The sponsor has submitted demographic information on 836 patients enrolled in the 5 clinical efficacy/activity trials and additional experience in expanded access studies. Safety data are provided for 620 patients enrolled in the 5 clinical efficacy/activity trials and interim data from the expanded access experience. FDA has chosen to conduct analyses primarily in the data derived from 229 of the 284 patients enrolled in the clinical studies and to utilize the expanded access data only to supplement targeted analyses of specific toxicities. The primary safety database is derived from the 5 clinical studies. The reasons for exclusion of patients from the database are summarized in the table below.

Study population for Integrated Summary of Safety

The baseline entry characteristics for the safety database are summarized below according to type of study (activity/efficacy vs. expanded access) for the 620 patients for whom safety data were provided. All studies enrolled patients with a diagnosis of follicular and/or low-grade non-Hodgkin’s lymphoma, with or without transformation to a higher grade histology, which had recurred after at least one prior cytotoxic chemotherapy regimen. The baseline entry characteristics of the two groups are generally similar, although those enrolled in the efficacy/activity studies were more heavily pretreated and had a higher proportion of intermediate grade histology and tumors with histologic transformation.

In the analysis of this application, it became apparent that there were significant amounts of missing study, in part due to a high rate of withdrawal from study, but also due to failure to collect data for patients who remained alive for analysis of survival. In an attempt to identify a subset of subjects with complete information for hematologic toxicity, the dose-limiting toxicity associated with Iodine I 131 tositumomab, FDA initially requested that the sponsor attempt to collect all possible information through a review of the primary medical records and to collect additional safety data from ongoing studies. In response, the sponsor submitted a safety database containing additional data from studies RIT-II-002, CP97-012, and CP98-020.

In review of this dataset, FDA determined that there were even greater amounts of missing information particularly for the patients who were enrolled in the expanded access protocol (CP 98-020). The proportion of patients in the expanded access experience with missing data for hematologic toxicity through the period at risk (weeks 5-9) and for documentation of recovery from hematologic toxicity (week 13) was higher than in the other studies. In addition, there were insufficient numbers of patients followed beyond 4 months post-treatment to permit accurate assessments of prolonged and persistent hematologic toxicity. There was also evidence of lack of reporting of non-hematologic adverse events. Specifically, the proportion of patients in whom any adverse events was reported in the aggregate, the proportion of patients with adverse events within organ system (e.g., GI, respiratory), and the number of adverse events per patient was lower in the expanded access dataset as compared to the efficacy/activity studies. Of note, the sponsor has not audited any of the study sites participating in the expanded access study or in sponsor-investigator studies/single patient INDs. Because of the concerns with regard to under-reporting of the adverse events, the data from CP 98-020 has not been included in the ISS with the following exceptions: Serious adverse events are included in the ISS and analysis presented as time to event data (development of HAMA seropositivity, development of hypothyroidism, development of myelodysplasia and/or secondary leukemia) include data from CP98-020.

Adverse Events- Overall

Ninety-five percent of the patients enrolled in the efficacy activity studies experienced one or more adverse events. The most common toxicities of any severity as well as the most common severe (NCI grade 3-4) toxicities were neutropenia, thrombocytopenia, and anemia. The hematologic toxicity will be presented as a separate section. The most common non-hematologic adverse events were asthenia, fever and chills, gastrointestinal toxicities (nausea, vomiting, anorexia, and diarrhea), musculoskeletal (myalgias, arthralgias), pain (unspecified and abdominal pain), headache, and rash. The most common serious adverse events were infections and second malignancies. Separate discussion will be provided for the following categories of adverse events: hematologic, infection, hemorrhagic events, infusion-related, gastrointestinal, hypersensitivity, thyroid, immune responses (HAMA), MDS and second malignancies.

Hematologic Adverse Events and Toxicity

The acute, dose-limiting toxicity of Iodine I 131 tositumomab therapy is severe neutropenia and/or thrombocytopenia with a median time from initiation of treatment (dosimetric dose) to nadir of 6 weeks (neutropenia) and 4.2 weeks (thrombocytopenia) and median duration of grade 3-4 toxicity of approximately 4 weeks. In order to achieve an accurate assessment of the depth and duration of the nadir and to confirm recovery from toxicity, FDA determined that subjects would need to be assessed at least weekly during 4 of the 5 weeks when the onset of the nadir was noted (weeks 5-9) and once at the recovery period (week 13). FDA reviewed the data from 620 patients, including 271 from studies RIT-II-000, 001, 002, 004 and CP 97-012 and 393 patients enrolled in the expanded access experience (6 patients in single patient INDs and 387 in the expanded access study CP98-020).

Patients Excluded from Hematology Safety Analyses:

Patients were excluded from all analyses of hematologic toxicity, including sensitivity analyses, if they had no laboratory data following study entry. There are nine patients in this category are summarized below. Of the 620 patients, 8 had no post-treatment platelet counts, 7 had no post-treatment hemoglobin values, and 9 had no post-treatment ANC values.

