A randomized study comparing yttrium-90 ibritumomab tiuxetan (Zevalin) and high-dose BEAM chemotherapy versus BEAM alone as the conditioning regimen before autologous stem cell transplantation in patients with aggressive lymphoma
High-dose chemotherapy combined with autologous stem-cell transplantation (ASCT) is the standard therapy for refractory/relapsed aggressive lymphoma. In the era of rituximab-containing frontline regimens, it is becoming more challenging to salvage patients in this setting, and novel approaches are required. This is a randomized study evaluating the safety and efficacy of standard-dose ibritumomab tiuxetan (Zevalin) combined with high-dose BEAM chemotherapy (Z-BEAM) and ASCT in refractory/relapsed aggressive lymphoma.
Forty-three patients with CD20+-aggressive lymphoma were randomized to a treatment arm (Z-BEAM, n = 22) or control arm (BEAM alone, n = 21). Ibritumomab tiuxetan was given at 0.4 mCi/kg on day −14 before ASCT.
Patient characteristics, engraftment kinetics, and toxicity profile were similar between the 2 groups. Two-year progression-free survival (PFS) for all patients was 48% (95% confidence interval, 32%-64%): 59% and 37% after Z-BEAM and BEAM alone, respectively (P = .2). Multivariate analysis identified advanced age (hazard ratio [HR], 8.3; P = .001), high-risk disease (relapse within 12 months of diagnosis and/or secondary International Prognostic Index >2; HR, 2.8; P = .04), positive positron emission tomography-computed tomography pretransplant (HR, 2.4; P = .07), and BEAM alone (HR, 2.8; P = .03) as poor prognostic factors. Intermediate-risk patients with 1 or 2 risk factors had better PFS with Z-BEAM compared with BEAM: 69% and 29%, respectively (P = .07). Two-year overall survival was 91% and 62% after Z-BEAM and BEAM, respectively (P = .05). Similar prognostic factors determined survival. The HR for BEAM alone in the multivariate analysis was 8.1 (P = .01).
High-dose chemotherapy combined with autologous stem cell transplantation (ASCT) is the standard therapy for patients with relapsed or refractory aggressive lymphoma who are chemosensitive to second-line chemotherapy. The PARMA study, reported in 1995, is the only randomized study comparing salvage chemotherapy with salvage chemotherapy followed by ASCT in this setting. The 5-year progression-free survival (PFS) in patients randomized to ASCT in this study was 46%, compared with 12% in the nontransplant group.1 However, the PARMA trial was conducted in the prerituximab era. The use of rituximab, an anti-CD20 monoclonal antibody, in frontline chemotherapy has dramatically changed the outcome of patients with aggressive lymphoma, increasing both response and survival rates.2-4 However, despite this progress a significant proportion of patients with aggressive lymphoma are still refractory or relapse after frontline rituximab-containing therapy. Moreover, it is increasingly more difficult to rescue these patients with current salvage chemotherapy and ASCT approaches.
The recently reported Collaborative Trial in Relapsed Aggressive Lymphoma (CORAL) identified prior treatment with rituximab as an adverse prognostic factor in relapsed/refractory aggressive lymphoma, especially in those relapsing within 12 months of diagnosis.5 The 3-year event-free survival was 21% in patients previously treated with rituximab, compared with 47% in rituximab-naive patients. Novel approaches are in need to improve outcome in these patients. Theoretically, these can include better salvage regimens, better conditioning regimens for ASCT, and/or maintenance treatment after ASCT. Rituximab-containing salvage regimens have been shown to be more effective, but mostly in rituximab-naive patients.5-8 Rituximab maintenance or consolidation after SCT to target minimal residual disease is another appealing approach.9-11 The results of the second randomization study in the CORAL trial are awaited to confirm this strategy.
