment could be used in determining clinical significance of an AE with respect to omitting or modifying the RSR13 dose.
Amendment #2	6/05/01	Sample size increased to 538 patients. Enrollment completion extended by 6 months. In addition to small cell lung cancer, extrapulmonary small cell carcinomas excluded from enrollment. Calcium channel blockers were added to the list of medicines that could potentiate or possibly interact with RSR13. Expanded warnings about use of concomitant CCBs and ACE inhibitors. A suggestion was added to start RSR13 dosing at 75 mg/kg in patients taking these classes of antihypertensive medications. An additional recommendation for patients who had a previous nephrectomy to start dosing a 75 mg/kg, to advise patients to avoid smoking during the RSR13 resaturation period. The Dosing Adjustment Guideline was expanded to include weight and gender. The scale for evaluation of hypoxemia AEs was initiated. Analysis of the NSCLC/breast population was incorporated.
Amendment #3	10/09/01	Included option to treat brain metastases with Cobalt 60. Clarified the conditions under which concurrent RT could be given to extracranial sites.
Date of Primary Analysis (Data Cutoff Date)	1/31/03	N/A
NDA submitted completed	12/4/03	N/A

Reviewer comments: The Sponsor stated that it was necessary to enroll 501 patients and observe 402 deaths to claim statistical significance in median survival time and rule out the null hypothesis. Total enrollment was later increased to 538 patients based on the percentage of patients enrolled with primary cancers other than lung and breast (sample size calculation allowed that if 25% of patients enrolled had “other” primary, a total of 501 patients would be enrolled. If “other” primary patients accounted for 30% of patients, then 538 patients would be enrolled).

1.0 Objectives

To determine the effect of RSR13 on primary and secondary efficacy endpoints in patients with brain metastases receiving daily intravenous doses of RSR13 administered immediately prior to standard WBRT/supplemental oxygen compared to patients receiving standard WBRT/supplemental oxygen.
To determine the safety of RSR13 in this patient population.
To assess the pharmacokinetics of RSR13 in the patient cohort receiving the study drug.
The primary efficacy endpoint in this study was survival in the total population. A secondary analysis of the NSCLC/breast primary tumor subpopulation was also planned with the addition of amendment # 2.
Secondary efficacy variables were time to radiographic tumor progression, time to clinical tumor progression in the brain, response rate in the brain, cause of death, and quality of life.

1.1 Overall Survival

The primary efficacy endpoint was overall survival using the log-rank statistic unadjusted for covariates. The primary final analyses of this study was undertaken when the planned number of deaths in both the total study population and the NSCLC/breast subpopulation was observed.

Reviewer comment: While overall survival in the intent to treat population was the primary efficacy endpoint in this study, amendment #2 made provisions for a secondary analysis of the NSCLC/breast primary tumor subpopulation as described above. One-hundred-seventy-three patients had been enrolled when amendment 2 was activated (protocol version 3) . See statistical review for further comments.

Secondary efficacy endpoints were time to radiographic tumor progression in the brain, time to clinical tumor progression in the brain, response rate in the brain, cause of death, and quality of life.

1.2 Time to Radiographic Tumor Progression in the Brain

Response was determined based upon evaluation of each contrast-enhanced MRI or CT scan performed after completion of the study treatment regimen. Time to radiographic tumor progression in the brain was reported by means of Kaplan-Meier estimates. Gray’s test was used to compare cumulative incidence between treatment arms. Time to first (cranial or extracranial) progression was estimated using Kaplan-Meier. Site of first progression (cranial, extracranial, simultaneous, died without documented progression, or alive without documented progression), as well as time to first failure, were summarized by treatment arm and primary site subpopulation.

1.3 Time to Clinical Progression in the Brain

Clinical progression was defined as either neurological progression, as assessed by the Neurological Function (NF) status, or as neurocognitive deterioration as measured by the Mini Mental State Examination (MMSE), or as the use of subsequent therapy for brain metastases such as radiation or surgery. An increase from baseline of 1 or more points in the NF status score indicated neurological disease progression. Neurocognitive deterioration was defined as a decrease from baseline in the MMSE score of 3 or more points. Time to clinical progression was summarized using Kaplan-Meier estimates and compared between treatment arms using cumulative incidence Grays test.

1.4 Response Rate in the Brain

Best response was determined for each patient from evaluation of MRI or CT scans.

It was projected at the outset that a differential treatment effect shown by improved

response would result in survival benefit. Treatment arms were compared using the

Cochran-Mantel-Haenszel test.

1.5 Cause of Death

The frequency of neurologic/non-neurologic/indistinguishable causes of death

was tabulated for each treatment arm and compared using the Cochran-Mantel-Haenszel test. Neurologic causes of death included such events as fatal cerebral edema, neurological deterioration, and convulsions. Non-neurologic causes of death included pneumonia, acute renal failure, cachexia, and pulmonary embolus. Patients in the indistinguishable category could not have their causes of death distinguished between neurologic and non-neurologic causes (see section 1.3, FDA Analysis).

1.6 Quality of Life

Quality of life was determined by means of the Spitzer Questionnaire and KPS assessment.

The frequency distributions were computed for each treatment by time of follow-up and focused on the 6 and 12-month time-points. The KPS score was categorized analyzed using the Cochran-Mantel-Haenszel test for each time-point.

2.0 Eligibility Criteria

Age ≥ 18 years of age.

Radiographic studies consistent with brain metastases and a histologically or cytologically confirmed primary malignancy, excluding small cell lung cancer and extrapulmonary small cell carcinomas, germ cell tumors, and lymphomas; or histologically or cytologically confirmed brain metastases consistent with a non-excluded primary malignancy. Patients with leptomeningeal metastases were not eligible.
Karnofsky Performance Status (KPS) ≥ 70 .
No prior treatment for brain metastases with WBRT, stereotactic radiosurgery, chemotherapy, hormonal therapy, immunotherapy, or biologic agents. Prior surgical resection was allowed if at least one measurable lesion remained. Prior and current corticosteroid therapy was allowed.
Adequate hematologic, hepatic, and renal function as defined by: hemoglobin ≥ 10 g/dL, WBC count ≥ 2000 cells/mm³ , platelet count ≥ 75,000/mm³, creatinine ≥ 2.0 mg/dL, bilirubin ≤ 2.0 mg/dL, alanine aminotransferase /serum glutamic-pyruvic transaminases and aspartate aminotransferase/serum glutamic-oxaloacetic transaminases ≤ 3.0 times the upper limit of normal.
Resting and exercise SpO2 while breathing room air ≥ 90%.
No other concurrent active malignancy from a second histologic site.
No use of any investigational drug, biologic, or device within 28 days prior to radiation therapy day 1.
Adequate pulmonary function tests by simple spirometry were required if the patient

had a pulmonary condition that might compromise oxygen loading in the lungs

Adequate PFTs were defined as forced vital capacity (FVC) and forced expiratory volume in 1 second (FEV₁) ≥ 50% of normal for that patient’s age, height, and race.

No chemotherapy for primary tumor or extracranial metastases within 7 days prior to WBRT day 1; no planned chemotherapy during WBRT; no planned additional therapy for brain metastases through the initial follow-up visit (1 month after completion of the RT course).
No previous exposure to RSR13
Able to provide written informed consent.
If female patients who are not post-menopausal (> 12 months since last menses) or surgically sterile, must have an negative serum β-hcg pregnancy test, and must be practicing a medically acceptable contraceptive regimen from the time of consent until the initial follow-up visit. All male patient who are not surgically sterile must be practicing a medically acceptable contraceptive regimen.

Reviewer comment: Inclusion and exclusion criteria were combined under the heading of Eligibility Criteria in Version 4 (final version) of the RT-009 protocol. Amendment #2 specified that in addition to small cell lung cancer, extrapulmonary small cell carcinomas were excluded from enrollment in the study.

3.0 Treatment Plan

Patients who were randomized to treatment arm A received daily RSR13 within 30 minutes prior to daily WBRT and supplemental oxygen. Patients in treatment arm B received WBRT and supplemental oxygen without placebo. RSR was administered using Dosing Adjustment Guidelines. Patients in treatment arm A received an initial 100 or 75 mg/kg dose of RSR13 at a concentration of 20 mg/mL over 30 minutes through a central venous catheter. All RSR13 infusions were administered using a volumetric pump. The RSR13 was given with supplemental oxygen beginning on Day 1 of radiation initiation and continued every radiation therapy day throughout the 10-day course of WBRT. WBRT was given within 30 minutes of completing the RSR13 infusion. Patients in both treatment arms received supplemental oxygen for at least 35 minutes prior to, during, and for at least 15 minutes after completion of daily WBRT. The flow rate of supplemental oxygen was 4 L/minute as needed to maintain a SpO₂ measurement ≥ 90% during and after RSR13 infusion.

Reviewer comment: According to the treatment protocol, supplemental oxygen was to be administered beginning at least 5 minutes prior to starting the RSR13 infusion. The supplemental oxygen was then continued throughout the duration of the RSR13 infusion and discontinued at least 15 minutes after completion of WBRT. Whole Brain Radiation Therapy was administered within 30 minutes after completion of the RSR13 infusion.

Whole brain radiation therapy was given as 30 Gy at 3 Gy fractions per day, 5 days per week over 10 days. Patients were stratified for enrollment by RPA (recursive partitioning analysis) Classes I and II according to the RTOG RPA of prognostic factors criteria (Table 3). RPA Class II patients were further stratified by site of the primary cancer (Table 4). The number of patients in each of the 4 strata was not predetermined.

Table 3: Recursive Partitioning Analysis (RPA)

	CLASS I	CLASS II
KPS > 70	Yes	Yes
Primary tumor	controlled	uncontrolled
Age	< 65 years	≥ 65 years
Metastases	Brain only	Brain and other

Table 4: Stratification at Randomization

Stratification (pre-defined subsets)	Total N	Control	RSR13
RPA Class I patients	57	28	29
RPA Class II NSCLC primary patients	263	131	132
RPA Class II breast primary patients	101	49	52
RPA Class II other primary patients	111	54	57
Totals	532	262	270

Reviewer comment: The protocol stated that 538 patients were analyzed for efficacy (mod 2- vol. 2, p. 14). However, only 532 patients were calculated from the stratification at randomization table as shown above. The sponsor was asked to clarify the discrepancy in numbers. They responded by explaining that 6 patients were improperly stratified based on their site of primary disease (patient numbers 2168, 3044, 3089, 4045, 4076, 4100), and were not included in the in-text table. These same 6 patients were included in the post-text table which provides summary information by stratum and site of primary within stratum.

4.0 Treatment Modifications

The selection of the RSR13 doses given in this study was based on the safety and efficacy results

obtained in the Phase 2 open-label studies in which over 270 cancer patients (which included 69 patients with brain metastases) received repetitive daily RSR13 infusions prior to RT. In the Phase 2 study RT-008, patients with brain metastases received WBRT with RSR13 at a dose of 100 mg/kg over 30 minutes. Adverse events leading to RSR13 dosing discontinuation were observed at 100 mg/kg in some patients and resulted in the initial development of dosing adjustments for individual patients to limit side effects that could result in early discontinuation of the study drug. Dose reductions to 75 or 50 mg/kg (or the withholding of doses) were allowed if clinical assessments or laboratory criteria indicated that the patient was experiencing exaggerated pharmacological effects or toxicities. Based upon these background data, the starting dose for RT-009 was 75 or 100 mg/kg . See Table 5.

