The Radioisotope Contributes Significantly to the Activity of Radioimmunotherapy

Agents targeting the CD20 antigen have shown significant activity in B-cell lymphomas. Rituximab, a chimeric monoclonal antibody (MAb) against CD20, induces response in ∼50% of patients with relapsed disease (1). The IgG2_a murine MAb, Tositumomab, also specifically binds to the CD20 antigen and has been shown to have in vitro and in vivo activity against human B-cell non-Hodgkin’s lymphoma (NHL) (2, 3). On binding to the CD20 antigen, both MAbs evoke immune effector functions as follows: antibody-dependent cellular cytotoxicity and complement-dependent cytotoxicity as well as apoptosis (4, 5, 6).

The efficacy of antibody therapy seems to be enhanced by conjugation of the MAb with a radionuclide (7, 8). However, the radioisotope also adds to the toxicity profile of these agents through marrow suppression and possible marrow dysplasia. The radiation emission from the isotope can provide not only direct toxicity to the cell bound by the antibody, but also toxicity to neighboring tumor and normal cells through “cross-fire radiation” that damages cells that are not accessible to the MAb, may not express the targeted antigen, and/or may be resistant to the effects of the unlabeled antibody (9).

The clinical experience with murine CD20 MAbs conjugated to iodine-131 (Iodine I 131 Tositumomab or Bexxar) or yttrium 90 (Y⁹⁰-ibritumomab tiuxetan or Zevalin) is quite extensive. Both agents have demonstrable efficacy in patients with relapsed or refractory indolent B-cell NHL (7, 8, 10, 11, 12), even after the use of rituximab (13) and are approved by the United States Food and Drug Administration.

Tositumomab and Iodine I 131 Tositumomab was introduced into clinical trials for the treatment of relapsed or refractory low-grade NHL in 1990. The regimen consists of the sequential administration of a dose of the “cold” or unlabeled MAb, Tositumomab, to improve tumor localization (14, 15) of the subsequent Iodine I 131 Tositumomab. The regimen includes a dosimetric dose to enable the calculation of a patient-specific therapeutic dose. This two-step regimen has been given the trade name BEXXAR (Corixa Corporation, South San Francisco, CA; GlaxoSmithKline, Philadelphia, PA) but is referred to herein by the generic name “Iodine I 131 Tositumomab.” Use of this regimen has resulted in significant durable clinical responses in some patients, supporting its approval for marketing in 2003. Approval was based on a study in 40 patients with relapsed/refractory disease after rituximab therapy and supported by the demonstration of durable responses in four other studies enrolling 190 patients with disease relapsed/refractory after chemotherapy. Tumor responses were documented in ∼60% of patients, with complete responses in ∼30%. Median response durations have been in excess of 12 months, with occasional durable complete tumor control (10, 11, 12, 16). Marrow suppression has been the dose-limiting toxicity.

Several observations led to interest in the possibility that the unlabeled Tositumomab could have innate antitumor effects. Clinical regressions have been observed after the imaging portion of the regimen but before the therapeutic dose of radiolabeled antibody. Because the radiation provided in the imaging dose is thought to be too small a dose to induce tumor regressions, it was hypothesized that the murine antibody alone might have significant antitumor effects. Additionally, in vitro testing has shown that Tositumomab can mediate growth suppression and apoptosis in CD20 expressing cell lines (2, 4), and the chimeric anti-CD20 antibody, rituximab, has significant antitumor activity (1, 17, 18). In addition, Tositumomab showed tumor growth inhibition in a human lymphoma xenograft model (3). Clinical and regulatory concerns provided the imperative to determine that the therapeutic benefit provided by radioconjugation justifies the increased toxicity, and to additionally document the activity of unlabeled Tositumomab. This study randomized patients with relapsed or refractory, indolent or transformed, CD20-positive NHL between unlabeled Tositumomab and Iodine I 131 Tositumomab, with each arm receiving identical doses of antibody. A unilateral crossover from unlabeled to radiolabeled Tositumomab was allowed.