No single preparative regimen before ASCT has been shown to have an advantage in disease eradication over the others. Lymphomas are inherently radiosensitive malignancies. Total body irradiation (TBI) has been used as part of conditioning regimens for ASCT. However, there are concerns about toxicity and long-term effects associated with TBI, and there are no data to support TBI-containing regimens over regimens containing high-dose chemotherapy alone. Radiolabeled antibodies such as yttrium-90 ibritumomab tiuxetan (Zevalin) and iodine-131 tositumomab (Bexxar) are increasingly used in indolent lymphomas, but are also effective in aggressive lymphomas.12 Several studies used radioimmunotherapy in combination with high-dose chemotherapy and ASCT with promising results (reviewed in Shimoni and Nagler13 and Gisselbrecht et al14).
In this randomized multicenter study, we evaluated the safety and efficacy of standard-dose ibritumomab tiuxetan combined with high-dose BEAM chemotherapy (Z-BEAM) and ASCT in patients with relapsed/refractory aggressive non-Hodgkin lymphoma in the era of rituximab-containing frontline therapy.
MATERIALS AND METHODS
The current study included patients with CD20+-aggressive non-Hodgkin lymphoma including diffuse large cell lymphoma or transformed follicular lymphoma. Patients were eligible if they failed to achieve complete remission (CR) with initial chemotherapy or they relapsed after CR but were chemosensitive to second-line chemotherapy. Fluorine-18-fluorodeoxyglucose positron emission tomography combined with computerized tomography (PET-CT) was performed before ASCT. Chemosensitivity was defined as having at least 50% reduction in the bidimensional measurements of the largest mass. Active central nervous system disease at the time of ASCT was not allowed. Unlike standard criteria for ibritumomab tiuxetan treatment, bone marrow involvement and pancytopenia did not exclude treatment, as patients were given autologous stem cells after completion of the treatment plan. A maximum of 2 prior chemotherapy regimens were allowed. Secondary International Prognostic Index (IPI) was calculated at disease relapse by standard criteria. In patients with primary refractory disease, IPI was calculated at the start of second-line chemotherapy. Patients were considered at high risk for relapse after ASCT if they had disease progression within 12 months of initial diagnosis or if they had secondary IPI >2. Patients were required to be free of marked organ dysfunction and to have an Eastern Cooperative Oncology Group performance score of 0 or 1, by standard institutional eligibility criteria for ASCT. Patients were required to have an autologous granulocyte colony-stimulating factor mobilized peripheral blood stem cell collection of at least 2 × 106 CD34+/kg. All patients gave written informed consent, and the study was approved by the institutional review board. The study was registered at ClinicalTrial.gov as NCT00491491.
This study was a prospective multicenter randomized study designed to compare 2 conditioning regimens given before ASCT for chemosensitive aggressive lymphoma, Z-BEAM and BEAM alone. Therefore, patients were registered and randomized only after completion of second-line chemotherapy and stem cell collection and within 2 weeks of admission for ASCT. On the basis of this design, all patients were given their assigned therapy, and there was no dropout after randomization. Prior therapy was determined by the referring physician and was usually R-CHOP for first-line therapy and platinum-based second-line therapy. Patients randomized to Z-BEAM were given rituximab 250 mg/m2 followed by ibritumomab tiuxetan 0.4 mCi/kg (capped at 32 mCi) on day −14 before ASCT. Dosimetry was not used, and there were no dose adjustments based on blood counts or other criteria. Both groups were given high-dose BEAM chemotherapy, which started on day −6 and included carmustine 300 mg/m2 on day −6, etoposide 200 mg/m2 daily, cytarabine 200 mg/m2 every 12 hours on days −5 to −2, and melphalan 140 mg/m2 on day −1. Chemotherapy doses were calculated by adjusted body weight. Autologous stem cells were infused on day 0. Filgrastim 5 μg/kg was administered from day +4 until engraftment. Standard institutional protocols were used for transfusions and antibiotic therapy for neutropenic fever. Valacyclovir was given for 1 month for prevention of herpes simplex reactivation, and trimetoprim-sulfamethoxasole was given for 6 months for prevention of Pneumocystis carinii pneumonia.
