Long-Term Outcomes, Secondary Malignancies, and Stem Cell Collection Following Bendamustine in Patients with Previously Treated Indolent Non-Hodgkin Lymphoma

Background

Given the widespread use of bendamustine, data on long-term outcomes are essential for patients and clinicians to understand potential risks and benefits of therapy.  Despite a long history as treatment for indolent non-Hodgkin lymphoma (iNHL), such information has been limited.  We retrospectively reviewed the registration SDX-105-01 and SDX-105-03 trials (bendamustine 120 mg/m² days 1+2 q21 days) and SDX-105-02 trial (bendamustine 90 mg/m² days 1+2 plus rituximab 375 mg/m² day 1 q28 days) to characterize long-term toxicity and efficacy of patients treated with bendamustine.

Methods

Patient level data was retrospectively collected from patients treated on the SDX-01, 02, and 03 trials. Descriptive statistics were used to summarize patient characteristics and events. The Kaplan-Meier method was used to report time-to-event outcomes. Wilcoxon Rank sum test was used to test the difference between events for continuous variables.

Results

Out of the total 245 subjects at 45 sites, data were  available for 149 subjects (60 men, 89 women; SDX-01 N = 40, SDX-02 N = 43, SDX-03 N = 66) at 21 sites (included based on willingness to participate).  The median age was 60 years at the start of bendamustine (range 39-84). The histologies included grade 1-2 follicular lymphoma (FL; N = 73), grade 3 FL (N = 23), SLL (N = 20), marginal zone lymphoma (N = 15), mantle cell lymphoma (N = 9), transformed lymphomas (N = 5), lymphoplasmacytic lymphoma (N = 2), and not reported (N = 2). The average time from diagnosis to study entry was 41 months (range 2-229). The median number of therapies prior to bendamustine was 2.5 (range 1-8).

Patients received a median of 6 cycles and a median total dose of bendamustine of 1408 mg (max 5216, min 240). With a median follow up of 8.8 years after study entry, 80 patients had experienced progression. The median PFS was 18.4 months (95% C.I. 11.9-27.8); the 3-year PFS was 37%. During follow up, 93 patients had died at a median time of 22.3 months after the start of bendamustine. The median OS after start of bendamustine was 65.9 months (95% C.I. 38.8-91.8). The causes of death were lymphoma (N = 45), bendamustine toxicity (N = 2), subsequent treatment toxicity (N = 8), MDS/AML (N = 5), other cancer (N = 2), other (N = 6), and unknown (N = 25).

A total of 98 patients received a median of 2 therapies following bendamustine (range 1-9), with the first treatment occurring a median of 13.2 months (range 0-111.3) following the final dose of bendamustine. The reported best response to the first subsequent treatment was CR (N = 11), PR (N = 6), SD (N = 21), PD (N = 12), not evaluable (N = 25), and unknown (N = 22) and the median OS of these patients was 51.3 months (95% C.I. 33.4-80.3).

Fourteen patients had attempted stem cell collection following bendamustine, 10 of which had stem cells collected successfully. Eight patients had stem cells collected with GCSF alone (N = 7) or GCSF plus chemotherapy (N = 1).

Twenty-three patients developed 25 cancers following bendamustine. Six patients developed MDS and 2 more developed AML. The median time to MDS/AML following bendamustine was 24 months (range 10-103) with an annualized incidence rate of 0.52%/year. One of patient had a prior myeloid neoplasm and one had a prior germ cell tumor. In univariate analysis, neither age at lymphoma diagnosis (P=0.438), nor total number of systemic regimens (P=0.443), nor total dose of bendamustine (P=0.291) was associated with MDS/AML. Other cancers included adenocarcinoma (colon N = 2; prostate N = 2; lung N = 2; breast N = 1), non-melanoma skin cancer (N = 6),  squamous cell carcinoma (N = 2), hepatocellular carcinoma (N = 1), and bladder cancer (N = 1). None of these occurred in the 12 patients with a history of solid tumor before bendamustine.

Conclusions

With a median follow up of survivors of > 8 years, there was no evidence that bendamustine in the setting of previously treated iNHL was associated with a high rate of long-term bone marrow toxicity. Rates of MDS/AML and failure to collect stem cells were lower than expected. However, roughly half of all patients died within 5 years of starting bendamustine, thereby limiting the long-term follow up. A small but meaningful number of patients achieved durable remissions following bendamustine. These rigorously collected, patient-level, long-term follow up data provide reassurance that bendamustine or bendamustine plus rituximab is associated with efficacy and safety for  many patients with relapsed or refractory iNHL.