The primary objective of the study was to compare the efficacy of Iodine I 131 Tositumomab to Tositumomab in patients with relapsed/refractory low-grade or transformed low-grade B-cell NHL. The primary endpoint was a comparison of complete response rates. The secondary endpoints were comparisons of overall response rate, duration of response, duration of complete response, progression-free survival, and safety. Patients were randomized to receive either Tositumomab and Iodine I 131 Tositumomab or two identical doses (485 mg) of unlabeled Tositumomab 7 to 14 days apart. Equal randomization was done and stratified by site. Patients treated with unlabeled Tositumomab who subsequently had documented tumor progression, still met the original entry criteria, and had not developed a human antimurine antibody response (HAMA) were eligible to cross over to treatment with Iodine I 131 Tositumomab. These patients were designated as Iodine I 131 Tositumomab-crossover. The study was conducted at 10 clinical sites after approval by their Institutional Review Board or Ethics Committee and written informed consent by the patients before study entry. An interim analysis was done by a Data Safety Monitoring Board who reviewed the safety and efficacy data on the first 33 patients on May 7, 1998; the Data Safety Monitoring Board recommended continued enrollment.

Enrolled patients were at least 18 years old and carried a histologically confirmed diagnosis of CD20-positive, low-grade or transformed low-grade NHL according to the classification scheme accepted at the time, the International Working Formulation (subgroups A–H; ref. 19). All of the patients had previously received at least one chemotherapy regimen that included an anthracycline or equivalent and had been refractory or experienced disease progression within 1 year of their most recent chemotherapy regimen. In addition, patients had an anticipated survival of at least 3 months, a Karnofsky performance status of >60%, an absolute granulocyte count of greater than 1,500 cells/mm³, a platelet count of >100,000/mm³, a serum creatinine of <1.5 × upper limits of normal, a total bilirubin of <1.5 × upper limits of normal, and an aspartate aminotransferase of <5 × upper limits of normal. All of the patients had bidimensionally measurable disease with at least one lesion ≥2 × 2 cm by computed tomography scans and no more than 25% of the hematopoietic marrow space occupied by NHL.

Exclusions included the following: previous stem cell transplantation, therapeutic antibodies, or radioimmunotherapy; previous or current investigational drug use; recent cytotoxic, radiation, or immunosuppressive therapy; obstructive hydronephrosis; active infection or heart disease; disease progression in a previously irradiated field within 1 year; prior malignancy other than lymphoma; previous allergy to iodine (excluding IV contrast materials); and pregnancy.

Tositumomab was produced by Coulter Pharmaceuticals, Inc. (Palo Alto, CA and now Corixa Corporation), Lonza Biologics plc (Slough, England), Cytogen Corporation (Princeton, NJ), and Boehringer Ingelheim Pharma KG (Biberach, Germany). For the initial 9 patients, radiolabeling of Tositumomab with I¹³¹ was done at each site with the Iodogen method (20). The sites did quality control testing for immunoreactivity and purity as described previously (16). For subsequent patients, radiolabeling of Tositumomab with I¹³¹ was done centrally (MDS Nordion Inc., Kanada, Canada).

The administration of Iodine I 131 Tositumomab followed previously described procedures, which included a dosimetric dose followed by a therapeutic dose 7 to 14 days later (11, 12, 14, 21).

Patients received a dose of unlabeled Tositumomab identical to that given in the radiolabeled regimen (450 mg infused over 60 minutes followed by an additional 35 mg infused over 20 minutes) on two occasions 7 to 14 days apart. The patients were premedicated in the same fashion as those in the Iodine I 131 Tositumomab arm.