Evaluation of Response and Follow-up
Neutrophil and platelet engraftment were defined as the first of 3 days with absolute neutrophil count (ANC) >0.5 × 109/L and the first of 7 days with an untransfused platelet count >20 × 109/L, respectively. Toxicity after ASCT was graded by the National Cancer Institute (Bethesda, Md) Common Toxicity Criteria. Disease response was evaluated by Cheson revised response criteria.15 CR was defined as the complete disappearance of all PET-positive masses and clearance of marrow when applicable. Patients were assessed for disease involvement by PET-CT and bone marrow biopsy (in patients with a history of marrow involvement) before ASCT, 1 and 3 months after ASCT, and then at 3- to 6-month intervals for 24 months after ASCT or as clinically indicated. Patients relapsing after ASCT were usually given chemotherapy selected at the discretion of the center and attending physician. Responding patients with adequate medical status were eligible for allogeneic transplantation.
The primary endpoint in this study was PFS at 2 years after ASCT. Relapse and progression were defined according to Cheson's revised response criteria as the appearance of any new lesion or increase by >50% of previously involved sites from nadir.15 Lesions were required to be PET positive to define progression. PFS was calculated from the day of ASCT until disease progression, death, or last follow-up. Overall survival (OS) was calculated from the day of ASCT until death or last follow-up. It was calculated that to detect an expected increase of 20% in PFS from 50% to 70%, with a 5% significance level, 100 patients would need to be randomized for 1 of the treatment groups, over 3 years. The study was stopped after enrollment of 43 patients because of slow accrual, as will be further discussed. Randomization was stratified by the participating center, but not by other prognostic factors because of the expected accrual number. Numerical variables between groups were compared with a t test and categorical variables with chi-square test. The probabilities of OS and PFS were estimated using the Kaplan-Meier method.16 The effect of various patient and disease categorical variables on survival probabilities was studied with log-rank test. A multivariate Cox proportional hazard model with hierarchical forward entering was constructed to assess prognostic factors.17 Survival and hazard ratio (HR) probabilities were presented with 95% confidence intervals. All reported P values are 2-sided, and P < .05 was considered significant.
Patient and Disease Characteristics
Forty-three patients were enrolled on the study. Patient and disease characteristics are outlined in Table 1. The median age was 55 years (range, 23-67 years). Disease histology was diffuse large B-cell lymphoma (n = 31), mediastinal lymphoma (n = 3), or transformed follicular lymphoma (n = 9). Forty-one patients had chemosensitive disease to second-line therapy after failure to achieve CR with first-line therapy (n = 5) or after disease relapse from a first remission (n = 36). Two patients with transformed follicular lymphoma were transplanted in first remission after only 1 line of therapy. In all but 1 patient, the first line of therapy included rituximab. PET-CT before ASCT was positive in 17 patients (40%). There were no significant differences in patient characteristics between the Z-BEAM and BEAM groups.
Table 1. Patient Characteristics
All Patients, n=43
Abbreviations: DLBCL, diffuse large B-cell lymphoma; FL, follicular lymphoma; NS, not significant; PET-CT, positron emission tomography combined with computerized tomography; SCT, stem cell transplantation; sIPI, secondary International Prognostic Index at relapse; Z-BEAM, standard-dose ibritumomab tiuxetan combined with high-dose BEAM.
Refractory to first-line therapy, but responsive to second-line therapy.
All patients had chemosensitive disease after 2 lines of therapy; 2 patients with transformed FL were transplanted in first remission after 1 line of therapy.
Patients refractory to first-line therapy were included with patients with no progression in the first 12 months from diagnosis.
In patients with primary refractory disease, IPI was determined at the start of second-line chemotherapy.
Patients having progression within 12 months from diagnosis and/or sIPI >2 were considered high risk; all others were low risk.
All 43 patients achieved primary engraftment. The median time to ANC 0.5 × 109/L was 10 days (range, 7-21 days). The median time to platelet count 20 × 109/L was 13 days (range, 8-32 days). There was no difference in engraftment kinetics between the 2 treatment groups. There were no infusion reactions associated with ibritumomab tiuxetan. The toxicity profile is outline in Table 2. Mucositis requiring narcotics was common in both groups, with a trend for a higher incidence after Z-BEAM. Other grade III toxicities were relatively rare and transient. Neutropenic fever was common and not different in both groups; however, there was a higher incidence of more serious infections, such as pneumonia or fungal infections, after Z-BEAM. All patients recovered from the early transplant-related toxicities, and there were no deaths within the first 100 days from any cause.
