- © 2002 by American Society of Clinical Oncology
Randomized Controlled Trial of Yttrium-90–Labeled Ibritumomab Tiuxetan Radioimmunotherapy Versus Rituximab Immunotherapy for Patients With Relapsed or Refractory Low-Grade, Follicular, or Transformed B-Cell Non-Hodgkin’s Lymphoma
- Thomas E. Witzig,
- Leo I. Gordon,
- Fernando Cabanillas,
- Myron S. Czuczman,
- Christos Emmanouilides,
- Robin Joyce,
- Brad L. Pohlman,
- Nancy L. Bartlett,
- Gregory A. Wiseman,
- Norman Padre,
- Antonio J. Grillo-López,
- Pratik Multani and
- Christine A. White
- From the Division of Internal Medicine and Hematology and Department of Radiology, Nuclear Medicine, Mayo Clinic and Mayo Foundation, Rochester, MN; Division of Hematology/Oncology and The Robert H. Lurie Comprehensive Cancer Center of Northwestern University, Chicago, IL; M.D. Anderson Cancer Center, Houston, TX; Roswell Park Cancer Center, Buffalo, NY; University of California Los Angeles Medical Center, Los Angeles, IDEC Pharmaceuticals Corporation, San Diego, and Clinical Research and Regulatory Strategy, Rancho Santa Fe, CA; Beth Israel Deaconess Medical Center, Boston, MA; The Cleveland Clinic Foundation, Cleveland, OH; and Washington University School of Medicine, St Louis, MO.
- Address reprint requests to Thomas E. Witzig, MD, Mayo Clinic, 620 Stabile Building, Rochester, MN 55905; email: witzig{at}mayo.edu This article was published online ahead of print at www.jco.org.
Abstract
PURPOSE: Radioimmunotherapy combines biologic and radiolytic mechanisms to target and destroy tumor cells, thus offering a needed therapeutic alternative for refractory non-Hodgkin’s lymphoma (NHL) patients. This phase III randomized study compares the novel radioimmunotherapy yttrium-90 (90Y) ibritumomab tiuxetan with a control immunotherapy, rituximab, in 143 patients with relapsed or refractory low-grade, follicular, or transformed CD20+ transformed NHL.
PATIENTS AND METHODS: Patients received either a single intravenous (IV) dose of 90Y ibritumomab tiuxetan 0.4 mCi/kg (n = 73) or rituximab 375 mg/m2 IV weekly for four doses (n = 70). The radioimmunotherapy group was pretreated with two rituximab doses (250 mg/m2) to improve biodistribution and one dose of indium-111 ibritumomab tiuxetan for imaging and dosimetry. The primary end point, overall response rate (ORR), was assessed by an independent, blinded, lymphoma expert panel.
RESULTS: ORR was 80% for the 90Y ibritumomab tiuxetan group versus 56% for the rituximab group (P = .002). Complete response (CR) rates were 30% and 16% in the 90Y ibritumomab tiuxetan and rituximab groups, respectively (P = .04). An additional 4% achieved an unconfirmed CR in each group. Kaplan-Meier estimated median duration of response was 14.2 months in the 90Y ibritumomab tiuxetan group versus 12.1 months in the control group (P = .6), and time to progression was 11.2 versus 10.1 months (P = .173) in all patients. Durable responses of ≥ 6 months were 64% versus 47% (P = .030). Reversible myelosuppression was the primary toxicity noted with 90Y ibritumomab tiuxetan.
CONCLUSION: Radioimmunotherapy with 90Y ibritumomab tiuxetan is well tolerated and produces statistically and clinically significant higher ORR and CR compared with rituximab alone.
THERAPEUTIC OPTIONS are limited in the treatment of relapsed or refractory, low-grade, follicular, or transformed non-Hodgkin’s lymphoma (NHL). Almost all patients relapse, regardless of the regimen used, and no single therapy shows definitive value in increasing survival for this patient population. Thus, in the absence of a curative therapy, a new agent that reduces tumor burden and conveys a prolonged, treatment-free period would be valuable. Radioimmunotherapy with yttrium-90 (90Y) ibritumomab tiuxetan (Zevalin; IDEC Pharmaceuticals, San Diego, CA) is a novel treatment that combines immunotherapeutic and radiation mechanisms. This is the first randomized phase III study to compare the efficacy and safety of a radiolabeled monoclonal antibody (90Y ibritumomab tiuxetan) with those of an unlabeled monoclonal antibody (rituximab).