Table 5: Dose Adjustment Guidelines

If the SpO₂ while breathing room air on any WBRT day was <90%, RSR13 was to be omitted.

DETERMINATION OF INITIAL DOSE OF RSR13

1. If SpO₂ while breathing room air at screening (at rest AND during exercise) AND on WBRT Day 1 was ≥ 93%, RSR13 was administered as follows:

a. Males

i. If weight ≤ 95 kg: 100 mg/kg

ii. If weight > 95 kg: 75 mg/kg

b. Females

i. If weight ≤ 70 kg: 100 mg/kg

ii. If weight > 70 kg: 75 mg/kg

2. If SpO₂ while breathing room air at screening (at rest OR during exercise) OR on WBRT day 1 was 90% - 92%: 75 mg/kg.

DOSE ADJUSTMENTS AFTER THE INITIAL DOSE

Down Titration

• Decrease from dose of 100 mg/kg to 75 mg/kg

• Omit RSR13 from dose of 75 mg/kg

DOWN TITRATION IF ANY OF THE FOLLOWING OCCURRED:

a. Duration of supplemental oxygen administration was >3 hours after end-infusion before SpO₂

while breathing room air returned to 90% on the previous dosing day.

b. The patient experienced nausea and/or vomiting (Grade 2 or higher) or clinically significant

(investigator judgment) hypotension associated with RSR13 within 12 hours after RSR13

administration on the previous dosing day.

c. The patient developed hypoxemia which required treatment after discharge on the previous

dosing day.

d. SpO₂ while breathing room air was 90% - 92% but had been ≥ 93% on the previous dosing day.

UP TITRATION

• Increase from dose of 75 mg/kg to 100 mg/kg

• Resume dosing at 75 mg/kg if RSR13 dose omitted

a. Increase from dose of 75 mg/kg administered on previous dosing day to 100 mg/kg if SpO₂ while breathing room air was 93% and none of the AE listed above a-c had occurred on the previous dosing day.

b. Resume dosing at 75 mg/kg after omitting RSR13 on the previous day:

• If SpO₂ while breathing room air was 90% - 92% and had been 90% - 92% on the dosing day that let to omission of RSR13 dose.

• If SpO₂ while breathing room air was ≥ 93%.

c. Dosing was not to be resumed after omitting RSR13 pm the previous day if SpO₂ while breathing room air was 90% - 92% but had been 93% on the dosing day that led to omitting the RSR13 dose.

(Derived from table 9.6, Final Study Report)

Reviewer comment: Amendment #1 changed the Dosing Adjustment Guideline to omit the use of RSR13 if SpO2 was < 90% while breathing room air on any day of radiation therapy. This was based on concerns that RSR13 could result in further hypoxemia in patients with compromised SpO₂ levels. The rationale for this amendment was appropriate for patient safety. Adverse events leading to dosing termination were observed in earlier trials of RSR13. This lead to the development of these Dosing Adjustment Guidelines that address efficacy and safety issues. These Guidelines were used throughout the duration of whole brain radiation therapy to determine on a daily basis whether a patient should be dosed at 100 mg/kg, 75 mg/kg, or have RSR13 held for that day. However, a dose of 50 mg/kg was permitted at the discretion of the investigator.

5.0 Safety Monitoring

All patients were assessed for safety, all adverse events, and all toxicities from randomization until the initial follow-up visit at 1 month after completion of the radiotherapy course. Standard follow-up visits were required 3 months after completion of the radiation therapy and every 3 months thereafter, until both radiographic and clinical progression were demonstrated and documented.

Table 6: Safety Monitoring

Evaluation	Screening Day (D) (D-21 to D0)	Baseline D-5 to D1	D2 to D9	D10	1 month FU (XRT completed)	3 month FU
Spitzer questionnaire		X		X	X	X
Resting SpO₂	X	X	X	X
Exercise SpO₂	X
PFTs	X
Physical exam	X	X		X	X	X
KPS	X	X		X	X	X
Neuro exam		X		X	X	X
MRI/CT	X				X	X
EKG	X
Hematology/chemistry	X	X		X	X
Serum pregnancy test	X
Supplemental O₆		X	X	X
Mini-MSE		X		X	X	X
AE check		X	X	X	X	X

(Derived from table 9.1, Final Study Report)

If any of the following occurred necessitating the early discontinuation of RSR13 in Treatment Arm A, the patient completed WBRT under Treatment Arm B procedures:

· The development of a significant adverse event/toxicity due to study participation as determined by the Investigator or patient.

· The development of an intercurrent illness, condition, or procedural complication that could interfere with the patient’s continuing to receive study drug.

· Voluntary patient withdrawal of consent to continue receiving study drug.

· The Investigator or Sponsor feels that it its medically in the best interest of the patient to discontinue receiving study drug.

6.0 Response Evaluation

Radiographic progression was defined by radiographic criteria which were evaluated by blinded

central review and determined from the date of randomization into the study. Determination of radiographic tumor progression in the brain was based on contrast enhanced MRI or CT scans taken at screening and compared to follow-up scans taken 1 month after the end of WBRT, 3 months after the end of WBRT, and every 3 months thereafter until death. The date of tumor progression was defined as the date of radiographic documentation that any treated lesion had enlarged by more than 25% in the bi-dimensional product. Maximum bi-dimensional measurements were used to compute the bi-dimensional product and for determination of response and radiographic progression (Table 7). The appearance of new lesions was not considered a sign of progression for the purpose of this study. However, the diagnosis of new lesions was collected.

The study protocol stated that predefined indicator lesions (the 3 largest well-defined lesions

identified before WBRT) would be followed for response to evaluate treatment effect. In

patients with 1 to 3 brain metastases, all treated lesions were followed for response. The central

reviewer could define a priori additional criteria for insuring the most appropriate assessment of

response and progression, including definitions of measurable and evaluable lesions.

Table 7: Response Rate and Radiographic Tumor Progression in the Brain

Response (Defined by Central Review)	Bi-dimensional Size of Residual Disease Compared to baseline.
Complete Response (CR)	0% for all indicator lesions, provided no treated lesion meets criteria for progression.
Partial Response (PR)	>0% to ≤ 50% for all indicator lesions, provided no treated lesion meets criteria for progression.
Stable disease (SD)	>50% to ≤ 125% for 1 or more indicator lesions, provided no treated lesion meets criteria for progression.
Progressive disease (PD)	>125% for any treated lesion.

(Derived from table 9.7, Final Study Report)

Reviewer comment: For this study, a partial response was defined as up to a 50% reduction in the bi-dimensional size of residual tumor compared to baseline, provided no treated lesion meets criteria for progression. Stable disease was defined as more than 50% tumor remaining after treatment, and not more than a 25% increase in the bi-dimensional size of residual tumor. Progressive disease represented more than a 25% increase in the bi-dimensional product from baseline. These parameters appear consistent with the WHO criteria of tumor response evaluation. We requested exactly how the response criteria for RT-009 was determined. The Sponsor replied stating that the criteria were established and agreed upon by the investigators, study chairs, Allos clinical and statistical personnel, and the head of the NeuroImaging Core Lab responsible for the central review of scans.

Allos did not require confirmation of response. The designation of CR or PR was based on “Best Response,” which was not defined in the protocol. The FDA sent Allos a query on how Best Response was determined. Allos replied stating that Best Response was determined by selecting the maximal response for a patient, starting at the one month follow-up visit and following over time until progressive disease or subsequent treatment of brain metastases (or death) occurred. Furthermore, the designation of CR or PR was made irrespective of the appearance of a new brain lesion or systemic progression.

7.0 Statistical Methods

The primary endpoint, overall survival, was compared between the treatment arms by unadjusted log-rank test. The primary analysis of efficacy endpoints was based on the intent-to-treat population. Enrolled patients could be in RPA Class I or II, which have distinct estimated

MST of 7.1 and 4.2 months, respectively. Based on an assumption that there would be a mix in this study of 20% RPA Class I and 80% RPA Class II patients, and based upon the RTOG brain metastases database, the estimated MST for patients treated with WBRT alone was 4.57 months.

A total of 402 events (deaths) were required to rule out the null hypothesis with 85% power that

the hazard ratio of the 2 treatment arms was 1 versus the alternative hypothesis that the HR

of the 2 arms was not equal to 1. Since it was assumed that up to 5% of patients could be ineligible for the analysis, it was necessary to enroll at least 501 patients initially. Total patient enrollment would be 501-538 patients depending on the percentage of patients with primary cancer other than NSCLC or breast. Sample size calculation allowed that if 25% of patients enrolled had other primary, a total of 501 patients would be enrolled. If other primary patients accounted for 30% of patients, then 538 patients would be enrolled.

In the subpopulation of patients with NSCLC and breast primary, a total of 308 deaths from both

treatment arms was required to provide 75% statistical power with a two-sided significance level

of 0.05 for estimation and hypothesis testing of treatment effect. If patient accrual was longer than the assumed 27 months, fewer patients would be required to observe 402 deaths in the total population and 308 deaths in the NSCLC/breast primary cohort. If accrual was shorter than the assumed 27 months, then either more patients or a longer follow-up period would be required to observe the required number of deaths. The NSCLC/breast primary site subpopulations were determined based on pre-randomization criteria. Allos medical monitors performed a treatment arm blinded review of all patients’ primary disease classification prior to the final analysis to assure consistent categorization of primary disease.

The hazard ratio (HR) was compared between treatment arms using the log-rank statistic

unadjusted for covariates. A modified Bonferroni adjustment was made for multiple

comparisons (co-primary analyses) . A p-value < 0.048 was required to reject the null

hypothesis that there was no difference in HR in the 2 treatment arms.

Analyses were performed for all randomized eligible patients, the NSCLC/breast

subpopulation, and by site of primary. Estimates of survival were calculated based

upon the number of RSR13 doses received. Survival was also estimated separately

by response category for each treatment arm.

The statistical analysis plan (SAP) specified 18 covariates that were collected prior to or at baseline. Ten of these are categorical variables. Two covariates (number of extracranial metastatic sites, and number of cranial lesions) are ordered and continuous. Five variables are analyzed multiple ways: KPS, age and Hgb are considered as both continuous and categorical, altitude is analyzed as categorical and continuous in both untransformed and log transformed scales, and area of cranial lesions is analyzed as ordered and continuous (3 levels), and in log transformed scale. These 5 covariates with multiple definitions, allow for 48 combinations. Cox multiple regression was performed on all 48 combinations using 17 of the 18 covariates that were defined at baseline. Cox single regression, multiple regression including all covariates, and stepwise Cox multiple regression models were performed. In addition to the baseline covariates, the Cox models were also run with subsequent therapy covariates per SAP.