Table 2. Engraftment and Early Toxicity
Abbreviations: ANC, absolute neutrophil count; NS, not significant; PLT, platelet count; Z-BEAM, standard-dose ibritumomab tiuxetan combined with high-dose BEAM.
At the time of first evaluation, approximately 1 month after ASCT, 42 patients were in CR with negative PET-CT, including 16 of the 17 patients with positive pretransplant PET-CT. With a median follow-up of 29 months (range, 12-50 months), 20 patients relapsed, and 1 died of a nonrelapse cause. The actuarial 2-year PFS is 48% (95% confidence interval [CI], 32%-64%). No difference in relapse patterns could be detected within this limited number of events. Table 3 outlines the univariate and multivariate analysis of factors predicting for PFS. There was a statistically nonsignificant trend in the univariate analysis for better PFS after Z-BEAM compared with BEAM alone: 59% and 37%, respectively (P = .2, Fig. 1, Top). The multivariate Cox proportional hazard model identified advanced age (≥55 years), high-risk disease (defined as relapse within 12 months of diagnosis and/or secondary IPI >2), positive PET-CT before ASCT (borderline), and treatment with BEAM alone as adverse prognostic factors, with HRs of 8.3, 2.8, 2.4, and 2.8, respectively (Table 3). Although based on small numbers in each subset, it seems that patients with none of the first 3 risk-factors had excellent outcome regardless of treatment group (5 of 5 alive and progression-free, Fig. 1, Middle). Conversely, patients with all 3 risk-factors had poor outcome (all 6 patients progressed). However, among patients with 1 or 2 risk factors (32 patients, 16 in each group), conditioning with Z-BEAM was associated with better outcome: 2-year PFS of 69% (95% CI, 46%-91%) and 29% (95% CI, 4%-55%), respectively (Fig. 1, Bottom; P = .07).
Table 3. Progression-Free Survival According to Prognostic Factors
At the time of writing, 32 patients were alive and 11 had died (10 of disease relapse, and 1 patient in the Z-BEAM group died 8 months after ASCT of JC virus encephalitis). The actuarial 2-year survival rate was 77% (95% CI, 63%-90%). OS was 91% (95% CI, 79%-100%) and 62% (95% CI, 39%-85%) after Z-BEAM and BEAM alone, respectively (Fig. 2, P = .05). The better survival after Z-BEAM was partly related to better salvage responses after relapse. Among relapsing patients, the 1-year OS after relapse was 87% and 47% after Z-BEAM and BEAM, respectively (P = .07). All patients were given salvage chemotherapy for relapse. Five of 8 patients relapsing after Z-BEAM ultimately had allogeneic transplant; 1 died early, and 4 became long-term survivors, although 2 later relapsed. Three of 12 patients after BEAM alone were given allogeneic transplants; 2 died early and 1 became a long-term survivor (P = not significant). Multivariate analysis identified high-risk disease and treatment with BEAM alone as adverse prognostic factors, with HRs of 7.8 and 8.1, respectively (Table 4). Advanced age and positive PET-CT before SCT had borderline significance. Similarly to the analysis of PFS, patients with 1 or 2 disease risk factors benefited more from Z-BEAM. Two-year OS in these patients was 100% and 63% after Z-BEAM and BEAM, respectively (P = .008).
Table 4. Overall Survival According to Prognostic Factors
Two patients, both in the Z-BEAM group, had protracted poor graft function after ASCT. Neither had marrow involvement at the time of ibritumomab tiuxetan administration or a poor graft that would predict this outcome. Both later relapsed and were salvaged with allogeneic SCT. None of the patients developed myelodysplastic syndrome or a secondary malignancy.