Rituximab (Rituxan; Genentech, South San Francisco, CA, and IDEC Pharmaceuticals, San Diego, CA), the anti-CD20 chimeric monoclonal antibody that was a control in this study, is a frequently used therapy in the relapsed and refractory NHL patient population. An immunoglobulin (Ig) G1 kappa monoclonal antibody, rituximab has mouse variable and human constant regions. In vitro studies have demonstrated that rituximab binds human complement through the Fc portion of the antibody and lyses lymphoid B-cell lines through complement-dependent cytotoxicity and antibody-dependent cell-mediated cytotoxicity.1 This chimeric antibody targets the CD20 antigen, which is an ideal target for immunotherapy of B-cell NHL because it is expressed in most cases of NHL and on normal B lymphocytes but not on stem cells, plasma cells, or nonhematopoietic tissues. In addition, the CD20 antigen does not shed from the cell surface to form free antigen that could compete with circulating antibody.2,3
Rituximab produced responses in approximately 50% of patients with relapsed or refractory, low-grade, or follicular NHL when administered as an intravenous (IV) infusion at a dose of 375 mg/m2 once a week for four doses, with a time to progression (TTP) in responders of approximately 13 months.4-7 Six percent of patients achieved a complete response (CR) and 44% a partial response (PR). Toxicity is primarily limited to infusion-related events, including fever, rigors, chills, and, less frequently, nausea, urticaria, fatigue, headache, pruritus, bronchospasm, tumor-related pain, angioedema, and hypotension and rare instances of tumor lysis, serious mucocutaneous, and cardiac events.8 In 1997, rituximab became the first monoclonal antibody to be approved by the United States Food and Drug Administration for the treatment of cancer.
Ibritumomab, the murine IgG1 anti-CD20 antibody that is the parent of the engineered chimeric antibody rituximab, targets the same epitope on the CD20 antigen. Both ibritumomab and rituximab have been shown to have antiproliferative effects in tritiated thymidine incorporation assays and to induce apoptosis directly.9 In addition to producing direct cytotoxicity, monoclonal antibodies such as ibritumomab can be used to target radionuclides to specific tumor cells. Radiolabeled monoclonal antibodies are attractive agents for NHL therapy because lymphoma cells are inherently sensitive to radiation.10-13
Ibritumomab can be covalently linked to the MX-DTPA linker-chelator14 tiuxetan, which provides a high-affinity chelation site for indium-111 (111In) or 90Y. Ibritumomab tiuxetan can chelate with 111In to form 111In ibritumomab tiuxetan (111In Zevalin) or with 90Y to form 90Y ibritumomab tiuxetan (90Y Zevalin). 90Y is a pure beta emitter with a physical half-life of 64 hours and a maximum energy of 2.3 MeV, resulting in a mean path length of 5 mm in soft tissue.15,16 The high energy and long beta path length of 90Y may be especially advantageous in treating bulky, poorly vascularized tumors and tumors with heterogeneous antigen expression. The characteristics of 90Y make it an ideal radionuclide for effective radioimmunotherapy, delivered on an outpatient basis.
In two separate phase I and I/II studies,15,17,18 use of 90Y ibritumomab tiuxetan yielded overall response rates (ORRs) of 64% to 67%, with CRs of 26% to 28% in patients with low-grade, intermediate-grade, or mantle-cell NHL and 82% in patients with low-grade NHL. In the phase I/II study, the median TTP in responders treated with 0.4 mCi/kg was 15.4 months, and the TTP for complete responders ranged from 28.3 to 36.4+ months (median not yet reached).
The primary goal of the present study was to compare the ORR of patients receiving 90Y ibritumomab tiuxetan with that of patients receiving rituximab as a control; response was evaluated by an independent panel, Lymphoma Expert’s Confirmation of Response (LEXCOR), that was blinded to the treatment received and to the investigator assessment of response. Secondary goals were to determine TTP, duration of response (DR), and time to next anticancer therapy (TTNT) and to confirm that dosimetry is not necessary for safe administration of 90Y ibritumomab tiuxetan. Prognostic factors that could affect response to treatment were analyzed.