Response criteria were defined prospectively at the start of the trial at a time when standardized criteria had not been defined. Response was assessed by physical examination and computed tomography scans at weeks 7 and 13, and then every 3 months for 2 years, then every 6 months until disease progression or death. A complete response was defined as complete disappearance of all of the disease-related physical or radiologic abnormalities and complete disappearance of all of the signs and symptoms related to disease, but the patient may have had residual radiographic or palpable abnormality(ies) thought to be residual scar tissue, such as an unchanging lesion of ≤2 cm in diameter by radiographic evaluation or ≤1 cm by physical examination. Patients who had demonstrable lymphoma in their baseline bone marrow biopsy were required to have a repeat bone marrow biopsy to show the absence of lymphoma for a designation of complete response. A partial response was defined as a 50% or greater reduction in the sum of the products of the longest perpendicular diameters of all of the measurable lesions, without the appearance of new lesions. Stable disease was defined as a <50% reduction and <25% increase in the sum of the products of the longest perpendicular diameters of measurable lesions, without the appearance of new lesions. Progressive disease was defined as a >25% increase from the nadir value of the sum of the products of the longest perpendicular diameters of all of the measurable lesions or the appearance of a new lesion (measuring >2 cm in diameter by radiographic evaluation or >1 cm in diameter by physical examination). All reported responses required confirmation by a repeat evaluation of the response at least 4 weeks after the initial documentation. An independent review of tumor response was done by two teams, each consisting of an oncologist and a radiologist who were not involved in the clinical trial, collectively known as the Masked Independent Randomized Radiology and Oncology Review (MIRROR) panel. The MIRROR panel reviewed radiologic and medical data through November 2001. Investigators continued to follow patients using the above criteria. Analyses of response rates, duration of response, and progression-free survival are presented based on the MIRROR panel assessments. Longer follow-up based on the investigator assessments are presented in the progression-free survival curves and in the discussion of ongoing durable responders. For these analyses, the investigator-assessed durations were used in place of the MIRROR panel durations for all of the MIRROR panel-assessed responders who continued in response at their last MIRROR panel assessment.

Beginning at 6 months, a physical examination, complete blood counts, serum chemistries, thyrotropin, and HAMA were evaluated every 3 months until the patient had progressive disease, expired, or had been followed for two years, whichever occurred first. After progressive disease or 2 years of follow-up, the patient entered a long-term follow-up phase where the use of thyroid medication (and later thyrotropin) was obtained every 6 months until death, along with a history of the development of any other malignancy and subsequent therapy for NHL.

Analyses were done with data from all of the enrolled patients (intent-to-treat). All statistical tests of treatment and subgroup differences were two-sided and set with a significance level of 0.05. Ps and confidence intervals were presented without adjustment for multiple comparisons, multiple outcomes, or multiple looks. However, in order for the treatment differences to be significant, the P for the final analysis must have been <0.049 to account for the interim analysis with the Lan-DeMets implementation of the O’Brien-Fleming sequential boundary technique.

Seventy-eight patients were enrolled into the trial from September, 1996 through June, 2000. The median follow-up was 42.6 months (range 1.9 to 71.5 months). Patient demographics and baseline characteristics for the two arms of the study were well balanced (Table 1).

Unlabeled Tositumomab was found to have a distinct activity, but a significant difference in favor of Iodine I 131 Tositumomab was detected for the primary and secondary efficacy endpoints (Table 2). The MIRROR panel-assessed confirmed complete response rate for patients treated with Iodine I 131 Tositumomab was 33% compared with 8% for patients treated with unlabeled Tositumomab (P = 0.012). The overall response rate (partial response and complete response) after Iodine I 131 Tositumomab, 55%, was significantly greater than after unlabeled Tositumomab, 19% (P = 0.002). The median duration of response was 28.1 months for patients treated with unlabeled Tositumomab and has not been reached for patients treated with Iodine I 131 Tositumomab. Resolution of B symptoms occurred in 9 of 11 patients treated with Iodine I 131 Tositumomab and in 7 of 7 patients treated with unlabeled Tositumomab. Kaplan-Meier estimates of time to progression are displayed in Fig. 1. Seventy-one percent (10 of 14) of patients who achieved a complete response after treatment with Iodine I 131 Tositumomab have not had disease progression, as assessed by either the investigator or the MIRROR panel, and remained in complete response at 29.8 to 71.1 months. Two of the three patients who achieved complete remission with unlabeled Tositumomab continued to have ongoing responses at 48.1 and 56.9 months.

Nineteen patients whose disease failed to respond to or progressed after treatment with the unlabeled antibody were subsequently crossed over to treatment with Iodine I 131 Tositumomab. Of note, only 3 of the 19 patients had achieved a partial response after unlabeled Tositumomab, and none of the patients had achieved a complete response. Resolution of B symptoms occurred in 2 of 2 patients. The efficacy data for the 19 patients in the Iodine I 131 Tositumomab-crossover arm are summarized in Table 3. Two crossover patients continue in complete response at 36 and 45 months. Of the 17 patients who did not cross over after Tositumomab therapy, 8 were ineligible because of positive HAMA, 1 has not received additional therapy, 5 received alternate therapy, and 1 died.

The overall concordance between the MIRROR panel and investigator-assessed confirmed response was 93% (95% confidence interval, 86–97%). Studies showed that site-radiolabeled antibody and centrally radiolabeled antibody were bioequivalent based on blood pharmacokinetics and had comparable clinical activity and safety.