This is the first reported randomized study to evaluate the safety and efficacy of the combination of radioimmunotherapy with high-dose chemotherapy in patients with relapsed/refractory aggressive lymphoma. It suggests that the addition of ibritumomab tiuxetan to BEAM high-dose chemotherapy is safe and possibly more effective in disease eradication. The 2-year PFS was 59% and 30% in the Z-BEAM and BEAM arms, respectively (P = .20). The 2-year OS was 91% and 62%, respectively (P = .05). There was no significant added toxicity with the Z-BEAM regimen.
There are 3 potential approaches for the use of radioimmunotherapy in ASCT. The radioconjugate can be administered in the standard treatment dose combined with high-dose chemotherapy, as in this study. Krishnan et al reported the use of Z-BEAM in 60 patients with various types of lymphoma.18 The 2-year PFS in the 20 patients with diffuse large B-cell lymphoma was 68%, compared with 48% in historical controls. In a following report, they compared 46 patients with aggressive lymphoma given Z-BEAM to a matched control cohort given TBI-based conditioning.19 Z-BEAM had a similar relapse incidence to TBI with lower toxicity, resulting in improved survival. The 4-year survival was 81% and 53%, respectively. This regimen was also used safely in patients with low-grade lymphoma.20
The second approach is to escalate the radioconjugate dose beyond the standard dose. Winter et al used an individualized ibritumomab tiuxetan dose calculated to deliver a maximum of 15 grays to critical organs.21 The 4-year PFS in a group of 44 patients, 30% with less than a partial response to prior chemotherapy, was 43%. This approach requires accurate dosimetry to calculate a dose that will not deliver excess radiation to critical organs. The third approach is to use high doses of the radioconjugate alone, followed by stem cell support with no additional high-dose chemotherapy.22, 23 This approach was used mostly in patients considered not eligible for standard ASCT. Devizzi et al used ibritumomab tiuxetan at 0.8 to 1.2 mCi/kg, 2 to 3× the standard dose, in 30 patients.23 The regimen was safe, with promising results: a 2-year PFS of 69%. Similar approaches have been used with iodine-131 tositumomab (Bexxar) as the radioconjugate, with similar outcomes.13, 14, 24 A large Blood and Marrow Transplant Clinical Trials Network randomized study comparing iodine-131 tositumomab-BEAM and BEAM has completed accrual, and results are eagerly awaited.
The first approach was used in the current study because there is no evidence to support 1 approach over the others, and we were seeking a regimen that will be widely applicable. Ibritumomab tiuxetan dosing was used without prior imaging, as is the common practice in Europe. Ibritumomab tiuxetan in standard dose is dosed based on body weight alone. Dosimetric studies demonstrated that in this dose there is minimal exposure to nontargeted tissues. No significant correlation was found between biodistribution of the radionuclide and toxicity. Therefore, ibritumomab tiuxetan was approved in the European Union without the necessity for imaging studies. The use of a fixed standard dose is easier, does not require specific expertise, allows ambulatory treatment, and limits the costs related to dosimetry and escalated doses of the radioconjugate, which may become prohibitive. Ibritumomab tiuxetan, being a pure beta emitter, is not associated with radiation risk to the patient family and health care providers, and can routinely be given to ambulatory patients.