PATIENTS AND METHODS
Patient Selection
The protocol required the patients to have histologically confirmed, relapsed or refractory low-grade or follicular or transformed NHL requiring treatment as determined by an increase in overall tumor size, the presence of B symptoms, and/or the presence of masses that were causing ongoing clinical symptomatology. The 143 patients enrolled were stratified by histology type as nonfollicular low grade (International Working Formulation type A in this protocol), follicular (International Working Formulation types B, C, or D, or Revised European-American Lymphoma classification, follicular grade I, II, III),19 or transformed. To meet eligibility criteria, rituximab-naïve patients had to have bidimensionally measurable disease ≥ 3 cm (later amended to ≥ 2 cm) by computed tomography scans or magnetic resonance imaging, had to be at least 18 years old, not pregnant or lactating, and following accepted birth control methods, had to have a prestudy performance status of 0, 1, or 2 according to the World Health Organization scale, and had to have a life expectancy of ≥ 3 months.
Within 2 weeks before initial treatment, the following were required: hemoglobin level greater than 8 g/dL, absolute neutrophil count ≥ 1,500 cells/mm3, platelet count ≥ 150,000 cells/mm3, total bilirubin level less than 2.0 mg/dL, alkaline phosphatase level less than four times normal, AST level less than four times normal, and serum creatinine level less than 2.0 mg/dL. Patients were excluded if their bone marrow biopsy or aspirate demonstrated ≥ 25% involvement with NHL, if their peripheral-blood lymphocyte count was greater than 5,000 cells/mm3, if they had prior external-beam radiation therapy to more than 25% of their bone marrow, or if they had a history of human antimurine antibodies (HAMA) or human antichimeric antibodies (HACA).
Patients with CNS NHL, chronic lymphocytic leukemia, mantle-cell lymphoma, human immunodeficiency virus, prior radioimmunotherapy, hematopoietic growth factors within 2 weeks before treatment, or prior autologous or allogeneic stem-cell transplant were not eligible. There was no limitation on bulky disease or the number of prior antineoplastic therapies or tumor relapses. All prior chemotherapy was required to have been completed at least 3 weeks (6 weeks if nitrosourea or mitomycin) before study treatment. The study was approved by the institutional review board at each study site, and written informed consent was obtained from all patients.
Study Design and Objectives
This was a prospectively randomized, controlled, phase III study of 90Y ibritumomab tiuxetan versus rituximab. Patients were randomly assigned to either of two regimens. Control patients were to receive four once-weekly doses of rituximab 375 mg/m2 IV as previously described.7 90Y ibritumomab tiuxetan patients were to receive a regimen of rituximab 250 mg/m2 IV on days 1 and 8, 111In ibritumomab tiuxetan (5 mCi of 111In, 1.6 mg of ibritumomab tiuxetan) IV (over 10 minutes) on day 1; and 90Y ibritumomab tiuxetan 0.4 mCi/kg IV (over 10 minutes) on day 8 after the day 8 rituximab, as previously described.18,20 The dose of 90Y was capped at 32 mCi. The primary study end point was ORR. Secondary study end points were DR and TTP (in all patients and in responders). Additional efficacy end points were CR rate, complete clinical response rate, PR rate, TTNT, and quality of life. The treatment period was defined as 13 weeks from the time of the first study drug infusion, and the follow-up period was defined as a maximum of 4 years in patients with a clinical response.
Laboratory studies were to include baseline and follow-up hematology, serum chemistry, Igs, HACA/HAMA, blood T- and B-lymphocyte and natural killer subsets by flow cytometry, and bcl-2 analysis of DNA from blood and marrow cells.21,22 During the treatment period, blood counts (complete blood count/differential/platelets) were collected weekly. The Functional Assessment of Cancer Therapy–General (FACT-G) survey23 was used to analyze the alleviation or continuation of symptoms common to NHL.
Dosimetry
Imaging and dosimetry were to be conducted at the clinical sites (methods published elsewhere).24,25 No treatment with 90Y ibritumomab tiuxetan was to be administered if the predicted delivered dose of radiation was more than 20 Gy to any nontumor organ or more than 3 Gy to the bone marrow.