Reviewer comment: The sample size was increased to a maximum of 538 patients with amendment #2 to allow for a statistically powered survival analysis of patients in the non-small cell lung cancer/breast cancer primary subpopulation , in addition to the survival analysis of all patients.

TRIAL RESULTS

*Informed consent

The individual investigator was responsible for preparing the written informed consent document for RT-009. A template for informed consent was provided by the Sponsor. The investigator was allowed to rearrange or reword the contents of the template, and add other elements or language, provided the meaning and content were not changed or deleted. The Sponsor reviewed the informed consent form used before any patient was enrolled. Written informed consent was obtained from all patients who participated in this study prior to enrollment.

*Randomization

Patients were randomized 1:1 to Treatment Arms A or B according to a permuted block design, balancing by institution within strata. The randomization was stratified by RPA Classes I and II and within RPA Class II by site of primary cancer (NSCLC vs. breast vs. other), for a total of 4 strata. Patients were assigned a 4-digit patient identification number with the first digit corresponding to the stratum number followed by the other 3 digits being numbered in the order of randomization sequentially within the stratum. The number served as patient identification for all data collected under the study.

A total of 2271 patients were screened in order to obtain the 538 patients who were randomized

into the study: 57 RPA Class I (10.6%) and 481 RPA Class II (89.4%) patients (Figure 10.1).

The most frequent reasons for failure to enroll screened patients were “patient unwilling to give

consent” and “KPS <70”, both accounting for 312 (17.9%) screen failures. RPA Class I patients

represented 10.6% of the total enrollment, thereby meeting the protocol projected mix of

10%-25% RPA Class I patients (Section 9.1). RPA Class II patients were stratified according to

the site of the primary cancer (NSCLC vs. breast vs. other) for a total of 4 strata.

Reviewer comment: Twenty-five patients were incorrectly classified at the site of accrual according to their primary tumor diagnosis. This error was captured only after central review of the Case Report Forms at an unspecified time point.

*Blinding

This study was open-label.

*Central review process

Radiographic data was forwarded to a central location for radiographic review. Radiographic data were forwarded by the investigational sites to Neuroimaging Core Laboratory

at the Cleveland Clinic Foundation for centralized radiological review.

Neuroimaging Core Laboratory

Cleveland Clinic Foundation

9500 Euclid Avenue

Section of Neurology/L10

Cleveland, Ohio 44195

Digital data, originals or duplicate originals of films of the magnetic resonance imaging or

computed tomography scans obtained at baseline and at follow-up visits were sent for

blinded central review for determination of radiographic response and progression in the brain.

Central Laboratory Facilities:

RSR13 Assays in Plasma and Red Blood Cells (RBCs)

Analytical Development Corporation

4405 N Chestnut Street

Colorado Springs, CO 80907

Routine Clinical Laboratory Tests:

Covance Central Laboratory Services SA

Rue Moise-Marcinhes 7

1217 Meyrin, Geneva

Switzerland

Covance Central Laboratory Services

8211 Scicor Drive

Indianapolis, IN 46214-2942

Sonic Clinical Trials

95 Epping Road

North Ryde, NSW 2113

Australia

*Protocol violations

A total of 202 exemptions were granted for protocol violations that occurred during the course of the study. A total of 151 exemptions were granted for failure to comply with protocol-defined time windows. Protocol deviations were defined a priori as violations in eligibility, disallowed medications, dosing violations, and patients who should have been withdrawn from the study but were not. Decisions to enroll patients who failed to meet all eligibility criteria were made on a

case-by-case basis:

Five patients in the Control arm were enrolled although they had received chemotherapy or hormonal therapy < 7 days prior to start of WBRT. According to the sponsor, neurological symptoms in theses patients indicated immediate need for whole brain radiation treatment. Three patients (2 in the Control arm and 1 in the RSR13 arm) were enrolled although their liver function tests were exclusionary. Elevations in ALT in 2 of the 3 patients were the result of metastases extending to the liver. In the third patient, elevated ALT was determined to be a temporary response to a previous biopsy under general anesthesia.
Two patients (both in the Control arm) were enrolled although they had received another investigational treatment within the previous month. Previous treatment had failed in both patients and neurological symptoms in these patients indicated immediate need for WBRT.
One patient was enrolled in the Control arm although FEV1 was 47%. This finding was determined a minor deviation since FVC was 62% and both resting and exercise SpO2 measurements were 94%.
One patient was enrolled in the RSR13 arm with a screening Hgb reported at 9.9 g/dL by a local laboratory and as 10.0 g/dL by central laboratory.

Reviewer comments: The five patients not meeting eligibility criteria because of prior chemotherapy or hormonal therapy within 7 days of RT day encompass the administration of 5-fluorouracil 6 days prior to RT day 1 (pt. # 1026), vinorelbine 5 days prior to RT day 1 (pt. # 3021), herceptin (pt. # 3077), letrozole (pt. # 3085), and gemcitabine (pt. # 4041). All five were in the control arm.

The 3 patients not meeting eligibility criteria because of inadequate hepatic function involved patient # 1052 (elevated ALT was attributed to recent biopsy of melanoma under general anesthesia- control arm), patient # 4007 (elevated ALT attributed to widespread metastases to the liver- control arm), and patient# 2016 (ALT value elevated, but value not reported – control arm).

Patient # 2012 and 4058 (both in the control arm) were given other investigational agents within 28 days prior to RT day 1: 2012 had been involved in a phase 2 study prior to entry, and # 4058 had been in a prior phase 3 experimental vaccine therapy for melanoma when this patient developed brain metastases.

Patient # 2006 (control arm) had a screening FEV1 below 50%, but FVC was 62% and resting and exercise SpO2 measurements were 94%.

Table 8: Ineligible Patients Identified by Blinded Central Review of Scans

(Derived from table10.5, Final Study Report)

Reason Ineligible	Primary Site	Control Patient #	RSR13 Patient #
Leptomeningeal mets	NSCLC	2163	2025
		2069	2101
		2190	2263
		2227
		1043
	Breast	3065	3016
		3092	3072
		3068
	Other	4055	4103
		4088
		4108
		4040
No measurable brain	NSCLC	2048
lesions (after resection)
	Breast	1020
		1025
Small cell lung cancer	Other	4012
Dural disease	Breast	3015

Reviewer comment: The table above identifies the patients found ineligible after central review of scans. Patients were required to have measurable disease for enrollment according to the protocol. Furthermore, patients with leptomeningeal metastases, dural disease or small cell lung cancer were ineligible according to protocol criteria.

*Enrollment

Table 8 lists the regions of accrual to both arms of the study. The top 4 accrual sites were Sheerbrooke, Canada (34), Phoenix, AZ (30), Tucson, AZ (41), and Cleveland, OH (36).

Table 9 lists the number of investigational sites per country.

Table 9: Number of Patients Enrolled in Each Region by Primary Site Subpopulation and Treatment Arm (ROW = rest of world)

Region

Primary Site

Control

(N=267)

N(%)

RSR13

(N=271)

N(%)

Canada

NSCLC

Breast

Other

Total

56(21.0)

11(4.1)

16(6.0)

83(31.1)

52(19.2)

9(3.3)

17(6.3)

78(28.8)

ROW

NSCLC

Breast

Other

Total

30(11.2)

14(5.2)

58(21.7)

31(11.4)

15(5.5)

14(5.2)

60(22.1)

USA

NSCLC

Breast

Other

Total

65(24.3)

30(11.2)

31(11.6)

126(47.2)

65(24.0)

36(13.3)

32(11.8)

133(49.1)

(Derived form table 10.1, Final Study Report)

Table 10: Number of Investigational Sites Per Country

COUNTRY	NUMBER OF INVESTIGATIONAL SITES
United States	40
Canada	15
Australia	4
Hungary	4
Belgium	3
Germany	3
Israel	3
France	3
Italy	2
Scotland	2
Spain	2
England	1

(Derived from table 6.1, Final Study report)

*Baseline Demographics

Table 11: Demographic Variables

(Derived from table 2.7.3.3.1, Summary of Clinical Efficacy)

Table 12. Reviewer’s Description of Tumor Histology

Histology

Control

267 patients

N(%)

RSR13

271 patients

N(%)

Lung

Breast

*Other:

Melanoma

Colorectal

Renal cell

151(56)

55(20)

61(23)

16(6)

10(4)

6(2)

148(55)

60(22)

63(23)

22(8)

9(3)

10(4)

*Predominant histology of “other” category.

Reviewer comment: The demographic variables and primary tumor histological types for patients enrolled in this study were evenly distributed between the two treatment arms.

Table 13: Distribution of Controlled and Uncontrolled Primary Tumors Between Treatment Arms

		Control		RSR13
Primary	N(%)	Controlled	Uncontrolled	Controlled	Uncontrolled
Site		N(%)	N(%)	N(%)	N(%)
Breast	115(21)	18(27)	37(18)	19(26)	41(21)
NSCLC	299( 56)	32(48)	119(60)	30(42)	118(59)
Other	124(23)	17(25)	44(22)	23(32)	40(20)
Total	538	67	200	72	199

Reviewer comment: The distribution of patients with controlled and uncontrolled primary tumors were even except within the “other” histological subgroup in which the RSR13 arm contained more controlled primary tumors than those in the control arm.

Table 14: Distribution of Breast Histology Between Treatment Arms

Primary Site

N(%)

Control

N(%)

RSR13

N(%)

Breast:

Infiltrating ductal:

Infiltrating lobular:

Other:

92(80)

4(3)

19(17)

46(84)

1(1)

8(15)

46(77)

3(5)

11(18)

Total

115

Reviewer comment: The various breast histological subtypes were evenly distributed between the two treatment arms.

Table 15: Distribution of KPS Score, Type of Treatment for Primary Malignancy, and Surgical Resection Across Treatment Arms.

Parameter

Control

RSR13

NSCLC

N=151

Breast

N=55

Other

N=61

NSCLC

N=148

Breast

N=60

Other

N=63

KPS:

90 – 100

< 90

Prior Treatment of the Primary Malignancy:

surgical resection

radiation Therapy

chemotherapy

hormonal Therapy

Surgical Resection of Brain metastases

Reviewer Comment: The NSCLC and Breast subgroups of the control arm had more radiation therapy as prior treatment of the primary malignancy than the corresponding subgroups in the RSR13 arm. This was also noted for hormonal therapy in the breast subgroup.

Table 16: Reviewer’s Table Demonstrating the Distribution of KPS in the Breast Subgroup

KPS	Control (N=55) N(%)	RSR13 (N=60) N(%)
60	0	1(2)
70	9(16)	9(15)
80	15(27)	14(23)
90	24(44)	28(47)
100	7(13)	8(13)

Reviewer comment: The distribution of KPS score was even in both treatment arms. This appears to be the case whether KPS is viewed as two categories or as five categories.