Among the remaining patients in the ISS database, 47 patients enrolled in RIT-II-000 who received a therapeutic dose below the MTD (which was based on hematologic toxicity) were excluded from analyses the analyses below. The remaining 229 subjects constitute the most complete dataset for assessment of efficacy

Missing Data

Based on the pattern of toxicity observed in individual patients and in a scatterplot of the study population, FDA considered that only those patients with a sufficient data obtained during the predicted likely period of hematologic toxicity could be adequately assessed. FDA defined sufficient data to assess for hematologic toxicity as having complete blood counts obtained in at least 4 of the 5 weeks (weeks 5-9) when the nadir might occur and at the time of the predicted recovery, which coincided with the end of the treatment period (week 13). Approximately 10% of the 229 patients enrolled in the activity/efficacy studies did not have CBC data during ³ 2 of the 5 weeks of expected toxicity (weeks 5-9) or a recovery time point (week 13). Approximately 15% of the 393 patients in the expanded access studies did not have CBC data during ³ 2 of the 5 weeks of expected toxicity.

Both infections and hemorrhagic events may occur as a complication of treatment induced cytopenias. The following analyses pooled preferred terms that may relate to either infection or to hemorrhagic events to obtain a clearer picture of the overall risks.

Infections

• fever reported in 84 patients (31%)
• infections (type not specified) reported in 47 patients (20%)
• pharyngitis reported in 27 patients (12%)
• pneumonia reported in 12 patients
• bronchitis reported in 9 patients
• Herpes zoster reported in 8 patients
• urinary tract infections reported in 7 patients
• sepsis reported in 7 patients
• sinusitis reported in 6 patients
• Herpes simplex reported in 4 patients
• cellulitis reported in 4 patients
• fungal dermatitis reported in 2 patients
• periodontal abscess reported in 1 patient

Hemorrhagic events

Analyses were conducted to assess the per-patient incidence of infections and of hemorrhagic events, which pooled the terms listed in the table below to avoid "double-counting" multiple infections in the same patient. The analysis of infectious events does not include fever as a term nor does it include febrile neutropenia. In FDA’s review, the incidence of febrile neutropenia has been under-reported in the database and the figures are not reliable. FDA will conduct an analysis of fevers occurring during a period of documented neutropenia in order to derive a more appropriate figure. An updated analysis will be available at the time of the Dec. 17, 2002, ODAC meeting.

The per-patient incidence of infection in the efficacy/activity studies was 48% (98/229) with 149 events reported in these 98 patients. The incidence in the expanded access (17%) is substantially lower and deemed unreliable by FDA. The per-patient incidence of hemorrhagic events is 12% (28/229) with 31 events reported among 28 patients enrolled in the efficacy/activity studies. The 5% incidence reported in the expanded access population is deemed unreliable.

B-cell lymphopenia.

The impact of Iodine I 131 tositumomab therapy on the number of circulating lymphocytes was assessed in patients enrolled in two studies: RIT-1-000, the Phase 1 study conducted previously treated subjects) and RIT-1-003, a single arm Phase 2 study conducted in patients with low grade NHL who had received no prior chemotherapy. As can be observed, there is considerable drop-off in the number of patients followed over time. The comparisons of time points is likely to be biased by selective retention of patients who are responding. Therefore, FDA will attempt to conduct analyses within patients over time in addition to the pooled analyses at various time points displayed below. Of note, the majority of the samples was obtained in a patient population (chemotherapy naïve) which differs from the population for which 131-Iodine tositumomab would be indicated. While the data may be qualitatively representative of the effects on CD20+ cells, the quantitative results would likely differ, as chemotherapy naive patients would be expected to have higher pretreatment counts.

CD20+ cells in the Peripheral blood Samples obtained

in Selected Patients with Sampling in RIT-I-000 & RIT-II-OO3

The sponsor cites a normal range for peripheral CD20+ cells as 14-246 cells/ m l

Infusional Toxicity

A constellation of symptoms, including fever, rigors or chills, hypotension, dyspnea, bronchospasm, and nausea, have been reported in the peri-infusional period. This constellation of adverse events is commonly observed with infusions of large proteins in doses of tens to hundreds of milligrams. All patients in the clinical studies received pretreatment with acetaminophen and an antihistamine. The value of premedication in preventing infusion-related toxicity was not evaluated in any of the clinical studies. Infusional toxicities were managed by slowing and/or temporarily interrupting the infusion. Symptomatic management was required in more severe cases.

The following table provides a listing of adverse events that occurred within 2 days of the dosimetric infusion.