Response Assessment
A LEXCOR team of one radiologist and one oncologist measured all lymph nodes on computed tomography scans of the neck, chest, abdomen, and pelvis and evaluated clinical information to determine response. Response was assessed according to both protocol-defined criteria and International Workshop NHL response criteria (IWRC) but is reported here as IWRC because those are the current standards.26 The investigative site performed response evaluations 1 month after treatment completion. Additional response evaluations were to be performed through year 4 or until disease progression. The decision to initiate lymphoma therapy and the type of treatment after 90Y ibritumomab tiuxetan or rituximab were determined by patients’ treating physicians and were not dictated by this study.
Safety
Safety variables were to include all adverse events within the 13-week treatment period. During follow-up, data were collected on serious adverse events and on adverse events that were possibly or probably related or of unknown relationship to study drug. Use of prophylactic growth factors was not permitted, but therapeutic administration of growth factors and/or transfusions to treat cytopenias was to be at the investigator’s discretion. Samples for the analysis of HAMA and HACA were to be collected at baseline, week 4 (all patients), week 8 (rituximab only), and week 12 (all patients). Toxicity was to be evaluated using the National Cancer Institute’s adult toxicity criteria (common toxicity criteria version 2.0).
Statistical Methods
Treatment comparisons for all response variables were analyzed in the intent-to-treat populations using the Cochran-Mantel-Haenszel test. All inferential analyses of safety and clinical activity were two-sided and were performed at the alpha = 0.05 level. Selected prognostic factors and response variables were evaluated at baseline with Fisher’s exact two-tailed test. Treatment differences, adjusted by prognostic factors, were tested with the Cochran-Mantel-Haenszel test, and the Breslow-Day test was used to detect the interaction between prognostic factor and treatment. Stepwise logistic regression analysis was performed to evaluate the clinical response difference between 90Y ibritumomab tiuxetan and rituximab when prognostic factors were included in the model.
Secondary efficacy variables were TTP, DR, and TTNT. TTP was measured from the date of first rituximab treatment to the date of disease progression, DR from the date of the first observation of response to the date of disease progression, and TTNT from the date of the first treatment to the date of new treatment. TTP and TTNT were evaluated in all patients, and TTP and DR in responders. These parameters were stratified by histology and analyzed with Cox proportional hazard models. Estimates of median TTP, DR, and TTNT were determined using the Kaplan-Meier estimation method.27 An exploratory analysis evaluated durable response at specific time points. Patients with durable response were defined as responders with TTP greater than or equal to the specified time period, ie, 6, 9, and 12 months. Safety profiles were compared using Fisher’s exact test. The paired t test was used to test for total score difference in the quality-of-life data (FACT-G). B symptom and disease-related symptom changes from baseline data were compared using the Wilcoxon rank sum test. All calculations were performed using SAS software (version 6.12)28 under the IDEC Hewlett Packard UNIX system.
The trial design was powered to detect a 25% higher ORR in the 90Y ibritumomab tiuxetan group compared with control. Given a 0.05 alpha level, a sample size of 150 patients would yield at least 80% statistical power. A prospectively defined interim analysis was performed when efficacy data for the first 90 patients became available.29 These analyses confirmed a minimum of 80% power to detect a 25% higher ORR in the 90Y ibritumomab tiuxetan group than in the control group, with a level of significance of .005 for the interim analysis and .048 for the final analysis. The trial was not designed to detect a difference in TTP; the TTP for the 90Y ibritumomab tiuxetan group was to be clinically equivalent to that of the control group (± 1.5 months).
RESULTS
A total of 143 patients from 27 institutions were enrolled onto the study and treated; 73 were randomized to 90Y ibritumomab tiuxetan and 70 to rituximab. Patient characteristics were well balanced. There were no statistically significant differences between treatment groups at baseline (Table 1⇓). All patients had advanced disease with a median of two prior regimens (range, one to six), including the following: alkylator treatment with or without prednisone (28%); cyclophosphamide, doxorubicin, vincristine, and prednisone (CHOP) (45%); cyclophosphamide, vincristine, and prednisone (32%); single-agent purine analogs (15%), corticosteroids (20%); and other aggressive regimens, including purine analog combinations, dexamethasone, cisplatin, and cytarabine; mesna, ifosfamide, mitoxantrone, and etoposide; prednisone, methotrexate, doxorubicin, cyclophosphamide, etoposide, cytarabine, bleomycin, vincristine, and methotrexate; and others (34%). Resistance to last prior chemotherapy, defined as no response or TTP of less than 6 months, was present in 49% and 47% of 90Y ibritumomab tiuxetan and rituximab control patients, respectively.
