Inferior Survival in Double Refractory Large B-Cell Lymphoma Eligible for Third-Line CD19 Chimeric Antigen T-Cell Therapy

Background: Outcomes following CD19 chimeric antigen T-cell (CAR-T) therapy for patients (pts) with large B-cell lymphoma (LBCL) refractory (rf) to both an anthracycline and platinum-based therapy, referred to as double refractory (DR), are not well described. It is also unclear if these patients may be less likely to proceed to CAR-T infusion. We reviewed the outcomes of pts referred for CAR-T at our centre. Methods: Retrospective review of the CAR-T database at the Princess Margaret Cancer Centre from April 2020 - December 2023. Anthracycline-refractoriness was defined as pts progressing/stable disease (PD/SD) on therapy or relapsing 3 months of completing treatment (typically R-CHOP). Platinum-refractoriness was defined as PD/SD while receiving platinum-based therapy (typically R-GDP). Inclusion criteria: adults 18 years from Ontario, Canada with LBCL referred for consideration of 3-line standard of care CD19 CAR-T therapy. Outcomes included progression free survival (PFS) defined from date of cell infusion to PD/death. Overall survival (OS) defined from date of initial CAR-T consultation to death. DR cohort outcomes were compared against non-double refractory (NDR) pts (defined as pts not relapsing 3 months or no PD/SD on anthracycline-based and no PD/SD on platinum-based treatment. Kaplan-Meier estimator was used to generate OS curves. Turnbull estimator was used to generate PFS curves. Results: A total of 174 pts were identified that met inclusion criteria (59: DR, 115: NDR). There were no difference in baseline characteristics (median age, sex, ECOG, stage, disease bulk, presence of MYC and BCL2/BCL6 translocation, extra-nodal involvement including CNS disease) between the 2 cohorts. Median age 61-years (range 20-83), 63% male, median prior lines 2 (range 1-3) for LBCL and 29% had an intake ECOG of 2. Histologic subtypes: de novo diffuse large B-cell lymphoma (DLBCL) (67%), transformed DLBCL (28%), primary mediastinal B-cell lymphoma (5%). MYC and BCL2/BCL6 translocations were detected by FISH in 41/114 pts tested (35.9%). The majority had stage IV disease (53%) with extra-nodal involvement in 58%, a history of central nervous system (CNS) involvement in 6% and bulky disease (7cm) in 43%. While no patients in the DR cohort had a prior autologous stem cell transplantation (ASCT), 43% of the NDR cohort had a prior ASCT. Bridging therapy (BT) was administered in 71% (n=122): radiotherapy (RT) (23%, n=39), chemotherapy (CT) (21%, n=36), RT and steroids (ST) (9%, n=16), ST only (9%, n=15), CT and ST (4%, n=7), CT, RT and ST (3%, n=5), CT and RT (2%, n=4), p=0.21. CAR-T product: axicabtagene ciloleucel 76%, tisagenlecleucel 24%. Of all 174 pts, 36 (20.6%) were unable to proceed with the intended CAR-T infusion due to ineligibility (poor ECOG or PD) (48%), pt decline (22%), active CNS disease (12%), manufacturing failure (8%) or death (6%) missing data 4%. The failure rate to proceed with CAR-T infusion in the DR and NDR cohorts were 30.5% versus 15.6%, p=0.036. Median follow-up for all pts was 17.5 months (range 14.7-21.1). In all pts the 12-month OS was 57% (95% CI 49-66.2%); DR 41.7% (95% CI 29.9-58%) versus NDR 65.2% (95% CI 55.6-76.6%), p=0.0046. In pts proceeding to CAR-T infusion, 12-month OS was 66.7% (95% CI 58-76.6); DR 57.2% (95% CI 42.9-76.4%) versus NDR 71% (95% CI 60.8-82.9%), p=0.23. In pts that proceeded to CAR-T infusion, 6-month PFS was 49.8% (95% CI 28.7-86.2%); DR 50.4% (95% CI 24.8-100%) versus NDR 51.7% (95% CI 29.2-91.8%), p=0.43. In pts that did not proceed to CAR-T infusion (DR: 18, NDR: 18) median OS was 2.56 months (95% CI 1.94-3.42); DR 1.94 months (95% CI 1.61-3.35) versus NDR 3.42 months (95% CI 2.56-NR), p=0.032. Univariable and multivariable analysis found DR status a predictor of inferior OS (adjusted p=0.034). Conclusion: In our analysis, DR pts had inferior OS to NDR pts, driven by higher failure rates to proceed with CAR-T infusion. The OS of DR and NDR pts proceeding to CAR-T infusion appears similar but may be influenced by sample size. Identifying DR status at time of referral as a determinate of outcome would be clinically beneficial. Approaches to bridge DR pts to CAR-T therapy and assessing outcomes of DR pts (those progressing on platinum-based BT) in second line CAR-T therapy need to be prospectively investigated.