Toxicities related to the antibody itself rather than the radioisotope were observed within 28 of days of the dosimetric infusion (21-14 days of the therapeutic infusion). These toxicities are attritubuted to infusion of a large protein load and to direct antibody binding. In assessing case reports, infusion-related toxicities included fever, chills, sweating, rigors, hypotension, and nausea. The table below provides the per-patient incidence for some of the commonly observed infusion-related toxicities. Analysis including a more comprehensive listing of the symptoms in this symptom complex that are temporally related to the dosimetric or therapeutic infusion, will be conducted. Based upon the list of preferred terms cited in the table below, and unrestricted by study day, the per-patient incidence of 40% for a pooled analysis of the preferred terms for fever, sweating, chills & fever and 23% for chills, sweating, and chills and fever. The latter grouping is probably more representative of the infusion-related events since fever is also a component on infectious events.

Hypersensitivity reactions

Tositumomab is a murine (mouse) antibody; administration of murine proteins to humans can result in the development of a serologic immune response commonly referred to as HAMA (human anti-murine antibody) response. Prior to the 2001 amendment for long-term follow-up, the clinical studies assessed patients for HAMA for a relatively limited period following treatment. Unfortunately, unlike antibodies directed against other targets, tositumomab therapy directly causes a reduction in the number of circulating CD20+ (B) lymphocytes, may transiently mask any immune response that may occur. This phenomenon has been observed with other CD20+-directed antibodies as well. In these circumstances, evidence of an immune response may not be detectable until the CD20+ cell population returns to pretreatment levels

A pooled analysis utilizing was conducted using only those preferred terms that may denote a severe hypersensitivity reaction. Specifically, the preferred terms were allergic reaction, face edema, injection site hypersensitivity, anaphylactoid reaction, laryngismus & serum sickness.

There were 14 patients in the efficacy/activity studies identified with one or more of these terms for a per-patient incidence of 6%. In the expanded access experience there were 10 events reported among 9 of the 391 patients for a per-patient incidence of 2%.

In review of the narrative summaries of the serious adverse events, there is one additional significant allergic reaction that was reported as hypotension in a single patient. The narrative summaries and CRFs are being re-assessed to identify any additional subjects with allergic reactions coded under other terms to further refine the estimated incidence.

Gastrointestinal Toxicity

The per-patient incidence of UGI adverse events is 38% (86/229) with 136 events observed among 86 patients. The per-patient incidence in the expanded access study is 20% (80/391). FDA believes that this figure is falsely low and is likely due to under-reporting of non-serious events. The per-patient incidence of LGI adverse events is 24% (55/229) with 78 events observed among 55 patients. The per-patient incidence in the expanded access study is 12% (48/391).

The expanded access experience includes serious adverse events reported among 387 subjects enrolled across 60 sites under Protocol CP 98-020 and 6 patients enrolled under single patient studies in investigator-sponsored INDs. The sponsor-investigator experience includes three patients treated at the University of Michigan Medical Center (Protocols ---------------------------------, two patients treated at Memorial Sloan-Kettering Cancer center (Protocols -----------------------------), and one patient treated at Stanford University Medical Center (Protocol---------------). None of these studies were audited by the sponsor. The protocol specified requirements for adverse event monitoring and reporting of adverse events were different from those in the activity and efficacy studies conducted by the sponsor, with the exception of the requirement for reporting of serious adverse events. Data from these studies are less reliable but can be included in limited safety assessments, specifically, reports of serious adverse events and time-to-event analyses (e.g., for HAMA, hypothyroidism).

Thyroid (TSH) Evaluation

The protocol-specified laboratory TSH schedule was Baseline, Month 6 and every 3 months up to year 2 (one year for RIT-II-001) for all the studies and additional week 7 and week 13 for the study RIT-I-000 and week 13 for the study RIT-II-002.

There were 598 patients (out of 620 patients in the Safety database) who had TSH measured at baseline. Forty-eight of 598 (8%) patients had an elevated TSH prior to the therapeutic dose, and an additional 22 patients had a history of thyroid medication. Thus 70 of 620 (11%) patients had a history of hypothyroidism prior to receiving their therapeutic dose. These patients were excluded from analyses of post–iodine I 131 tositumomab hypothyroidism. There were 528 patients who had normal TSH values at the baseline and did not have Thyroid medication prior to Iodine I 131 tositumomab treatment. The data are summarized below:

There were 362 patients (out of 620 patients in the Safety database) who had a TSH value after treatment. There were 34 patients who had an elevated TSH (event) during the course of follow-up. For these 34 patients, the median time to TSH elevation 10.9 months (95% CI on median 6.0 to 13.6 months; range: 1.8 months to 76.3 months, IQ range 5.7 to 18.6 months).

Once patients become hypothyroid, they continue to be hypothyroid. Therefore, the event was assumed to have occurred the first time a patient had elevated TSH for these 34 patients. The remaining 328 patients are assumed to have non-elevated TSH at their last day of TSH evaluation during the TSH follow-up, and are censored at individual patient’s last evaluation day of TSH measurements.