Patient Characteristics and Disease Status by Treatment Group at Study Entry
(Continued)
Dosimetry
Dosimetry using 111In ibritumomab tiuxetan was conducted for all 90Y ibritumomab tiuxetan patients. Figure 1 shows a gamma scan from a representative patient. All patients were eligible for treatment with 90Y ibritumomab tiuxetan based on MIRDOSE3-estimated radiation-absorbed doses to normal organs and red marrow.24 Dosimetry results, reported elsewhere, demonstrate that 90Y ibritumomab tiuxetan delivers acceptable radiation-absorbed doses to uninvolved organs in all cases.25
Fig 1. 111In-labeled ibritumomab tiuxetan images from a representative patient. The 4-hour scan demonstrates the residual blood pool. The 63-hour and 140-hour scans demonstrate axillary, inguinal, and periaortic lymphoma, as well as uptake in the liver.
Efficacy
LEXCOR assessed response and durable response at 6, 9, and 12 months by study treatment, as displayed in Table 2.30 A statistically significant increase in ORR was noted for the 90Y ibritumomab tiuxetan group versus the rituximab control group when evaluated by both protocol-defined criteria and IWRC. The combined CR/CR unconfirmed (CRu) rate was 34% in the 90Y ibritumomab tiuxetan group versus 20% in the rituximab control (P = .063). In 113 patients with follicular low-grade NHL, the ORR was 86% in the 90Y ibritumomab tiuxetan group versus 55% in the rituximab control (P < .001). For 17 patients with nonfollicular low-grade NHL, the ORR was 67% (six of nine patients) in the 90Y ibritumomab tiuxetan group (PR in three of six patients with small lymphocytic lymphoma; CR in one patient with marginal zone NHL; PR in one patient with lymphoplasmacytic NHL; CR in one patient with maltoma) and 50% (four of eight patients) in the rituximab control group (PR in one of four patients with small lymphocytic lymphoma; CR in one patient; CRu in one patient; and PR in one of three patients with maltoma). In the 13 transformed patients, the ORR was 56% (five of nine patients) in the 90Y ibritumomab tiuxetan group versus 75% (three of four patients) in the rituximab control group.
LEXCOR Response Assessment and Durable Response
The Kaplan-Meier–estimated median TTP in all patients was 11.2 months (range, 0.8 to 31.5+ months) for the 90Y ibritumomab tiuxetan group compared with 10.1 months (range, 0.7 to 26.1 months) for the rituximab control (P = .173; Fig 2A). Patients with follicular histology had an estimated median TTP of 12.6 months for the 90Y ibritumomab tiuxetan group and 10.2 months for the rituximab control group (P = .062; Fig 2B). For patients achieving a CR or CRu, the median TTP could not be estimated because data from 73% of patients were censored (Fig 2C)
Fig 2. Kaplan-Meier analysis of the TTP (C, censored): (A) intent-to-treat patients; (B) intent-to-treat patients with follicular NHL; (C) patients achieving a CR or CRu.
The estimated median DR in the 90Y ibritumomab tiuxetan group was 14.2 months (range, 0.9 to 28.9 months) versus 12.1 months (range, 2.1 to 24.5+ months) for the rituximab control group (P = .644). The estimated median DR for follicular patients treated with 90Y ibritumomab tiuxetan was 18.5 months (range, 1.7 to 28.9+ months) versus 12.1 months (range, 2.7 to 24.5 months) for the rituximab control group (P = .371).
The median TTNT for the 90Y ibritumomab tiuxetan group has not been reached (range, 1.2 to 31.5+ months), as 56% of patient data are censored (Fig 3). The median TTNT after rituximab was 13.1 months (range, 0.8 to 26.1+ months), with 44% (31 of 70 patients) censored (P = .084). In patients with nontransformed histology, a significantly longer TTNT was estimated for the ibritumomab tiuxetan group (median not reached, as 61% were censored; range, 2.1 to 21.7+ months) compared with the rituximab control group (13.1 months [46% censored]; range, 1.3 to 26.1+ months) using the log-rank test (P = .040).
