Clinical and pathological features of patients with large B-cell lymphoma relapsing after CAR T-cell therapy

Abstract Introduction: Chimeric antigen receptor (CAR) T-cell therapy has become a standard treatment for relapsed/refractory large B-cell lymphoma (LBCL), and it is considered a potentially curative option. However, the prognosis of patients relapsing after CAR T-cell therapy is extremely poor. Studies investigating the association between pathological antigen expression profiles and clinical treatment outcomes in bispecific antibodies (BsAbs) era at the time of relapse after CAR T-cell therapy are limited. Methods: We conducted a retrospective analysis of patients with LBCL who received CAR T-cell therapy at the National Cancer Center Hospital, Tokyo, Japan between December 2020 and July 2025 and subsequently experienced disease progression. A total of 146 patients underwent CAR T-cell therapy: 50 received axicabtagene ciloleucel (axi-cel), 75 lisocabtagene maraleucel (liso-cel), and 21 tisagen lecleucel (tisa-cel). Among them, 48 patients (16 with axi-cel, 20 with liso-cel, and 12 with tisa-cel) developed progressive disease. We evaluated patient characteristics, subsequent treatments, prognosis and pathological findings which compared the expression of CD19 and CD20 before and after relapse. Results: Among the 48 patients, the median age was 66 years (range, 20–77), and 23 (48%) were female. 15 patients (31%) received CAR T-cell therapy as second-line therapy. The median number of prior treatment lines was 2 (range,1–8). Histologic subtypes included diffuse large B-cell lymphoma, not otherwise specified in 27 (56%), transformed follicular lymphoma in 8 (17%), high-grade B-cell lymphoma, in 12 (25%), and primary mediastinal large B-cell lymphoma in 1 (2%). At the time of progression, the median lactate dehydrogenase level was 267.5 IU/L (range, 173–3371), and the median metabolic tumor volume was 44.78 mL (range, 0-2967). Disease progression occurred within 3 months of infusion in 25 patients (52%), between 3–6 months in 13 (27%), and at or beyond 6 months in 10 (21%). The median time to progression after CAR T-cell infusion was 2.8 months (range, 0.1–21.2). Thirty-eight of the 48 patients (79%) received post-relapse systemic therapy that included BsAb in 19, conventional systemic therapy in 18 (rituximab and lenalidomide n=5, Polatuzumab vedotin, bendamustine and rituximab n=7, other cytotoxic chemotherapy n=6) and one small molecule. Seven patients (15%) underwent allogeneic hematopoietic stem cell transplantation. The median overall survival (OS) who received systemic therapy was 9.8 months. Before the approval of BsAb, 14 of 23 patients (61%) received conventional systemic therapies. Following the approval, 17 of 25 patients (68%) received BsAb therapy, while only 4 patients (16%) received conventional systemic therapies. Among those who received conventional systemic therapy post-BsAb-approval, loss of CD20 expression was the primary reason for treatment selection. The 1-year OS rate was 71% for patients treated with BsAb compared to 23% in those who received conventional systemic therapies (Hazard ratio HR=6.195, p=0.0062). When stratified by timing of relapse, the median OS was 10.1 months for patients who relapsed within 3 months of CAR T-cell infusion (n=25), 15.4 months for those who relapsed between 3-6 months (n=13), and not reached for those relapsing after 6 months (n=10). Among the 10 patients who did not receive post-relapse systemic therapy, 8 experienced relapse within 3 months of infusion. The primary reasons for not receiving further treatment were rapidly progressive disease (n=5), cytopenia (n=3), ongoing infection (n=1) and patient refusal (n=1). Repeat biopsies were performed in 39 of 48 patients. Loss of CD19 was observed in 6 patients (15%), and loss of CD20 in 14 (36%), including 3 patients who lost both antigens. The median OS was 6.0 months in CD20-negative patients and 12.4 months in CD20-positive patients (HR=3.344, p=0.0161). CD20 expression patterns were heterogeneous; 3 patients who were initially CD20-negative became positive at relapse, while 5 patients lost CD20 expression despite retaining CD19. Conclusion: The introduction of BsAbs has improved outcomes of patients with post-CAR-T relapse, particularly in those with CD20-positive disease. However, loss of CD20 and/or CD19 is common and significantly influences both treatment selection and prognosis. Repeat biopsy at relapse is essential to guide optimal post-CAR-T therapeutic strategies.