Table 17: Summary of Prior Treatment by Treatment Arm (Breast Subpopulation)

Location of Malignancy

Treatment

Breast Cancer Subpopulation

Control = 55 patients

N(%)

RSR13 = 60 patients

N(%)

Primary malignancy

Surgical resection

Radiation Therapy

Chemotherapy

Hormonal therapy

50(91)

39(71)

48(87)

34(62)

54(90)

39(65)

57(95)

32(53)

Extracranial metastases

Surgical resection

Radiation Therapy

Chemotherapy

Hormonal Therapy

9(16)

14(25)

30(55)

13(24)

6(10)

17(28)

32(53)

7(12)

Brain metastases

Surgical resection

4(7)

2(3)

Reviewer comment: There were some differences in the distribution of patients exposed to prior treatment of extracranial metastases in the breast cancer subgroup, the most notable of which appear to be in prior hormonal therapy. The number of patients in each subgroup is too small to make a statistical judgment.

Table 18: Number of Brain Lesions According to Baseline Scans (ITT Population)

Number of Brain Lesions	N	Control N(%)	RSR13 N(%)
1	98	53(26)	45(17)
2-3	162	81(31)	81(31)
>3	266	127(49)	139(52)
Total	526	261	265

Reviewer comment: Although the incidence of brain lesions appear evenly distributed between the control and RSR13 arms in the ITT population (Table 18), this did not seem to be the case for the breast subgroup or “other” subgroup (Table 19).

Table 19: Number of Brain Lesions by Primary Site of Disease in Each Treatment Group

Primary Site	Number of brain mets	Control N(%)	RSR13 N(%)
Breast (N=114)	1 2-3 >3	7(13) 9(16) 40(71)	13(22) 14(24) 31(53)
Sub-total		56	58
NSCLC (N=298)	1 2-3 >3	35(23) 51(34) 64(43)	24(16) 53(36) 71(48)
Sub-total		150	148
Other (N=114)	1 2-3 >3	11(20) 21(38) 23(42)	8(14) 14(24) 37(63)
Sub-total		55	59
Total (526)		261	265

(Derived from primary.xpt and scans.xpt datasets)

Reviewer’s comment: The sponsor was queried about the total number of patients in this table adding up to 526, rather than 538. The sponsor explained 12 patients are not included for baseline scans. Nine patients (2126, 2127, 2131, 2232, 3045, 3065, 4012, 4015, and 4113) were categorized as “scans not done”, or “scans not evaluable” and considered as patients not evaluable since baseline information was not reliable. The other three patient (3025, 2025, and 4040) had missing values for baseline scan information.

Within the breast subpopulation, the number of patients with ≥3 brain lesions was higher in the control arm than the RSR13 arm. In addition, the number of patients with only one brain lesion was higher in the treatment arm. This suggests a greater tumor burden in breast cancer patients within the control arm than the RSR13 arm, which could influence outcome. In the “other” subgroup, the control arm appears to have a greater proportion of patients with 2 to 3 brain lesions.

EFFICACY RESULT – SPONSOR’S ASSESSMENT

1.0 Primary Endpoint

1.1 Survival

Overall survival was calculated from the time of randomization into the study until death or

31 Jan 2003, whichever occurred first. All randomized patients in both treatment arms were

followed for survival until death or for a minimum of 6 months and patients that were still alive

were considered censored. The hazard rate was compared between treatment arms using the

log-rank test (unadjusted for covariates).

Log-rank Test

The observed MST for the Control arm was 4.47 months (n = 267) compared to 5.26 months for

the RSR13 treatment arm (n = 271), and no statistically significant difference was detected

between the survival distribution functions of the 2 arms using the unadjusted log-rank test

(HR = 0.877, p = 0.1688). For all eligible patients (N = 515), the observed MST for the Control arm was 4.37 months compared to 5.39 months for the RSR13 arm, a difference that was not statistically significant by log-rank test (p = 0.1549). There was also no statistically significant difference in survival between the 2 arms for randomized patients in the NSCLC/breast subpopulation (HR = 0.844, p = 0.1217), nor was there a statistically significant difference for patients in Strata 1, 2, or 4 (RPA Class I patients, RPA Class II patients with NSCLC primary, and RPA Class II patients with other primary, respectively).

The sponsor detected a significant difference between survival of the 2 arms for patients in Stratum 3 (RPA Class II patients with breast primary; HR = 0.542, p = 0.0061). There was also a significant difference between the 2 treatment arms in which patients with metachronous brain metastases in the RSR13 arm had a longer MST than metachronous patients in the Control arm (HR = 0.731, p = 0.0069). However, there was no significant difference between the 2 treatment arms for patients with synchronous brain metastases (HR = 1.267, p = 0.1598).

Table 20: Summary of Applicant’s Primary Analysis

(Derived from table 14.2.2.1.1, Final Study Report)

Population	Control		RSR13
Population	N(%)	MST	N(%)	MST	HR	95% CI	p-value
Patients: ITT Eligible	267(100) 250(94)	4.47 4.37	271(100) 265(98)	5.26 5.39	0.877 0.871	0.727, 1.057 0.719, 1.054	0.1688 0.1549
Breast and Lung	206(77)	4.47	208(77)	5.91	0.844	0.680, 1.047	0.1217
Breast	55(21)	4.57	60(22)	8.67	0.552	0.359, 0.850	0.0061
NSCLC	151(57)	4.37	148(55)	4.94	0.991	0.771, 1.273	0.9426
Other	61(23)	3.75	63(23)	4.01	1.029	0.708, 1.496	0.8812

ITT=intent to treat

MST=median survival time

Reviewer comment: There was no significant difference in overall survival in the intent to treat population using the logrank test. A significant difference in overall survival was noted in the subpopulation of breast cancer patients; however, this subpopulation was a predefined subset of patients identified for stratification purposes only. Any subgroup analysis of breast cancer patients or patients with metachronous disease should be considered exploratory.

Analyses per SAP

The estimated increase in survival of patients enrolled in arm A (RSR13 arm) was based on

the assumption that RSR13 would increase the efficacy of whole brain radiation. In this study, RSR13 was not intended to affect extra-cranial cancers. Given concern that patients may die of progression due to primary or metastatic extra-cranial tumors and that those deaths could decrease overall survival time in both treatment arms, the sponsor analyzed the survival data based on additional subsets of patients. The log-rank test was performed on the subsets of patients with controlled primary cancer and no extra-cranial metastases. It was anticipated that these patients would have a higher probability of death due to neurological progression and therefore are patients where RSR13 may have the greatest impact on survival. Estimates of survival for treatment arm A patients was also provided for each category of number of RSR13 doses received: 0-6 and ≥ 7. One hundred thirty-nine patients were classified as “Primary Disease Controlled”: 67 in the Control arm and 72 in the RSR13 arm. No statistically significant difference in survival was detected between the treatment arms in this subset (HR = 1.006, 95% CI: 0.682-1.484). One hundred eighty patients were classified as having no extracranial metastases: 96 in the Control arm and 84 in the RSR13 arm. No statistically significant difference in survival was detected between the treatment arms (HR = 1.008, 95% CI: 0.718-1.414). Two hundred eighteen (80.4%) of the patients in the RSR13 arm received at least 7 doses of RSR13. Patients in this group had a statistically significant increase in survival as compared to the RSR13 arm receiving fewer than 7 doses (HR = 0.636, p = 0.0060).

Analyses for Patients with NSCLC as the Site of Primary

The observed MST for NSCLC patients in the Control arm (n = 151) was 4.37 months compared

to 4.94 months in the RSR13 arm (n = 148), and no statistically significant difference was

detected between the survival distribution functions of the 2 arms using the unadjusted log-rank

test (HR = 0.991, p = 0.9426).

Reviewer comment: This was not a prespecified analysis. At best, this analysis can only be regarded as exploratory. As outlined in amendment #2, a secondary analysis for the NSCLC/breast primary tumor subpopulation was made at a later date (6/05/01), and did not demonstrate a significant difference in survival.

Analyses for Patients with the Breast as the Site of Primary

The observed MST for breast patients in the Control arm (n = 55) was 4.57 months compared to

8.67 months for the RSR13 arm (n = 60), and the sponsor reported a significant difference between the survival distribution functions of the 2 arms (HR = 0.552, p = 0.0061).

The sponsor also reported significant difference between the survival distribution functions of the 2 arms for patients in Stratum 3 (RPA Class II patients with breast primary; HR = 0.542, p = 0.0061).

Analyses for Patients with Other Primary Site

The observed MST for the patients with other primary in the Control arm (n = 61) was

3.75 months compared to 4.01 months for the RSR13 arm (n = 63), and no statistically

significant difference was detected between the survival distribution functions of the 2 arms

using the unadjusted log-rank test (HR = 1.029, p = 0.8812).

Reviewer comment: This was not a prespecified analysis.

Cox Regression Models

Of the 538 randomized patients, 10 patients were excluded from the Cox model analysis due to

missing values for baseline MRI/CT information (9 patients) and for missing baseline weight

(1 patient). A Cox multiple regression model was run for each of the 48 possible models (every variable plus every combination of the five variables with different possible values) for all randomized patients and by site of primary. Table 21 lists the 17 covariates used by the sponsor.

Table 21: Covariates Included in Cox Multiple Regression Models

Covariate

Site of primary*

KPS*

RPA Class*

Presence of extracranial mets*

Number of metastatic lesions*

Control of primary*

Age*

Presence of liver mets

Timing of diagnosis

Prior cranial met treatment

High enrolling center

Gender

Baseline Hgb

Altitude

Location of center

Dosing algorithm category

BDP total area

*Covariates mentioned in the original protocol as important covariates to test the relative importance of these factors for survival.

Reviewer comment: The sponsor points out that the log-rank test does not adjust for these 17 covariates and that there were imbalances in the prognostic factors between the two treatment arms. After applying Cox multiple regression models to adjust for these imbalances, the sponsor found a statistically significant difference in survival favoring the RSR13 arm. While some of these covariates may influence drug effect (e.g. higher altitude causing more release of oxygen to tissue), a literature review did not find support for the natural history of brain metastases being altered by whether one is from a high enrolling center, center in high altitude center, or any particular center location in general.

Seven covariates (RPA class, site of primary cancer, primary tumor control, age, presence of extracranial metastases, KPS, and number of metastatic lesions) were alluded to in the original protocol of RT-009. Furthermore, there is overlap of these covariates. For instance, KPS already encompasses age and KPS.

For All Randomized Patients

There was no statistically significant difference for RSR13 effect between the treatment arms for all randomized patients when there were no adjustments for covariates. However, the RSR13 indicator variable was statistically significant in 100% (48/48) of the Cox models where RSR13 treatment effect was adjusted for all other covariates. According to the Cox regression analyses, the most important non-stratification prognostic factors (ie, those that were statistically significant in all 48 Cox models) for predicting survival were: KPS, previous treatment for brain metastases indicator, number of extra-cranial metastases, gender, age, and baseline Hgb. The Control arm had a higher relative frequency of patients with the more favorable level of these covariates for all prognostic factors except KPS. The sponsor feels that this helps to explain why the Cox multiple regression model analyses were able to detect a statistically significant survival advantage for patients in the RSR13 arm compared to patients in the Control arm that the unadjusted log-rank failed to detect.