The most commonly reported nonhematologic adverse events regardless of relationship were nausea (48%), asthenia (40%), fever (33%), and rash (31%) (Table 4). The incidence of the nonhematologic adverse events was similar between the two arms with the exception of nausea, which occurred more frequently after Iodine I 131 Tositumomab. This may have been related to the administration of iodide medications for thyroid blockade. The nonhematologic adverse events were typically transient for both groups.

Infusion-related adverse events considered by the investigators to be possibly or probably related to study drug occurred in 13 of 42 (31%), 11 of 36 (31%), and 7 of 19 (37%) of patients on the day of the dosimetric dose and 8 of 41 (20%), 6 of 36 (17%), and 2 of 19 (11%) on the day of the therapeutic dose for Iodine I 131 Tositumomab, unlabeled Tositumomab, and Iodine I 131 Tositumomab crossover, respectively. This is consistent with experience with other B-cell-targeted antibodies, where initial exposure induces more frequent and severe infusion reactions than later doses.

Hematologic toxicity is summarized in Table 5.

Forty-three percent (18 of 42) of patients treated with Iodine I 131 Tositumomab experienced a total of 25 infections. Eight cultured sources produced six identifiable pathogens. An antibiotic was administered for 88% of the infections. Twenty-five percent (9 of 36) of patients treated with unlabeled Tositumomab experienced a total of 16 infections: 12 were cultured, and 8 grew identifiable pathogens. An antibiotic was administered for all 16 infections. Forty-two percent (8 of 19) of patients treated in the Iodine I 131 Tositumomab-crossover group experienced a total of 10 infectious events. One cultured source produced an identifiable pathogen. An antibiotic was administered in 70% of these infections.

Two patients treated with Iodine I 131 Tositumomab, one patient treated with unlabeled Tositumomab and one from the Iodine I 131 Tositumomab-crossover arm, developed elevated thyrotropin levels. One of the patients treated with Iodine I 131 Tositumomab and one patient from the Iodine I 131 Tositumomab-crossover arm received thyroid medication (on days 545 and 446). Three additional patients (two treated with Iodine I 131 Tositumomab and one treated in the Iodine I 131 Tositumomab-crossover arm) were started on thyroid medication without reported thyrotropin levels.

The concordance between site and central HAMA results was 94.4%. The percentage of patients who became HAMA positive, based on a central assay, was 27% versus 19%, for Iodine I 131 Tositumomab versus unlabeled Tositumomab, respectively (P = 0.810). The median time to a positive HAMA was 96 days after treatment with Iodine I 131 Tositumomab and 49 days after treatment with Tositumomab alone. None of the 11 patients in the Iodine I 131 Tositumomab-crossover group became HAMA positive after treatment. The one-year and two-year cumulative incidence rates for patients becoming HAMA positive were 27% in patients who received Iodine I 131 Tositumomab and 19% in patients who received Tositumomab alone. No patient converted after 9 months, with a maximum follow-up of 51 months. Of the 25 patients who converted to positive HAMA, 7 developed symptoms that could be associated with serum sickness, but the frequency was identical in the patients who did not convert to positive HAMA. Of 2 patients followed for more than 12 months from HAMA positivity, both converted to undetectable within that time frame.

Four of 61 patients treated with Iodine I 131 Tositumomab were reported to have developed myelodysplastic syndrome. One of the patients was found on independent review to have myelodysplasia in a marrow specimen taken before the administration of Iodine I 131 Tositumomab. Thus, 3 patients (5%) developed myelodysplasia after Iodine I 131 Tositumomab, for an annualized incidence of 1.76%/year. These patients had been previously treated with a median of four chemotherapeutic agents, including alkylating agents and topoisomerase II inhibitors.

Antibody and radiolabeled antibody therapy represent important advances in the treatment of malignancies, supported by the commercial approvals of rituximab and two radiolabeled anti-CD20 antibodies. However, the specific activity of unlabeled Tositumomab and the contribution of radioiodine in Iodine I 131 Tositumomab has not previously been defined. It is assumed that the conjugation of a radioisotope to the antibody adds significantly to the activity, and it is clear that it adds to the toxicity. In addition, long-term toxicity to the thyroid gland and myelodysplasia have been seen. A previous study comparing the therapeutic regimen of rituximab with ⁹⁰yttrium ibritumomab tiuxetan (the Zevalin regimen) showed an advantage for the radiolabeled treatment regimen (22), but that comparison did not directly measure the contribution of the radioisotope, as significantly different doses of unlabeled antibody were given in each treatment arm.

