Transformed FL shows improved survival after CAR T-cell therapy compared to de novo LBCL, but early transformation (<12 months) defines a poor-risk subgroup: Results from a nationwide cohort

Abstract

Abstract Background Large B-cell lymphoma (LBCL) can arise from transformation of indolent lymphomas, most commonly follicular lymphoma (FL). While only 5-10% of 1st line LBCL cases are transformed FL (tFL), this increases to ~30% in the ≥3rd line CAR T-cell therapy (CART) setting, representing a clinically relevant subgroup (Vaughn, Blood Cancer Journal 2024; Spanjaart, Cancers 2023). However, registrational studies reported limited data on tFL outcomes. Prior real-world data showed superior outcomes after CART for transformed indolent lymphomas versus de novo LBCL (dnLBCL) (Thiruvengadam, Am J Hematol 2025; Stephan, Blood Adv 2025). No study has yet evaluated tFL patients treated uniformly with a single CAR T-cell product or examined outcome variation across different subgroups of tFL based on time to transformation. This study compares outcomes after axicabtagene ciloleucel (axi-cel) in tFL versus dnLBCL, with a specific focus on the impact of transformation timing. Methods All consecutive patients ≥18 years with refractory/relapsed LBCL after ≥2 prior lines who received axi-cel as standard of care between May 2020 and December 2023 at any of the 7 CART centers in The Netherlands were included. Data were obtained from the Follow-that-CAR registry and completed by chart review. tFL was defined as LBCL with a prior diagnosis of FL (sequential) or a concurrent diagnosis of LBCL and occult FL. Sequential tFL was subdivided based on time to transformation < or > 12 months after FL diagnosis. All other cases were classified as dnLBCL. No other transformed indolent lymphomas were treated during the study period. Endpoints included response, progression free survival (PFS) and overall survival (OS). CRS and ICANS were graded per ASTCT guidelines. Day 7 CAR T-cell expansion was measured by flow cytometry as % CAR+ cells in CD3+ cells in a subset of patients. Results 272 patients were included: 90 (33%) with tFL and 182 (67%) with dnLBCL. Median follow up was 29 months. Baseline characteristics were comparable (tFL vs dnLBCL: median age 61 vs 62; male 67% vs 66%; WHO PS <2 92% vs 91%). More patients with tFL had received ≥3 total prior lines of systemic therapy compared to those with dnLBCL (47% vs 12%, p<0.001). Among patients with tFL, 58% had received at least 1 prior line of therapy for indolent FL before transformation. Time since last therapy before CART was longer in tFL (>6 months: 24% vs 14%, p=0.036). Fewer tFL patients had extranodal disease (27% vs 43%, p=0.002), but ≥2 extranodal sites were equally common (both 32%). Bridging therapy use was similar (77% vs 78%). Day 7 CAR T-cell expansion was significantly higher in tFL (median 62% CAR+ cells in CD3+ cells, IQR 43–70%, n=8) than in dnLBCL (29%, IQR 12–45%, n=10), despite comparable disease burden (reflected by similar pre-lymphodepletion LDH, ferritin and CRP, bulky disease (18% vs 21%) and Ann Arbor stage III–IV (76% vs 82%), A significantly greater proportion of patients with tFL achieved a complete response (CR) compared to those with de novo LBCL (82% vs 66%, p=0.008). Median PFS was 25 months for tFL vs 6 months for dnLBCL (2-yr PFS 53% vs 37%, p=0.015). Median OS was not reached for tFL and was 17 months for dnLBCL (2-yr OS 69% vs 46%, p=0.001). There were no significant differences in CRS (92% vs 97%, with grades ≥3 in 2% and 8%) and ICANS ( 66% vs 61%, with grades ≥3 in 26% and 23%). Among tFL patients, 26 (29%) had concurrent FL/LBCL at diagnosis, while 64 (71%) had sequential transformation. Of the latter, 7 had very early transformation (<12 months from FL diagnosis). Patients with concurrent FL/LBCL or transformation <12 months had worse outcomes: 2-yr PFS 37% vs 59% (p=0.038), and 2-yr OS 55% vs 78% (p=0.011), compared to those with transformation >12 months. FL pretreatment (wait and see, radiotherapy or systemic therapy) did not affect CART response or survival. Conclusion This nationwide population-based cohort study shows superior CR rates, PFS and OS after axi-cel in patients with tFL compared to dnLBCL. Outcomes were especially favorable in patients with transformation >12 months after initial FL diagnosis. Those with concurrent transformation or transformation <12 months showed survival patterns more closely resembling dnLBCL. Higher CAR T-cell expansion in tFL despite comparable disease burden suggests biological differences, such as composition of the tumor microenvironment or target antigen expression, which may contribute to improved efficacy.