By Site of Primary

In patients with breast primary, the RSR13 indicator variable was statistically significant in

100% (48/48) of the Cox multiple regression models where RSR13 treatment effect was adjusted

for all other covariates as well as the Cox single regression model (HR = 0.552, p = 0.0069). In NSCLC and other primary patients, the RSR13 indicator variable was not statistically significant in any of the Cox multiple regression models nor the Cox single regression model.

Reviewer comment: See statistical review for further discussion of covariate analysis.

2.0 Secondary Endpoints

2.1 Time to Radiographic Tumor Progression in the Brain

Time to radiographic tumor progression (TTRP), as determined by blinded Central Radiology Review, was estimated for all patients using cumulative incidence analysis and Kaplan-Meier methods and tested between treatment arms using Gray’s test. Death was recorded as a competing risk when it occurred prior to diagnosed radiographic progression.

All Randomized Patients

There was not a statistically significant difference in the cumulative incidence of radiographic

progression between the Control arm and the RSR13 arm (χ²=0.458, p=0.4986).

By Site of Primary

There was not a statistically significant difference in the cumulative incidence of radiographic

progression between the Control arm and the RSR13 arm in the subset of patients with NSCLC

(χ²=0.055, p=0.8142), breast (χ²= 0.063, p=0.8023), or other primary (χ²=0.839, p=0.3597).

2.2 Time to Clinical Progression in the Brain

Time to clinical tumor progression (TTCP), was estimated for all patients using cumulative incidence analysis and Kaplan-Meier methods and tested between treatment arms

using Gray’s, which is used for comparing the cumulative incidence of a particular type of failure among different groups (9). Clinical progression was defined as either neurological progression, as assessed by the Neurological Function (NF) status score, or as neurocognitive deterioration as measured by the Mini Mental State Examination (MMSE) score, or as the use of subsequent therapy for brain metastases such as radiation, surgery, and/or SRS. An increase from baseline of 1 or more points in the NF status score indicated neurological disease progression. Neurocognitive deterioration was defined as a decrease from baseline in the MMSE score of 3 or more points.

Reviewer comment: This is a composite endpoint with subjective measures which can only be considered exploratory in this non-blinded clinical trial. Neurological assessments such as the Neurological Function Status Score and Mini-Mental Status Examination are of limited objectivity, especially in the non-blinded setting. The decision as to the nature and timing of subsequent treatment can be influenced by a number of variables, making interpretation of this composite endpoint even more difficult.

All Randomized Patients

There was not a statistically significant difference in the cumulative incidence of clinical

progression between the Control arm and the RSR13 arm (χ²=0.595, p=0.4407).

By Site of Primary

There was not a statistically significant difference in the cumulative incidence of clinical

progression between the Control arm and the RSR13 arm in the subset of patients with NSCLC

(χ²=1.541, p=0.2145), breast (χ²=0.846, p=0.3577), or other primary (χ²=0.377, p=0.5393).

2.3 Response Rate in the Brain

The distribution of best response in the brain was compared between RSR13 arms using the

Cochran-Mantel-Haenszel test.

All Randomized Patients

Four hundred forty-five patients had a scan after the baseline scan from which to assess

response; 216 patients in the Control arm and 229 patients in the RSR13 arm. For all randomized patients, there was not a statistically significant difference in the distribution

of response between the treatment arms (χ²= 2.3839, p = 0.1226). The point estimates of

response rate (complete plus partial response) were 37.5% in the Control arm and 45.4% in the

RSR13 arm. The estimated increase in response rate in patients receiving RSR13 was 7.9% with

an associated 95% confidence interval of –0.4% to 16.3% (p = 0.0609).

Patients with NSCLC as the Site of Primary

For patients with NSCLC primary, there was not a statistically significant difference between the

arms in distribution of response (χ² = 1.4216, p = 0.2331) . The point estimates of response rate (CR or PR) were 37.7% in the Control arm and 45.3% in the RSR13 arm. The estimated increase in response rate in patients receiving RSR13 was 7.5% with an associated 95% confidence interval of –3.6% to 18.7% (p = 0.1857).

Patients with Breast as the Site of Primary

For patients with breast primary, there was a statistically significant difference between the arms

in distribution of response (χ²= 5.8617, p = 0.0155). The point estimates of response rate (CR or PR) were 49.1% in the Control arm and 71.7% in the RSR13 arm. The estimated increase in response rate in patients receiving RSR13 was 22.6% with an associated 95% confidence interval of 5.1% to 40.0% (p = 0.0112). There were 2 covariates that were statistically significant for predicting response (CR or PR) when logistic multiple regression was performed for breast primary patients:

RSR13 treatment effect (odds ratio = 2.622 [95% CI: 1.157-5.942], p = 0.0209) and patients with

a baseline KPS ≥90 versus <90 (odds ratio = 3.806 [95% CI: 1.680-8.624], p = 0.0014).

For patients with breast primary, the number of patients in continuous remission (CR or PR)

declined over time of follow-up in the Control arm (21, 11, and 11 patients at 1, 3, and 6 months,

respectively) versus the RSR13 arm (22, 26, and 20 patients at 1, 3, and 6 months,

respectively).

Patients with Other Sites of Primary

For patients with other sites of primary, there was not a statistically significant difference

between the arms in distribution of response (χ²= 1.1994, p = 0.2735). The point estimates of response rate (complete plus partial response) were 26.2% in the Control arm and 20.6% in the RSR13 arm. The estimated increase in response rate in patients receiving RSR13 was -5.6% with an associated 95% confidence interval of -20.5% to 9.3% (p = 0.4615).

Reviewer comment: In assessing response to treatment, the FDA has the following concerns:

No predefined criteria for determining Best Response in the protocol
Confirmatory scans were not a protocol requirement
The designation of Complete Response or Partial Response was given regardless of the appearance of a new brain lesion.

Refer to section 1.2, FDA Analysis.

2.4 Cause of Death

Cause of death was determined by the investigator and attributed to 1 of 3 categories: neurologic,

non-neurologic, or indistinguishable. Patients with unknown cause of death were assigned a neurologic cause of death for calculation of all statistical tests.

All Randomized Patients

Four hundred forty-one patients died by the time of data cutoff: 221 in the Control arm and

220 in the RSR13 arm. Three patients withdrew consent and subsequently died, and therefore, have missing values for cause of death. The Cochran-Mantel-Haenszel test did not detect a difference in the distribution of cause of death between the treatment arms (χ² = 0.4361,

p = 0.5090). The Cochran-Mantel-Haenszel test did not detect a difference in the distribution of cause of death between the treatment arms in the subset of patients with NSCLC primary

(χ² = 0.0562, p = 0.8127), breast primary (χ² = 1.4692, p = 0.2255), or other primary (χ²= 0.0079,

p = 0.9292).

2.5 Quality of Life

This was determined with the KPS assessment and the Spitzer Questionnaire that were performed at baseline, WBRT day 10, and all routine follow-up visits. Comparisons of QOL measures between treatment arms focused on the 6-month and 12-month time-points and did not include WBRT day 10. KPS measurements were used to evaluate a patient’s condition. A KPS score could range from 100 (normal, no complaints, no evidence of disease) to 0 (death), thus a decrease in score indicated a worsening or deterioration in the patient’s condition. Patients must have had a KPS score of at least 70 to be eligible for enrollment. Spitzer Questionnaire (SQ) scores were based on 5 questions each worth 0-2 points for a total of 10 possible points. Patients with at least 3 of the 5 questions answered were given a scaled total score equivalent to the average score per question multiply by 5. The SQ scores at the 6-month and 1-year follow-up visits were compared to baseline for each patient and categorized as one of the following: stable or increasing, decreased by 1-2 points inclusive, or decreased by more than 2 points. The distribution of SQ categories at 6-months and at 1-year was compared between treatment arms using the Cochran-Mantel-Haenzel test.

KPS: All Randomized Patients

For all randomized patients, the distributions of KPS scores were similar at all time-points

between the 2 treatment arms, and no statistically significant difference was

detected in the distribution of KPS score categories between treatment arms at 6 months or

1 year using the Cochran-Mantel-Haenzel test: χ²= 2.0318, p = 0.1540 and χ² = 1.7727,

KPS: Patients with NSCLC as the Site of Primary

For patients with NSCLC primary, no statistically significant difference was detected in the

distribution of KPS score categories between treatment arms at 6 months and 1 year using the

Cochran-Mantel-Haenzel test: χ²= 0.2992, p = 0.5844 and χ² = 0.1221, p = 0.7268, respectively.

KPS: Patients with Breast as the Site of Primary

For patients with breast primary, a statistically significant difference was detected in the

distribution of KPS score categories between treatment arms at 6 months and 1 year using the

Cochran-Mantel-Haenzel test: χ² = 8.0212, p = 0.0046 and χ² =7.2717, p = 0.0070, respectively.

The percentages of patients with breast primary in the RSR13 arm who had a stable or an

increasing KPS score at the 6-month interval (30% [18/60]) was higher than in patients with

NSCLC primary (16% [24/148]) and patients with other sites of primary (11% [7/63]). The

percentages of Control arm patients with a stable or an increasing KPS score were similar at

every time-point across the 3 “Site of Primary” categories but lower than the breast patients in

the RSR13 arm.

KPS: Patients with Other Sites of Primary

For patients with other primary, no statistically significant difference was detected in the

distribution of KPS score categories between treatment arms at 6 months and 1 year using the

Cochran-Mantel-Haenzel test: χ² = 0.9718, p = 0.3242 and χ² = 0.2715,

Spitzer Questionnaire: All Randomized Patients

For all randomized patients, the distributions of SQ scores were similar at all time-points

between the 2 treatment arms and there was not a statistically significant

difference in the distribution of SQ scores between the treatment arms at 6 months or 1 year

using the Cochran-Mantel-Haenzel test: χ²= 1.0232, p = 0.3118 and χ² = 1.6712, p = 0.1961,

respectively. The percentages of patients in the Control arm who had a stable or an increasing

SQ score at the 6-month and 1-year intervals (15% [39/267] and 6% [15/267], respectively) were

comparable to the RSR13 arm (16% [43/271] and 9% [24/271], respectively).

Spitzer Questionnaire: By Site of Primary

There was not a statistically significant difference in the distribution of SQ scores at the 6-month

or 1-year intervals between the Control arm and the RSR13 arm in the subset of patients with

NSCLC (χ² = 1.8099, p = 0.1785 and χ²= 0.7259, p = 0.3942), or other primary (χ² = 0.8519,

p = 0.3560 and χ² = 0.2258, p = 0.6347, respectively) and at the 6-month interval in the subset of

patients with breast primary (χ²= 0.2107, p = 0.6462)(chi-square and a p-value were not

calculated at 1-year due to missing data).

EFFICACY RESULTS – FDA ASSESSMENT

1.1 Primary Endpoint – Survival

Of the 538 patients randomized at study entry, 23 were subsequently labeled ineligible (refer back to table 8), leaving 515 evaluable patients from the intent to treat population. Amendment #2 provided that the combined results of the NSCLC and breast primary tumor subpopulation would also be analyzed for efficacy. One-hundred-seventy-three patients had been enrolled by the time of this amendment.