The present multicenter study was designed to estimate the innate activity of Tositumomab and to assess the relative contribution of conjugation with ¹³¹I to the efficacy and safety of the antibody. Important in the study design was the administration of the same total dose of Tositumomab in the two arms, with ¹³¹I conjugated to the antibody in only one arm. Although this dose of unlabeled antibody was not optimized for biological or therapeutic activity or tolerability, this study showed that the murine antibody Tositumomab, when given without conjugation, was clinically active and can control signs and symptoms of NHL in some patients for a significant period of time. The confirmed complete response rate of 8% associated with two doses of unlabeled Tositumomab compares favorably to the reported 5 to 10% complete response rate after treatment with 4 weekly doses of rituximab (1, 17). The activity of the unlabeled antibody is also manifested by its ability to eliminate B symptoms in 7 of 7 patients with these symptoms. Of note, 2 patients treated with unlabeled Tositumomab remain in remission over 4 years (Fig. 1).

The conjugation of ¹³¹I to Tositumomab, however, provided a significant and meaningful improvement in the therapeutic activity of this MAb in the treatment of patients with relapsed and refractory low-grade and transformed low-grade NHL. The results show superior activity for Iodine I 131 Tositumomab, with improvement in overall response, complete response, and TTP for the radiolabeled regimen (Table 2). The crossover arm is of interest as 19 patients who were initially treated with the unlabeled antibody serve as their own control for subsequent Iodine I 131 Tositumomab treatment. None of these 19 patients initially treated with the unlabeled antibody achieved a complete response. After disease progression, crossover to Iodine I 131 Tositumomab resulted in significantly improved activity in all of the variables: overall response, complete response, median duration of response, and TTP (Table 3). The response rate in patients on crossover was higher than seen with radioimmunotherapy as initial therapy, likely representing chance variation with small sample size.

A comparison of the toxicities shows that infusion-related reactions were similar in the two arms and were typically transient and mild to moderate in nature (Table 4). Bone marrow suppression was more common in the Iodine I 131 Tositumomab and Iodine I 131 Tositumomab-crossover arms than in the unlabeled antibody arm, presumably directly related to marrow irradiation by the conjugated antibody. The Iodine I 131 Tositumomab-crossover group had twice the frequency of grade IV neutropenia (37% versus 17%) and thrombocytopenia (26% versus 12%), respectively, compared with the initial Iodine I 131 Tositumomab treatment group, but there were no serious infectious or bleeding complications.

The frequency of development of HAMA was similar in the two arms and consistent with previous experience after the Iodine I 131 Tositumomab regimen (10, 11, 12). There were no conversions to detectable HAMA in the group of patients included in the Iodine I 131 Tositumomab-crossover population. It seems that patients who do not develop a humoral response against Tositumomab after initial therapy are unlikely to develop HAMA after retreatment with Tositumomab at a time point remote from the initial dosing. There was no correlation between the development of HAMA and any clinical symptoms. Thyroid function abnormalities have been infrequent to date and also occurred after unlabeled Tositumomab. Myelodysplasia/acute myeloid leukemia is the most serious long-term toxicity reported after radioimmunotherapy, and although it remains a concern, the incidence to date does not clearly exceed that reported for patients treated extensively with chemotherapy and external beam radiotherapy to marrow producing regions (23).

This study documents that the radioimmunoconjugate provided a therapeutic effect over and above that provided by unlabeled Tositumomab. Toxicity attributable to the radiation component of this therapy was predominantly transient and manageable marrow suppression. The significantly improved frequency and durability of response would seem to justify the increased but manageable toxicity associated with radioimmunotherapy in patients with relapsed or refractory low-grade or transformed NHL.

Grateful appreciation is extended to Drs. Stephanie Gregory (Rush-Presbyterian-St. Luke’s Medical Center, Chicago, IL), Stanley Frankel and Philip Cohen (Georgetown University Medical Center), and Aldo Serafini (University of Miami, Miami, FL) for their contribution of patients to this trial.