Unmet Medical Need Among Elderly Patients with Previously Untreated DLBCL Characterized Using Real-World Data in the United States

Introduction: Diffuse large B-cell lymphoma (DLBCL) is a significant source of cancer morbidity and mortality. More than half of all newly diagnosed patients are older than 65 years, among whom the 5-year relative survival rate is 54% (SEER 2020). Prior research has shown that not all elderly patients (≥80 years old) receive R-CHOP or mini-R-CHOP regimens as the first line of therapy (LoT); those who do not may have suboptimal outcomes (Williams, et al. Cancer 2015; Hamlin, et al. Oncologist 2014; Juul, et al. Eur J Cancer 2018). This study leverages two real-world data (RWD) sources, Flatiron Health (FH) electronic health record-derived de-identified database and SEER-Medicare (SEER-M) to characterize elderly patients with DLBCL (including observed treatment patterns), summarize overall survival (OS) outcomes, and identify unmet medical needs in this population.

Methods: RWD from FH included patients with a DLBCL diagnosis on or after January 1, 2011, with follow-up until May 31, 2020. The SEER-M database is a linkage of two population-based RWD sources: the SEER Cancer Registry and Medicare claims database. RWD from SEER-M for this study included patients with a DLBCL diagnosis between January 1, 2011 and December 31, 2015, with follow-up until December 31, 2016. All fee-for-service Medicare enrollees in SEER-M had to have complete claims. RWD for basic demographics, treatments and outcomes were analyzed from both datasets; FH’s database included data on certain clinical characteristics including granularity for dosing data when available, in comparison to SEER-M. This descriptive analysis included patients who were aged ≥80 years at diagnosis. Among patients in the FH database who received R-CHOP as first LoT and had available dosing data, those who received <80% of standard full doses for cyclophosphamide (750mg/m²) and doxorubicin (50mg/m²) at first administration were classified as “reduced-dose” R-CHOP. OS data were summarized using an unadjusted Kaplan-Meier survival function and 95% confidence intervals (CI).

Results: The study included 725 patients from the FH database and 2613 patients from the SEER-M database; patient characteristics and outcomes were generally consistent between the two datasets. In total, 16% and 35% of the elderly patients had no record of systemic treatment in FH and SEER-M respectively (Table). More than half of the treated patients received R-CHOP in the first LoT (63% and 53% in FH and SEER-M, respectively); other patients received attenuated regimens, including rituximab plus bendamustine (R-Benda), rituximab plus cyclophosphamide, vincristine and prednisolone (R-CVP), and rituximab (R) monotherapy. Patients who received R-CHOP in the first LoT had numerically longer median OS (Flatiron: 55.0 months [95% CI: 41.8–NA]; SEER-M: 50.7 months [95% CI: 45.9–62.9]) compared with those who received other regimens (Figure A, B). Untreated patients had a median survival of 3.1 months (95% CI: 2.3–5.2) in the Flatiron dataset and 2.0 months (95% CI: 1.8–2.2) in the SEER-M dataset. Among those who received R-CHOP and with available dosing data, 51% received reduced-dose R-CHOP in the first LoT and OS appeared shorter than for patients who received full-dose R-CHOP (Figure C).

Conclusions: Despite differences between the databases, RWD in FH and SEER-M both demonstrate considerable variation in the regimens received by elderly DLBCL patients, with 16–35% receiving no treatment and >50% receiving attenuated regimens including reduced-dose R-CHOP. Patients receiving regimens other than R-CHOP had a numerically lower survival probability compared with the standard of care (SoC) R-CHOP/reduced-dose R-CHOP. These data show a high unmet medical need among elderly patients with DLBCL who may not be able to tolerate immunochemotherapy regimens that have been evaluated in trials for a carefully selected patient population. Further research will aim to assess prognostic factors at the time of treatment initiation, as well as gather information on comorbidities and other factors that may prevent elderly patients from receiving SoC R-CHOP; these patients may be candidates for better-tolerated novel approaches.