The sponsor’s proposed indication for the use of RSR13 is as adjunctive therapy to whole brain radiation therapy in the treatment of brain metastases originating from breast cancer. The Sponsor did not find a statistically significant difference in survival between the two treatment arms when analyzed using the log-rank test (median survival time: control=4.47 months vs. RSR13=5.26 months, p=0.169). There was also no statistically significant difference in survival between the two arms for randomized patients in the NSCLC/Breast subpopulation (HR=0.877, p=0.1217).

The Sponsor retrospectively analyzed the collected data and noted significant p values for overall survival in the non-prespecified breast cancer subpopulation using the log-rank test (control arm=4.57 months vs. 8.67 months, p-0.0061).

Primary efficacy analysis per original protocol, comparing overall survival between WBRT and RSR13 + WBRT, in the ITT population using unadjusted log-rank test is presented in Table 22. There were a total of 441/538 patients who had events (deaths) at the time of the final analysis. The Kaplan-Meier curves for the ITT population are illustrated in Figure 1. The efficacy analysis in the subgroup of NSCLC/Breast primary patients is presented in Table 23. The Kaplan-Meier curves for the NSCLC/Breast subgroup is presented in Figure 2. There were 331/414 deaths in this subgroup at the time of the final analysis.

Table 22: Primary Efficacy Survival Analysis in ITT Population

Treatment

Number of Deaths

Median Survival in Months¹

(95% C.I.)

Hazard Ratio²

(95% C.I.)

P-value³

WBRT

221/267

4.5 (3.7, 5.4)

0.877

(0.727, 1.057)

0.1688

RSR13 + WBRT

220/271

5.3 (4.5, 6.2)

¹: Kaplan-Meier Estimates; ²: Hazard Ratio of RSR13 + WBRT/ WBRT;

³: unadjusted log-rank test.

Figure 1: Kaplan-Meier Survival Curves in the ITT Population

The FDA analysis confirmed the sponsor’s findings that there was no statistically significant difference in overall survival between the two treatment arms in the intent to treat population.

Table 23: Co-Primary Efficacy Survival Analysis in NSCLC/Breast Primary Cancer Subgroup*

Treatment

Number of Deaths

Median Survival in Months¹

(95% C.I.)

Hazard Ratio²

(95% C.I.)

P-value³

WBRT

167/206

4.5 (3.8, 5.4)

0.844

(0.680, 1.048)

0.1217

RSR13 + WBRT

164/208

5.9 (4.7, 7.0)

*: Corrected for mis-classification (i.e., non-randomized subgroup);

¹: Kaplan-Meier Estimates; ²: Hazard Ratio of RSR13 + WBRT/ WBRT;

³: unadjusted log-rank test.

Figure 2: Kaplan-Meier Survival Curves in the Subgroup of Patients with NSCLC/Breast Primary

The FDA analysis confirmed the sponsor’s findings that there was no statistically significant difference in overall survival between the two treatment arms in the NSCLC/Breast groups combined.

For enrollment, all patients had to have a KPS score of ≥ 70 (Table 16). As for prior (initial) treatment of the primary malignancy, the distribution of surgical resections performed, radiation therapy given, chemotherapy, and hormonal therapy administered was fairly balanced between the two treatment arms, except in the control arm for hormonal therapy (Table 17). More patients had received hormonal therapy in the control arm of the breast cancer subpopulation.

Patients were stratified at the time of enrollment by RPA Classes I and II to balance both treatment arms. RPA Class II patients were further stratified by site of the primary cancer (NSCLC vs. breast vs. other). When these results were analyzed using the log-rank test, no statistically significant difference in overall survival was observed between treatment arms. After study completion, a statistically significant difference was observed in the subgroup of breast cancer patients. However, this was a subgroup established for stratification purposes, not as a prespecified endpoint to test survival as a hypothesis in this specific subgroup. The finding in the breast subpopulation can only be considered exploratory at this time. Table 24 reveals the exploratory survival analysis in the subgroup of patients with primary breast cancer. The International Conference on Harmonisation – Guideline for Industry, section 11.4.2.8 (Examination of Subgroups), states that subgroup analyses “are not intended to ‘salvage’ an otherwise nonsupportive study but may suggest hypotheses worth examining in other studies or be helpful in refining labeling information, patient selection, or dose escalation. When there is a prior hypothesis of a differential effect in a particular subgroup, this hypothesis and its assessment should be part of the planned statistical analyses.” Please see the statistical review for further discussion.

Table 24: Exploratory Survival Analysis in the Subgroup of Patients with Primary Breast Cancer

Treatment

Number of Deaths

Median Survival in Months¹

(95% C.I.)

Hazard Ratio²

(95% C.I.)

P-value³

WBRT

47/55

4.6 (3.8, 6.2)

0.552

(0.359, 0.850)

0.0061

RSR13 + WBRT

39/60

8.7 (6.0, 11.3)

¹: Kaplan-Meier Estimates; ²: Hazard Ratio of RSR13 + WBRT/ WBRT;

³: unadjusted log-rank test and not adjusted for multiple analyses

The sponsor used the Cox multiple regression model to adjust for potential imbalances within the two treatment arms. A reference is given to Akazawa et al. (6), highlighting the regression model’s ability to adjust for the imbalance of prognostic factors between two treatment groups. Such a strategy is not intended to be used as a substitute when the primary analysis has failed according to the log-rank test.

Table 21 listed the seventeen covariates identified by the sponsor as potential imbalances between the control and RSR13 treatment groups. Only seven covariates (site of primary, KPS, RPA class, presence of extracranial metastases, number of metastatic lesions, control of primary malignancy, and age) were mentioned in Version 1 of RT-009. Furthermore, there is overlap of these covariates. For instance, KPS already encompasses age and KPS. Again, although covariates such as a high enrolling center and center location may influence the quality of clinical trial conducted, no data is available to support the notion that patients with brain metastases from high enrolling centers have more favorable outcomes than those enrolled from lower enrolling centers. Furthermore, while a high altitude may influence release of drug such as RSR13 to tissue, there is no supportive evidence that patients with brain metastases have more favorable outcomes based on the altitude at which they live.

The FDA has concerns over the existence of imbalances in the number of brain lesions between the two treatment arms in the breast subgroup as presented in Table 19. It appears the control arm had a higher percentage of patients with three or more documented brain lesions (71% in the control arm versus 53% in the RSR13 arm). This suggests a greater tumor burden in patients on the control arm within the breast subgroup that were already destined to have a shorter survival when compared to patients with fewer and possibly smaller brain lesions.

As for subsequent treatments - defined as any form of palliative therapy administered after exposure to RSR13, the distribution of surgery, radiation therapy, chemotherapy, hormonal therapy, stereotactic procedures, other research studies, unknown therapies, and no further treatment – several imbalances were noted as outlined under Tables 25 through 32. Tables 25 through 27 focuses on the Intent To Treat population broken down by subsequent treatment of extracranial metastases, primary malignancy, and brain metastases. Tables 28 through 30 focuses on the Breast subpopulation broken down by subsequent treatment of extracranial metastases, primary malignancy, and brain metastases. Table 31 and Table 32 combines subsequent treatment of extracranial metastases and primary malignancy into the category of Systemic Treatment for simplification.

Table 25: Intent to Treat Population- Subsequent Treatment of Extracranial Metastases

Treatment Type	Control (267 patients) N(%)	RSR13 (271 patients) N(%)
Surgical resection	7(3)	4(1)
Radiation therapy	47(18)	51(19)
Chemotherapy	37(14)	39(14)
Hormonal therapy	7(3)	10(4)
Stereotactic radiosurgery	0	0
Other research study	2(1)	0
Other therapy	13(5)	18(7)
Unknown	0	0
No treatment	113(42)	124(46)

The distribution of subsequent treatment types for extracranial metastases was even in both treatment arms in the intent to treat population.

Table 26: Intent to Treat Population- Subsequent Treatment of Primary Malignancy

Treatment Type	Control (267 patients) N(%)	RSR13 (271 patients) N(%)
Surgical resection	2(<1)	4(1)
Radiation therapy	39(15)	48(18)
Chemotherapy	62(23)	76(28)
Hormonal therapy	7(3)	14(5)
Stereotactic radiosurgery	0	0
Other research study	4(1)	3(1)
Other therapy	8(3)	2(1)
Unknown	0	0
No treatment	161(60)	153(56)

In general, subsequent treatment type (of the primary malignancy) were evenly distributed in the two study arms. Whether the numerically increased percentage of patients receiving chemotherapy in the RSR13 arm (23% vs. 28%) is of significance is difficult to assess.

Table 27: Intent to Treat Population- Subsequent Treatment of Brain Metastases

Treatment Type	Control (267 patients) N(%)	RSR13 (271 patients) N(%)
Surgical resection	9(3)	4(1)
Radiation therapy	8(3)	11(4)
Chemotherapy	7(3)	3(1)
Hormonal therapy	0	0
Stereotactic radiosurgery	13(5)	18(7)
Other research study	0	0
Other therapy	0	0
Unknown	0	1(<1)
No treatment	236(88)	234(86)

The distribution of subsequent treatment types (for brain metastases) was even in both arms.

Table 28: Breast Subpopulation- Subsequent Treatment of Extracranial Metastases

Treatment Type	Control (55 patients) N(%)	RSR13 (60 patients) N(%)
Surgical resection	3(5)	1(2)
Radiation therapy	11(20)	16(27)
Chemotherapy	16(30)	21(35)
Hormonal therapy	6(11)	9(15)
Stereotactic radiosurgery	0	0
Other research study	0	0
Other therapy	11(20)	12(20)
Unknown	0	0
No treatment	24(44)	21(35)

Subsequent treatment of extracranial metastases in the breast subpopulation with radiation therapy, chemotherapy, or hormonal therapy was numerically greater in the RSR13 arm. It is difficult to assess the significance of this finding given the small number of patients involved.

Table 29: Breast Subpopulation- Subsequent Treatment of Primary Malignancy

Treatment Type	Control (55 patients) N(%)	RSR13 (60 patients) N(%)
Surgical resection	1(2)	1(2)
Radiation therapy	2(4)	2(5)
Chemotherapy	11(20)	12(20)
Hormonal therapy	7(13)	12(20)
Stereotactic radiosurgery	0	0
Other research study	0	0
Other therapy	1(2)	1(2)
Unknown	0	0
No treatment	35(64)	36(60)

Subsequent treatment of the primary malignancy with hormonal therapy was numerically greater in the RSR13 treatment arm compared to control.

Table 30: Breast Subpopulation- Subsequent Treatment of Brain Metastases

Treatment Type	Control (55 patients) N(%)	RSR13 (60 patients) N(%)
Surgical resection	1(2)	1(2)
Radiation therapy	1(2)	2(3)
Chemotherapy	2(4)	1(2)
Hormonal therapy	0	0
Stereotactic radiosurgery	4(7)	3(5)
Other research study	0	0
Other therapy	0	0
Unknown	0	0
No treatment	49(89)	54(90)

The distribution of subsequent treatment types for brain metastases was even in both arms of the study.