Fig 3. Kaplan-Meier analysis of the time to next cancer therapy in intent-to-treat patients (C, censored).
After patients were stratified by prognostic variables, 90Y ibritumomab tiuxetan produced a significantly superior response compared with rituximab. The ORR in the 90Y ibritumomab tiuxetan group was higher compared with the ORR in the rituximab control group in patients resistant to any chemotherapy (63% v 43%, P = .078) and in patients resistant to the last course of therapy (64% v 36%, P = .045). Also, the ORR was higher in the 90Y ibritumomab tiuxetan group versus the rituximab control group for patients with bulky disease ≥ 5 cm (67% v 45%, P = .130) and in patients with follicular lymphoma (76% v 47%, P < .001).
Stepwise logistic regression analysis demonstrated that no prognostic factor significantly affected ORR. The following factors were included in the model: age, sex, histology, bone marrow involvement, splenomegaly, hepatomegaly, extranodal disease, tumor bulk, number of prior regimens, International Prognostic Index score, and bcl-2 status. Therefore, the statistical analysis was performed without adjustment for any prognostic variable, which resulted in an odds ratio for ORR of 2.89 in favor of 90Y ibritumomab tiuxetan (95% confidence interval, 1.44 to 5.8; P = .003).
Eighty-one patients completed the FACT-G survey at baseline and week 12. In the 90Y ibritumomab tiuxetan group (n = 45), the mean FACT-G score improved significantly from 86.9 at baseline to 93.3 at week 12 (P = .001). The baseline score in the rituximab group (n = 36) also increased, from 90.7 at baseline to 93.4 at week 12 (P = .185).
Safety
The overall incidence of nonhematologic adverse events with a probable, possible, or unknown relationship to study treatment was similar between the treatment groups during the treatment period (P = .36). All patients received rituximab on day 1 and experienced similar adverse events on that day. Figure 4 shows the incidence of the most frequent, related nonhematologic adverse events by treatment group. Higher incidences were noted in the 90Y ibritumomab tiuxetan group than in the rituximab group for grades 1 and 2 cough (15% v 7%), bronchospasm (6% v 4%), and dyspnea (15% v 7%) and for grades 1 and 2 nausea (43% v 19%), vomiting (19% v 7%), and anorexia (11% v 3%).
Fig 4. Incidence of most frequent nonhematologic adverse events by grade (N = 143).
Five patients (7%) in the 90Y ibritumomab tiuxetan group were hospitalized with infection or febrile neutropenia: one with sepsis from an obstructed biliary stent (with a normal absolute neutrophil count), one with febrile neutropenia, two with urinary tract infection, and one with gastroenteritis. In the rituximab control group, one patient (1%) was hospitalized with gastroenteritis. All patients recovered.
Figure 5 shows the median nadir for absolute neutrophil count, hemoglobin concentration, and platelet count for the 90Y ibritumomab tiuxetan group during the treatment period, and Table 3 shows the extent and duration of hematologic toxicity. Duration of hematologic toxicity for each hematologic variable (absolute neutrophil count, platelet count, and hemoglobin concentration) was calculated using two methods. In method A, duration was measured from the date of the last laboratory value before development of grade 3 or 4 toxicity to the day of the first value in grade 2 after recovery from nadir. In method B, duration was measured from the date of the first laboratory value in grade 3 or 4 toxicity to the date of the last value in grade 3 or 4 after nadir. In the 90Y ibritumomab tiuxetan group, seven patients were treated with granulocyte colony-stimulating factor and five received prophylactic oral antibiotics while neutropenic. Thirteen patients received platelet transfusions and one received platelet growth factor (oprelvekin). Nine patients received RBC transfusions, and four received erythropoietin. In the rituximab control group, one patient received a RBC transfusion during follow-up.
Fig 5. Median absolute neutrophil count, platelet count, and hemoglobin concentration over time in 90Y ibritumomab tiuxetan–treated patients.
Integrated Safety: Hematologic Toxicity in 90Y Ibritumomab Tiuxetan Group
Comparison of survival curves is not yet meaningful. Twelve patients in the 90Y ibritumomab tiuxetan group died of disease progression, and seven of these 12 patients received additional NHL therapy before death. Ten patients in the rituximab group died: eight of disease progression, one of pancreatic cancer, and one of neutropenic sepsis after subsequent chemotherapy (Fig 6).