This randomized study confirmed the previously reported observations that the addition of ibritumomab tiuxetan to high-dose chemotherapy is safe and not associated with excess toxicity. Although there was a trend for more mucositis and more serious infections in the Z-BEAM arm, there was no excess in critical organ toxicities, and all toxicities were reversible, with no early deaths. The cause of increased rate of infections is not clear, but they did not result in any mortality or irreversible morbidity. Engraftment was not significantly affected by ibritumomab tiuxetan, as the residual radiation dose at the time of stem cell infusion, 14 days after treatment, or >5 half-lives of the agent, was reported by other groups to be very low.21, 23 Two patients in the Z-BEAM group had protracted poor graft function. Theoretically, targeting radiation to the marrow in patients with active marrow involvement at ASCT may cause stromal damage. However, none of these patients had such involvement, and other causes may have contributed to this finding. Long-term safety requires longer follow-up. Several group reported few patients with myelodysplastic syndrome, but this can be related to ASCT itself. The risk of myelodysplastic syndrome after ibritumomab tiuxetan therapy was reported to be low,25 and Devizzi et al reported no cytogenetic abnormalities after high-dose ibritumomab tiuxetan and ASCT.23
PFS after ASCT is highly dependant on several prognostic factors. This makes the results of different studies difficult to compare and strengthens the need for randomized studies. Chemosensitivity to the salvage regimen is the most important factor for the success of ASCT, such that patients who do not have at least a partial response to second-line chemotherapy are not generally accepted as ASCT candidates.26 Z-BEAM may be able to partially overcome the poor prognosis of chemorefractoriness.21, 27 When the secondary IPI was retrospectively applied to patients included in the PARMA study, it could distinguish patients with different survival probabilities.28 Hamlin et al applied the secondary age-adjusted IPI to 150 patients treated for relapsed/refractory aggressive lymphoma.29 Among the patients with chemosensitive disease (ASCT candidates), 4-year PFS was 69%, 46%, and 25% for patients with low, intermediate, and high secondary IPI, respectively. Several studies have shown that positive fluorodeoxyglucose (FDG)-PET before ASCT is associated with a poor outcome. Derenzini et al showed a 3-year PFS of 87% and 35% in patients with negative and positive FDG-PET scans before ASCT.30 These observations were recently confirmed in a meta-analysis.31 Schot et al combined FDG-PET with the clinical risk scores. Both were independent predicting factors, and the combined score separated patients into 4 groups with PFS of 5% to 100%.32 Relapse within 12 months from diagnosis has been shown in the PARMA trial as a very poor prognostic factor.33 Similarly, the CORAL study identified early relapse, secondary IPI >1, and prior rituximab therapy as poor prognostic factors in relapsed lymphoma.5 Three major risk factors were identified in the current study. Advanced age was a dominant poor prognostic factor. Disease risk defined as relapse within 12 months of diagnosis and/or secondary IPI >2 was another significant risk factor. These 2 known risk factors were combined to reduce the number of subsets in this limited patient group. Positive PET-CT was a third factor. The effect of prior rituximab could not be analyzed, as all but 1 patient had rituximab with frontline therapy. These risk factors were combined, similarly to the approach of Schot et al, showing a wide separation in prognosis. Patients with none of these factors had an excellent outcome, whereas patients with all 3 factors had a dismal outcome regardless of the regimen used. The intermediate group with 1 or 2 risk factors had a PFS of 48%. In this group, Z-BEAM was shown to have an advantage over BEAM alone, with 2-year PFS of 69% and 29%, respectively (P = .07) and 2-year OS of 100% and 63%, respectively (P = .008). It is possible that higher than standard doses of the radioconjugate will benefit higher-risk patients, but this has not yet been tested. The current observations must be considered with caution, as they are based on small patient subsets. We cannot rule out that subtle differences in the risk profile between the groups that were not statistically significant because of patient numbers contributed to the difference in survival. However, these observations can serve as hypothesis generating for further studies. Results of any future clinical study should be interpreted in light of these prognostic factors.
The major limitation of this study is the low patient number, resulting from early closure of the study because of low accrual. Rituximab in frontline chemotherapy leads to selection of a poor-risk population at relapse that is becoming a major challenge. Other studies have shown a lower response rate at relapse for patients with prior rituximab therapy. The study was designed to compare 2 conditioning regimens; therefore, patients were randomized only after documented response to second-line therapy. We do not have data on response rate before randomization, but based on the above discussion, we can expect that improved outcome after first-line therapy and low response rate after relapse led to the low accrual. It seems that even for responding patients, results are currently inferior to those of patients in the prerituximab era, as PFS in the control group was only 37%. This emphasizes the need for better regimens. Larger, international studies with longer follow-up will be needed to confirm these observations before the new Z-BEAM regimen can be accepted as a standard of care for ASCT in the era of rituximab-containing front-line and second-line chemotherapy.