Table 31: Breast Subpopulation: Subsequent Systemic Treatment (Extracranial Metastases and for Primary Malignancy)

Treatment Type	Control (55 patients) N(%)	RSR13 (60 patients) N(%)
Surgical resection	4(7)	2(3)
Radiation therapy	13(24)	18(30)
Chemotherapy	25(45)	32(53)
Hormonal therapy	13(24)	18(30)
Stereotactic radiosurgery	0	0
Other research study	0	0
Other therapy	11(20)	13(22)
Unknown	0	0
No treatment	46(84)	45(75)

Subsequent exposure to radiation therapy, chemotherapy, and hormonal therapy were more frequent in the RSR13 arm. The percentage of patients having no further systemic therapy in the breast subpopulation was lower in the RSR13 treatment arm.

Table 32: Intent to Treat Population: Subsequent Systemic Treatment (Extracranial Metastases and for Primary Malignancy)

Treatment Type	Control (267 patients) N(%)	RSR13 (271 patients) N(%)
Surgical resection	9(3)	8(3)
Radiation therapy	81(30)	89(33)
Chemotherapy	91(34)	105(39)
Hormonal therapy	14(5)	20(7)
Stereotactic radiosurgery	0	0
Other research study	5(2)	3(1)
Other therapy	20(7)	24(9)
Unknown	0	0
No treatment	207(77)	202(74)

Within the intent to treat population, the distribution of subsequent systemic therapy types was even between both arms of the study.

Cause of death was to be determined by the investigator and documented on the individual CRF

according to 1 of 3 categories:

Neurologic cause of death: The patient had stable systemic disease and progressive disease in the brain.
Non-neurologic cause of death: death was not caused by progressive brain disease; the death was further attributed to systemic cancer if extracranial progression occurred (primary or extracranial metastases), or to other causes including unknown.
Indistinguishable cause of death: Death could have been caused by documented progressive disease in the brain and/or by documented extracranial progression.

Table 33 illustrates the distribution of neurologic and non-neurologic causes of death in the treatment arms. Neurologic causes of death included cerebral edema, neurological deterioration, and convulsions. The non-neurologic causes of death included pneumonia, acute renal failure, cachexia, and pulmonary embolus. These findings suggest that the majority of breast cancer patients with brain metastases died of non-neurologic causes of death, causes that were not influenced by RSR13. Furthermore, a notable proportion of patients died of causes that were indistinguishable.

Table 33: Cause of Death - Neurologic vs. Non-neurologic

Cause of Death	Control N(%)	RSR13 N(%)	Total N(%)
Neurologic	34(13)	36(13)	70(13)
Non-neurologic	128(48)	128(47)	256(47)
Indistinguishable	58(22)	53(19)	111(20)
Alive/NA	47(18)	53(19)	100(18)
Unknown	0	1	1
Total	267	271	538

1.2 Secondary Endpoint - Response Rates in the Brain

Table 34: Reported Response Rates in the Brain According to Sponsor (ITT)

Response	Control 267 patients N(%)	RSR13 271 patients N(%)
CR	16(6)	28(10)
PR	84(31)	95(35)
Cr+PR	100(37) CI: 0.32, 0.44	123(45) CI: 0.39, 0.52

Table 34 shows the response rates in the brain within the intent to treat population according to the sponsor’s analysis.

As already stated in this review, the FDA has concerns regarding this analysis. First, the method for determining Best Response was not given in the protocol. The sponsor replied to a query dated 2-22-04 that Best Response was determined by selecting the maximal response for a patient, starting at the 1-month follow-up visit and following overtime until progressive disease or subsequent treatment of brain metastases (or death) occurred. This is explained further in Table 35.

Table 35: Method of Determining Best Response (Sponsor’s Table)

As already stated in this review, the appearance of a new brain parenchymal lesion during or proceeding the treatment was recorded in RT-009, but was not considered a sign of progression. Table 36 removes four patients (three in the control arm and one in the RSR13 arm) labeled as a CR or PR on the same date as documentation of a new brain lesion. The percentages of actual CR’s and PR’s do not change.

Table 36: Response Rates in the Brain According to FDA Analysis

Response	Control 267 patients N(%)	RSR13 271 patients N(%)
CR	15(6)	27(10)
PR	82(31)	95(35)
CR+PR	97(36) CI: 0.31, 0.42	122(45) CI: 0.39, 0.51

The sponsor stated in response to our query that confirmatory imaging was not required according to the protocol; however, they provided estimates of confirmed responses as illustrated in Table 37. Confirmation of response was assessed by comparing the response of the first scan after the best response to the best response. If the response was the same as best response, response was considered confirmed.

Table 37: Confirmed Best Response in the Brain According to Sponsor*^

RESPONSE

CONTROL

267 PATIENTS

N(%)

RSR13

271 PATIENTS

N(%)

Best / Confirmed

NSCLC

151(56)

Breast

55(20)

Other

61(23)

NSCLC

148(55)

Breast

60(22)

Other

63(23)

CR / CR

CR / PR

PR / PR

8(5)

1(1)

22(14)

3(5)

2(4)

6(11)

1(2)

0(0)

3(5)

12(8)

5(3)

19(13)

4(6)

3(5)

18(30)

1(1)

0(0)

7(11)

Total

31(20)

11(20)

4(6)

36(24)

25(42)

8(12)

(Table provided by sponsor)

*Assessed by comparing the response of the first scan after best response to the best response. If the response was the same as best, response was considered confirmed.

^Median time to confirmation ~ 2.3 months

Because confirmatory imaging studies were not required, it is difficult to interpret the findings shown in Table 37. Furthermore, the FDA cannot adequately assess duration of response due to the lack of confirmatory scans. Given that both oxygen and radiation therapy were part of the treatment in both arms and given the issues discussed above, there is uncertainty as to the contribution of RSR13 to tumor response. Therefore, it is not likely that response rate in the brain could be used as a surrogate to predict clinical benefit in this case.

1.3 Other Secondary Endpoints

There were no statistically significant findings in Time to Radiographic Tumor Progression in the Brain and Time to Clinical Tumor Progression in the Brain.

· Cause of Death

This is discussed under section 1.1.

RT-008:

A PHASE 2 STUDY TO EVALUATE THE EFFICACY AND SAFETY OF RSR13 ADMINISTERED TO PATIENTS RECEIVING STANDARD CRANIAL RADIATION THERAPY FOR BRAIN METASTASES

PROTOCOL REVIEW

Table 38: Summary of Protocol Amendments

(Derived from Table 9.4, Final Study Report RT-008)

1.0 Objectives

Evaluate overall median survival time (MST), response rate (CR and PR in the brain), and time to tumor progression in the brain in patients after receiving daily IV doses of 100 mg/kg RSR13 administered over 30 minutes with standard WBRT for brain metastases.
Evaluate the safety of daily IV doses of 100 mg/kg RSR13 administered over 30 minutes to patients receiving standard WBRT for brain metastases.
Determine the PK/PD profile of daily IV doses of 100 mg/kg RSR13 administered over 30 minutes in this patient population.

1.1 Survival

Survival time was defined as the period from Radiation Therapy (RT) day 1 to death. All patients in this study were followed for survival until death or for a minimum of 24 months.

1.2 Response Rate in the Brain

Response was determined based upon evaluation of each patient’s MRI or CT.

1.3 Time to Tumor Progression in the Brain

Time to tumor progression was defined as the time from RT day 1 to documented disease progression.

2.0 Eligibility Criteria

Patients must have been at least 18 years of age.
Patients with histologically or cytologically confirmed breast, NSCLC primary carcinoma, melanoma, GU, or GI primary carcinomas. The type of primary carcinoma may have included the following: invasive ductal or invasive lobular adenocarcinoma of the breast; or large cell carcinoma, adenocarcinoma (including bronchoalveolar carcinoma), squamous or epidermoid carcinoma of the lung; or any melanoma, GU, or GI carcinoma.
Patients must have had either histologically or cytologically confirmed brain metastases or radiographic studies consistent with brain metastases and a histologically or cytologically confirmed malignancy as defined above. If no obvious primary cancer was seen, then a histological diagnosis consistent with a breast, NSCLC, melanoma, GU, or GI primary was sufficient for entry.
KPS ≥ 70.
Patients must have met the RTOG criteria for RPA Class I or Class II.
Patients must have had no prior treatment for brain metastases with RT, surgical resection, chemotherapy, hormonal therapy, immunotherapy, or biologic agents. Corticosteroid therapy was allowed.
Patients must not have received chemotherapy within 1 week before the start of RT. Patients may not have received chemotherapy during RT and RSR13 administration in the study.
Patients must have had a baseline resting SpO2 ≥ 90% on room air.
Patients must have had adequate hematologic, hepatic, and renal function as defined by:
WBC count ≥ 2,000 cells/mm³, hemoglobin ≥ 10 g/dL, platelet count ≥ 100,000 cells/mm³, bilirubin ≤ 2.0 mg/dL, alkaline phosphatase and transaminases ≤ times the upper limit of normal, and creatinine ≤ 2.0 mg/dL.
Patients must not have used any investigational drug, biologic, or device within 3 weeks before study initiation.
Patients who had a pulmonary condition that may have compromised oxygen loading in the lungs (eg, significant intrathoracic tumor involvement, COPD, interstitial lung disease, pulmonary embolism) must have met the following requirements: a) adequate pulmonary function tests as defined by forced vital capacity (FVC) and forced expiratory volume in 1 second (FEV1) ≥ 60% of normal for that patient’s age, height, and race; and b) an exercise SpO2 on room air ≥ 90%.
Patients, if female and not post menopausal (>12 months since last menses) or surgically sterile, must have had a negative serum beta-human chorionic gonadotropin pregnancy test, must not have been breast-feeding, and must have been practicing a medically acceptable contraceptive regimen.

3.0 Treatment Plan

The patient population consisted of patients with brain metastases who were scheduled to receive a standard 2-week course of WBRT.

A total of 69 patients were enrolled into the study: 12 RPA Class I and 57 RPA Class II patients.

Patients were enrolled from 17 investigational sites during the period from 24 Feb 1998 to

28 May 1999. The first patient consent was received on 24 Feb 1998, the date the final RSR13

treatment was administered occurred on 16 Jun 1999, and the date of the last initial (1-month)

follow-up was 26 July 1999. All patients in this study were followed for survival until death or

for a minimum of 24 months. Data were transferred to RTOG as of 23 Jul 2001. This date was

used as the censoring date for analysis purposes. As of that date, 3 patients remained alive and

each had been followed for a minimum of 24 months. The database for the study was locked as

of 23 Apr 2002.

