Analysis of real-world overall survival (rwOS) among patients with third-line or later (3L+) diffuse large B-cell lymphoma (DLBCL) with multiple qualifying index dates treated in the United States

Abstract Introduction: Standard of care (SOC) first line (1L) therapy for diffuse large B-cell lymphoma (DLBCL) includes rituximab, cyclophosphamide, doxorubicin, vincristine, and prednisone (R-CHOP) and polatuzumab vedotin plus R-CHP (Pola-R-CHP), among other regimens. For eligible patients who experience relapsed/refractory (R/R) disease, chimeric antigen receptor T-cell (CAR-T) therapies or high-dose chemotherapy with hematopoietic stem cell transplant (HSCT) are standard treatment options. For patients ineligible for CAR-T or HSCT or those who progress to 3L therapy, many therapeutic options exist; however, there is no single SOC. Given the rapidly evolving treatment landscape, including options like CAR-T, challenges exist in estimating real-world overall survival (rwOS) for patients receiving standard care in recent years. Our study contributes the latest data on rwOS for patients receiving 3L+ therapy for R/R DLBCL who are not eligible for HSCT or CAR-T at index line of therapy (LOT). Methods: Using the COTA electronic health record-based database and inclusion and exclusion criteria of the ECHELON-3 trial, eligible adult patients were identified who had received at least 2 prior LOTs. Qualifying patient-LOTs were identified based on the following criteria: initiation of 3L+ between OCT/18/2017 and DEC/31/2023, no documentation of an Eastern Cooperative Oncology Group (ECOG) performance status score of 3 or greater within the baseline period, no evidence of solid tumor malignancy within two years of index LOT, no exclusionary treatment history, and sufficiently precise (i.e., at least month and year precision) key study dates. Multiple index date qualification methodology was used to identify all qualifying LOTs (multiple per patient), first qualifying LOT (one per patient), and last qualifying LOT (one per patient), and characteristics and outcomes were assessed by each cohort. The baseline period was defined as the earlier of: 30 days prior to index or the end of prior LOT not exceeding 90 days until 7 days post-index. rwOS was evaluated using the Kaplan-Meier method. COTA's mortality variable includes dates of death abstracted from the health record and is supplemented with dates from third-party obituary data. The resulting composite variable has been validated in prior publications. Results: In the COTA database, 476 unique patients qualified for the study with a total of 716 qualifying patient-LOTs. The median age at diagnosis and first qualifying index was 64 (IQR: 55, 73) and 67 years (IQR: 57, 76), respectively, and the patients were majority male (59%), White (69%), and treated in the community setting (66%). Of 357 patients with known cell of origin, 176 (49%) had GCB and 181 (51%) had non-GCB DLBCL at first qualifying index LOT. Median follow-up time (reverse KM) among all qualifying LOTs (N=716) was 26.7 months (mos) (95% CI: 23.7, 31.9), with 480 death events (299 unique patient death events). Median rwOS across all qualifying LOTs was 8.0 mos (95% CI: 7.1, 10.1). Median rwOS among unique patients from first qualifying LOT and last qualifying LOT was 11.4 (95% CI: 9.7, 13.6) and 6.3 mos (95% CI: 5.3, 8.0), respectively. For patients with documented ECOG score 0-2 (n=408 qualifying-LOTs, n=308 with missing/unknown ECOG score), the median rwOS was 6.9 mos (95% CI: 6.0, 8.5). Among those who had received CAR-T therapy (n=72 qualifying-LOTs) or SCT (n=98 qualifying-LOTs) prior to index (which was used as a proxy for ineligibility at index), median rwOS was 7.2 (95% CI: 5.5, 13.3) and 8.0 mos (95% CI: 5.9, 13.6), respectively. Among all qualifying-LOTs with GCB (n=252) and non-GCB (n=278) DLBCL, median rwOS was 8.1 (95% CI:7.0-10.8,) and 7.3 mos (95% CI: 6.4, 10.6), respectively. Conclusions: Patients in this analysis with 3L+ DLBCL had a median age of 67 years at first index and experienced a median rwOS of 8 mos from all qualifying index dates. This study demonstrates the feasibility of using real-world data to contextualize the results reported from a clinical trial and highlights the need for novel treatment options for patients with R/R DLBCL receiving 3L+ therapy who may be ineligible for HSCT and CAR-T therapy. Results of the study also suggest that considering only the first qualifying index date may result in overestimation of outcomes, while considering only the last qualifying index date may underestimate study outcomes.