Fig 6. Overall survival analysis for 90Y ibritumomab tiuxetan and rituximab.
Median CD19+ B-lymphocyte counts in peripheral blood declined by week 4 and recovered just past month 6 in the 90Y ibritumomab tiuxetan group; counts fell by week 4 and recovered between months 9 and 12 in the rituximab control group (Fig 7). Median CD3+/CD4+ cell counts and mean/median IgM, IgA, and IgG levels remained within normal limits through month 12 in both groups. One 90Y ibritumomab tiuxetan–treated patient developed a posttreatment HAMA on day 39, and one rituximab control patient developed a posttreatment HACA on day 227; no related clinical or laboratory manifestations were noted in either patient.
Fig 7. Summary of median peripheral-blood B-cell (CD19) counts by visit and treatment group (N = 143).
One patient in each group developed secondary malignancies during follow-up. One 90Y ibritumomab tiuxetan–treated patient developed myelodysplastic syndrome on day 347; chromosome analysis showed 5q−. This patient was heavily treated with alkylating agents (cyclophosphamide, vincristine, and prednisone, and lomustine, vincristine, procarbazine, and prednisone) before 90Y ibritumomab tiuxetan treatment. A patient in the rituximab control group developed pancreatic adenocarcinoma on day 70.
DISCUSSION
Immunotherapy for B-cell NHL with the human chimeric anti-CD20 antibody, rituximab, has been an important advance in the treatment of B-cell malignancies. The unique targeting of monoclonal antibodies can be harnessed to carry radionuclides,31-39 enhancing cytotoxicity to tumor cells with relative sparing of normal body tissues. Radioimmunoconjugates such as 90Y ibritumomab tiuxetan may play an important role in the treatment of patients with B-cell NHL.
Although single-arm phase I/II studies of 90Y ibritumomab tiuxetan suggested a higher response rate for 90Y ibritumomab tiuxetan than for rituximab,15,17,18 this advantage had not yet been demonstrated in a randomized, controlled study. This is the first study reported in which an unlabeled monoclonal antibody has been compared with a radioimmunoconjugate with the same specificity for the target antigen. Results demonstrate that treatment with 90Y-labeled immunotherapy increased the ORR from 56% to 80%, produced a higher rate of CRs, and although not statistically significant, prolonged TTP in follicular and CR groups. Also, time to next therapy was longer in the 90Y ibritumomab tiuxetan–treated patients. Excellent response rates were achieved in chemoresistant patients, and patient tolerance was such that this becomes an attractive treatment strategy for elderly patients who are not acceptable candidates for chemotherapy.
The patients in this trial had incurable relapsed or refractory B-cell NHL and were in need of therapy because of B symptoms, organ dysfunction, pain, and/or rapidly progressing tumor (see Patients and Methods). In this context, a clinically meaningful outcome is a significant reduction in tumor burden (as evidenced by a PR, complete clinical response, or CR) and resolution of disease-related symptoms sustained for a period of time during which further treatment is not required. As O’Shaughnessy et al40 noted, “the primary aim of cancer treatment is prolongation of life, but the demonstration that a new agent causes tumor regression and improves patients’ clinical condition also supports approval of a new agent, even in the absence of improved survival.”
Although the ORR was significantly higher in the 90Y ibritumomab tiuxetan group compared with the rituximab control group, this did not translate into a longer TTP (Fig 2) in this study. There are several possible explanations for this finding. First, the trial was not powered to assess differences in TTP. Second, 90Y ibritumomab tiuxetan patients had a greater absolute and percent shrinkage in sum of the product of the perpendicular diameters of measurable lesions (SPD), resulting in patients with small nadir SPD that, on follow-up examinations, met the criteria for progression with minor increases in tumor size. This definition of disease progression (increase of 50% or more in SPD from nadir)41,42 may explain the more apparent difference in TTNT relative to TTP. Third, a few patients receiving rituximab who were stable but were treated eventually with other regimens were not counted as progressions. However, durable responses of 6 months’ duration were significantly higher, and at the time of this analysis, there was already a numerical benefit in durable response at 9 months and at 12 months. TTP was longer in the CR/CRu patients treated with 90Y ibritumomab tiuxetan, which suggests that future studies of radioimmunotherapy should aim to increase the CR rate.