Patients were stratified upon enrollment into RTOG RPA Class I or II because of the very

different expected survival between classes (MSTs of 7.1 and 4.2 months for Classes I and II,

respectively). Separate sample size calculations were performed by stratum: planned enrollment

was 54 RPA Class I and 50 RPA Class II patients to reach 51 and 48 evaluable patients,

respectively. Study enrollment was closed shortly after the Class II enrollment target was met; at

that time only 12 Class I patients had been enrolled. Enrollment of RPA Class I patients

proceeded slowly because of the smaller proportion of Class I patients (20%) compared to

Class II patients (65%) in the overall population of brain metastases patients.4 In addition,

potential Class I participants often received surgery, SRS, or a different RT regimen, all of

which would preclude their participation. Of the 69 patients enrolled, 55 patients completed

evaluations through the 1-month follow-up visit. Of these 55 patients, there were 4 patients who

stopped receiving RSR13 due to AEs. These patients continued their participation in the study

by completing the routine follow-up evaluations.

A total of 16 patients terminated their participation in the study. These patients terminated from the study completely as opposed to patients who terminated RSR13 dosing and remained in the study by completing the follow-up visits. There were 10 patients who terminated the study during the RSR13/RT dosing phase: 7 due to AEs, 1 due to death, 1 due to a reason specified as other, and 1 was lost to follow-up. There were 4 patients who terminated the study after completing the RSR13/RT dosing phase but prior to the 1-month follow-up: 3 due to death and 1 due to unsatisfactory response. Two additional patients terminated the study early but following the initial 1-month follow-up: 1 patient for unsatisfactory response and 1 patient for non-compliance.

RSR13 Injection was supplied by Almedica and was tested and released according to Allos’

specifications. RSR13 was formulated as a sterile solution for injection and was supplied in

single-use glass bottles as 2 g of RSR13 in 100 mL of 0.225% NaCl at a concentration of

20 mg/mL. The osmolality of 20 mg/mL RSR13 in diluent is approximately equivalent to

0.45% NaCl (half-normal saline). The dose of RSR13 in this study was to be 100 mg/kg (dosing

reductions permitted) infused at a concentration of 20 mg/mL through a central venous access

device over 30 minutes. The study drug solution for IV administration was prepared by the

pharmacist or qualified chemotherapy nurse at the study site. The RSR13 stock solution was

removed from the 100 mL glass bottles with a syringe and then passed through a 0.8 or

5.0 micron filter, with adequate capacity, directly into a commercial sterile infusion bag. One

filter was to have been used for each 100 mL of RSR13 stock solution. The RSR13 infusion

solution was prepared in the sterile infusion bag within 6 hours prior to infusion.

RSR13 was administered at a concentration of 20 mg/mL over 30 minutes through a central

venous access device at a constant rate using a volumetric pump. If the administration of RSR13

was interrupted or delayed, the infusion was to have been resumed but the total infusion duration

was not to have exceeded 45 minutes.

4.0 Treatment Modifications

Early termination from the study by a patient may have been required due to any of the following

circumstances:

1. The development of a significant adverse event/toxicity due to study participation as

determined by the investigator or the patient.

2. The development of an intercurrent illness, condition, or procedural complication that could

have interfered with the patient’s continued participation.

3. Voluntary patient withdrawal.

4. The investigator or Allos felt that it was medically in the best interest of the patient to

terminate participation in the study.

Procedures listed under 1-month follow-up/early termination in the Schedule of Events

were to have been completed in the case of early withdrawal/termination. The reason

for early termination was to have been recorded on the termination page of the case report form . Patients who terminated drug dosing, but continued to have routine follow-up visits,

were considered to have terminated dosing, but not the study.

All patients were free to withdraw from participation in this study at any time, for any reason,

specified or unspecified, and without prejudice.

5.0 Safety Monitoring

Table 39 : Schedule of procedures for RT-008

RT week

Minus

14 days

Week 1

Week 2

1 month

routine

RT Day

Screen

RSR13 administration

Oxygen administration

Brain CT/MRI

PFT/exercise SpO2

Pulse oximetry

Physical exam

CXR

Neurological Assessment

MMSE

KPS

Hematology/coags/

chemistry

Urine

PKPD

6.0 Response Evaluation

CR was defined as a disappearance of all brain lesions seen on CT scans or MRI for at least 1 month with stable or decreasing steroid dose. PR required at least a 50% decrease in all lesions for at least 1 month with a stable or decreasing steroid dose. A response of Stable Disease was defined as any lesion with shrinkage less than 50% or growth less than 25% (includes all lesions with no change in growth). Disease progression was defined as any lesion in the brain enlarged by more than 25% with a stable or increased steroid dose, any new lesion, or clinical deterioration with a stable scan image. For measurable disease, standard biperpendicular diameters of the 2-dimensional tumor image at maximum dimension were applied. For patients with more than one lesion in the brain, all lesions must have demonstrated a decrease in size with a stable or decreasing steroid dose to meet the criteria for CR or PR.

7.0 Statistical Methods

The primary objective of this study was to estimate the median survival time (MST) of patients with brain metastases treated with RSR13 and RT. Since MST for patients with brain metastases may be influenced by prognostic factors, sample sizes were calculated for each of the RPA classes addressed in this study. RPA Classes I and II formed the strata for this study.

A two-sample test of significance at 0.10 (one-sided) and a detectable improvement of at least

55% would have had a statistical power of 88% in RPA Class I with 51 evaluable patients

compared to the historical control. For RPA Class II patients, the required sample size was

48 evaluable patients for a one-sided significance level of 0.05 and a detectable difference of at

least 67%. The target sample sizes of 54 RPA Class I and 50 RPA Class II patients allowed for a

5% rate of unevaluable patients. Patients who received 7 or more doses of RSR13 were

considered evaluable, however, the criterion was not used in analyses.

No substitutions were made for missing or poor quality samples. No formal interim analyses of study data were performed.

TRIAL RESULTS

*Informed consent

Each patient gave his or her written informed consent to participate in the study prior to or

during the screening visit. The consent was witnessed, dated, and retained as part of the study

records. A second original of the consent form was given to the patient.

*Randomization

This was an open-label study. There was no placebo control.

*Blinding

This was unblinded.

*Central review process

RSR13 Assays in Plasma and Red Blood Cells

Analytical Development Corporation

4405 N Chestnut Street

Colorado Springs, CO 80907

Pharmacodynamic (PD) Determinations

Allos Pharmacodynamic Laboratory

Virginia Biotech Research Park

800 Leigh Street, Suite 212

Richmond, VA 23219

Routine Clinical Laboratory Tests (Hematology, Chemistry, Coagulation, and

Urinalysis Parameters)

Covance Classical Laboratory Services

8211 SciCore Drive

Indianapolis, IN 46214

Study site monitoring was conducted at regular intervals by Allos Clinical Development staff:

Carrie VanDuym, Marilyn Craig, Margie Suhs, and Catherine Feutz. Monitoring was also

conducted by clinical research associates (CRAs) of Endpoint Research Limited and Health

Research Management, Inc. Monitoring was performed in accordance with applicable

regulations and Good Clinical Practice guidelines.

Data management and analyses for the final report were provided by Allos. The lead Data

Manager was Karen Guisinger. Adam Boyd, John Hackman, and Jim Kennedy performed

analyses and produced tables. Allos Clinical Data Management personnel performed a Quality

Control (QC) audit of the database for final reporting.

An analysis comparing results of this study with those of the Radiation Therapy Oncology Group

Brain Metastases Database (RTOG BMD) was conducted by Charles Scott, PhD, Associate

Director, Quality of Life Research, American College of Radiology; 1101 Market Street,

14th Floor; Philadelphia, PA 19107.

*Protocol violations

Major protocol deviations were defined as violations in eligibility, disallowed medications,

dosing violations, and patients who should have been withdrawn from the study but were not.

None of the patients with protocol deviations were excluded from analysis.

A total of 12 patients had protocol deviations in 6 different categories of eligibility. For the majority of these deviations, Allos granted an exemption to allow the patient to enter the study. Only 1 patient did not meet more than 1 eligibility criteria (Patient 224). Having prior treatment for brain metastases was the most common violation in eligibility.

Two patients had protocol deviations related to the requirement of having a MRI/CT scan within

2 weeks of the projected start of RT:

· Patient 229 had a CT scan performed 3 weeks before RT.

· Patient 242 had an MRI scan performed 7 weeks before the start of RT. The deviation for Patient 242 was not discovered until after the patient was treated.

Two patients who had serious adverse events (SAEs) also had protocol deviations related to SpO2 readings. Because these patients experienced SAEs following protocol deviations

they are also being noted here:

· Patient 101 (enrolled prior to Amendment 2) was discharged on RT day 2 with an SpO2 of 87% while breathing supplemental oxygen at 2 L/minute (protocol discharge criteria required SpO2 ≥ 87% while breathing room air). On RT day 5, the RSR13 infusion was started even though the patient’s preinfusion SpO2 ranged from 85-89% (protocol required preinfusion SpO2 of ≥ 87% while breathing room air). This patient experienced an SAE on RT day 5 (hospitalization for hypotension, hypoxia, and acute renal failure), and RSR13 dosing was subsequently terminated.

· Patient 215 was discharged on RT days 1, 2, and 3 with SpO2 values of 88%, 87%, and 74%, respectively, while breathing room air (following IND Safety Letter and Amendment 2, discharge SpO2 was to be ≥ 90% while breathing room air). On RT day 4 the patient experienced the first of 2 SAEs (hospitalization for nausea, vomiting, increased intracranial pressure, and cerebral edema). The second SAE occurred on RT day 5 (hospitalization for weakness, dizziness, and hyponatremia). In addition, the patient was discharged on RT day 5 with a SpO2 of 72%.

Minor protocol deviations (eg, not performing scheduled tests, taking blood samples outside

scheduled time window, not taking scheduled blood samples) also occurred, but were not

deemed to have affected the medical status of the patient and were therefore not quantified.

Most exemptions related to dosing adjustments were granted to have RSR13 held on the first day

of RT due to procedural/timing difficulties with PICC line placement or completion of laboratory

test results. Prior to Amendment 3, exemptions were also granted for dose reductions from 100

to 75 or 50 mg/kg due to results from clinical or laboratory assessments.

Table 40

(Derived from Table 10.1, Final Study Report)

*Enrollment

A total of 69 patients were enrolled from 16 study centers in the United States and one center in Canada.

*Baseline Demographics

Table 41: Demographic Variables

Parameter

RSR13

Total N=69 patients

N(%)

Sex

Male

Female

31(45)

38(55)

Race

Caucasian

Black

Native American

Asian

Hispanic

Other

62(90)

TABLE OF CONTENTS FOR BRIEFING DOCUMENT

Clinical Review

I. Introduction and Background

A. Drug Established and Proposed Trade Name, Drug Class, Sponsor’s Proposed Indication(s), Dose, Regimens, Age Groups

B. State of Armamentarium for Indication

C. Important Milestones in Product Development

D. Other Relevant Information

II. Description of Clinical Data and Sources

A. Overall Data

C. Post-marketing Experience

D. Literature Review – An extensive literature review, including a review of some of the sources listed below, was performed by the Sponsor.

III. Efficacy