This randomized controlled trial offers the opportunity to distinguish adverse events associated with rituximab, a naked monoclonal antibody therapy, from those specific to radioimmunotherapy with 90Y ibritumomab tiuxetan. In both groups, the majority of nonhematologic adverse events were grade 1 or 2.
The only significant differences in nonhematologic adverse events were in the digestive and respiratory body systems. In the digestive system, the primary difference consisted of a greater incidence of grades 1 and 2 nausea, vomiting, and anorexia in the 90Y ibritumomab tiuxetan group. However, the incidence of digestive system adverse events considered by the investigator to be possibly related to the study drug was not statistically different between groups. The majority of the difference in respiratory system adverse events was due to an increased incidence of cough and of dyspnea associated with anemia in the 90Y ibritumomab tiuxetan group. No long-term pulmonary toxicity has been reported.
The majority of related grade 3 and 4 toxicities associated with 90Y ibritumomab tiuxetan therapy consisted of reversible depletion of one or more hematologic cell lines, which has been noted to be correlated with percentage of bone marrow involvement with NHL18 and with the number and types of prior chemotherapies.
Patients were required to have less than 25% marrow involvement with NHL because of concern about excessive myelosuppression. This restriction eliminated many patients with small lymphocytic NHL who typically have bone marrow extensively involved with NHL. Further studies in small lymphocytic lymphoma/chronic lymphocytic leukemia are indicated, perhaps after reduction of marrow lymphoma with rituximab or chemotherapy for patients with more extensive marrow involvement.
There have been other single-arm studies of radioimmunoconjugates for NHL, including studies of iodine-131 (131I)–labeled tositumomab,39,43 131I-labeled anti-CD22,38 and 131I-labeled anti-HLA-DR.31 90Y has potential advantages for radioimmunotherapy over other 131I. As a high-energy, beta-emitting isotope, 90Y delivers higher beta energy radiation (2.3 MeV v 0.6 MeV) at a longer path length (5 mm v 1 mm).44-46 These characteristics may be especially advantageous in treating bulky, poorly vascularized tumors and tumors with heterogeneous antigen expression. In a mathematical model of tumor curability, the optimal target diameter and range for 90Y were calculated to be 3.4 cm (range, 2.8 to 4.2 cm) versus 0.34 cm (range, 0.26 to 0.5 cm) for 131I.47 Because 90Y is a pure beta emitter, 90Y ibritumomab tiuxetan can be administered as outpatient therapy without significant patient restrictions.
90Y ibritumomab tiuxetan treatment does not preclude patients from receiving subsequent chemotherapy. Patients have received subsequent treatment, with therapies such as cyclophosphamide, doxorubicin, vincristine, and prednisone; fludarabine; dexamethasone, cisplatin, and cytarabine; and etoposide, methylprednisolone, cytarabine, and cisplatin, including five patients who had high-dose treatment with stem-cell or bone marrow transplantation.48
Building on the success of single-agent radioimmunotherapy will require further clinical trials. Rituximab immunotherapy is now being combined with multiagent chemotherapy, and results indicate an improved ORR and overall survival.49 Maintenance rituximab after 90Y ibritumomab tiuxetan or repeat courses of 90Y ibritumomab tiuxetan may prolong DR. 90Y ibritumomab tiuxetan, given either before or at the end of induction chemotherapy for low- or intermediate-grade NHL, may be useful, possibly improving the CR rate and prolonging DR. Radioimmunotherapy may be combined with high-dose chemotherapy with stem-cell rescue.36 Also, further trials will need to explore retreatment and combined-modality dose and schedule.
90Y ibritumomab tiuxetan radioimmunotherapy is safe and effective and represents a clinically meaningful advance in the treatment of relapsed or refractory low-grade, follicular, or transformed NHL. The improved ORR found in this randomized study of a radiolabeled antibody versus rituximab lays the foundation for the next generation of radioimmunotherapy clinical trials.
APPENDIX
The appendix listing the investigators who participated in the trial is available online at www.jco.org.
We thank the following investigators, who participated in this trial:
- Received November 16, 2001.
- Accepted March 12, 2002.
