CD19 chimeric antigen receptor T‐cell efficacy and toxicity in adults with Richter's transformation and response to bridging therapy

Résumé fourni par la source

Despite novel agents, Richter's transformation (RT) still carries a dismal prognosis with a median overall survival (OS) of 6–12 months.1, 2 While CD19-targeting chimeric antigen receptor (CAR)-T therapy (CD19CAR-T) has transformed outcomes in relapsed/refractory (r/r) large B-cell lymphoma (LBCL), CAR-T efficacy in RT is incompletely understood as RT patients were excluded from pivotal trials.3, 4 However, two emerging real-world datasets of CD19CAR-T for r/r RT demonstrate impressive overall response rates (ORR) and median OS rates of 63% and 8.5 months (n = 63 patients),5 and 57% and 9.9 months (n = 30 patients)6 respectively. In the latter analysis, only 14/30 patients received commercial CAR-T products and experienced poorer median progression-free survival (PFS) and OS (2.9 and 4.3 months) compared to the total cohort, which included patients receiving experimental CAR-T products (6.7 and 9.9 months).6 Separately, a DESCAR-T registry study of CD19CAR-T for RT demonstrated slightly lower ORR and complete response (CR) rates (50%; 42%) with higher toxicity compared to de novo LBCL.7 In the United Kingdom (UK), CD19CAR-T was approved for r/r RT in December 2018. Here, we evaluate the feasibility, safety and efficacy of commercially available CD19CAR-T for r/r RT in the UK, highlighting clinical features associated with PFS and OS. This retrospective intention-to-treat (ITT) analysis included r/r RT patients approved for third-line axicabtagene ciloleucel (axi-cel) or tisagenlecleucel (tisa-cel) across nine sites by the UK National CAR-T Clinical Panel (NCCP) from 2018 to 2023. Further details on methods and statistics are found in Supporting Information. Regarding patient and disease characteristics, 27 patients were approved for CD19CAR-T therapy and all underwent leukapheresis (Table 1; Figure S1). Median age at CAR-T approval was 63 years (range: 25–78). Of 27 patients, 21 (78%) had a preceding diagnosis of chronic lymphocytic leukaemia (CLL) with a median interval of 4.4 years to RT diagnosis (interquartile range [IQR]: 2.0–6.1; range: 0.7–11.7). Of 27 patients, 25 had histology demonstrating non-germinal centre LBCL and/or CD5 positivity. Central histology reports were not available for two patients (Table 1). Six patients (22%) had a concomitant diagnosis of CLL at time of RT diagnosis. Where available, cytogenetic and molecular status for CLL/RT revealed 17p deletion (del(17p)) or a TP53 mutation in 8/22 patients (36%) and unmutated immunoglobulin heavy chain variable region (IGHV) gene status in 8/10 patients (80%). Clonality assessment is not routinely available or performed in the UK and results were only available in three patients. Patients received a median of one prior line of therapy for CLL (range: 0–5) including Bruton's tyrosine kinase inhibitor (BTKi) therapy in 10/23 patients (43%), B-cell lymphoma 2 inhibitor (BCL2i) therapy in 9/23 patients (39%) and phosphoinositide 3-kinase inhibitor (PI3ki) therapy in 2/23 patients (7%) (Table 1; Table S1). Patients received a median of two prior lines of therapy (range: 2–5) for RT, including allogeneic stem cell transplantation in two patients. At leukapheresis, the median peripheral blood lymphocyte count was 1.57 × 109/L (range: 0.4–6.8). Of 27 CAR-T products, 26 (96%) met release criteria. One product failed on cell viability criteria by 2%. All 27 patients received bridging therapy (BT): 23/27 (85%) received polatuzumab-based chemo-immunotherapy (CIT); 2/27 (7%) received non-polatuzumab CIT; 2/27 (7%) received radiotherapy and 1/27 (4%) received BTKi, which was ceased prior to lymphodepletion (Table 1). Response assessment post-BT revealed complete response (CR) in 2/27 patients (7%), partial response (PR) in 9/27 (33%), stable disease (SD) in 1/27 (4%) and progressive disease (PD) in 15/27 (56%). Four patients (15%) did not proceed to CAR-T due to rapid PD post-BT (all received rituximab + bendamustine + polatuzumab vedotin [RBP]) and all died within 3 months of approval. Post-BT lactate dehydrogenase (LDH) was elevated in 13/20 patients (65%) and baseline haematotoxicity (HT) score was ‘high risk’ in 12/23 (52.2%). Of 27 patients, 23 (85%) received CAR-T infusion, of whom 20/23 (87%) received axi-cel and 3/23 (13%) received tisa-cel, with a median vein to vein time of 38 days (range: 27–87). In contrast to other reports, patients did not receive concurrent BTKi therapy.5 Regarding toxicity, any grade cytokine release syndrome (CRS) or immune effector cell-associated neurotoxicity syndrome (ICANS) affected 21/23 (91%) and 8/23 (35%) patients respectively, with ≥Grade 3 CRS and ICANS in 3/23 (13%) and 1/23 (4%) patients respectively, primarily in patients with PD post-BT and elevated LDH pre-lymphodepletion (median: 365; range: 204–1124). At month 1 (M1) post-CAR-T infusion, ≥Grade 3 neutropenia and thrombocytopenia affected 12/22 (54.5%) and 8/22 (36.3%) patients respectively. This resolved in all but two cases by month 3 (M3). Higher HT score was associated with ≥Grade 3 thrombocytopenia at M1 (high vs. low risk, 4/4 vs. 0/9; p = 0.029). Infections affected 10/22 (45.5%) patients and are detailed in the Supporting Information: ‘Results’ section and outlined in Table 2. For 1- and 3-month cytopenia; patients with progressive disease (PD) at 1 or 3 months have been excluded. No haemophagocytic lymphohistiocytosis (HLH) reported. Regarding response rates and survival, in the ITT cohort (N = 27), the 2-year OS from CAR-T approval was 42.6% (95% confidence interval [CI]: 21.6–62.2). For infused patients, ORR at M1 was 69.6%, with CR in 12/23 (52%), PR in 4/23 (17%), SD in 1/23 (4%) and PD in 5/23 (22%). By M3, three CR and PR patients developed PD and two PR patients converted to CR (Table S2). With a median follow-up of 22.3 months (IQR: 12.7–23.5), the median PFS and OS from infusion were 14.3 months (IQR: 1.3–not reached [NR]) and 19.4 months (IQR: 2.5–NR). At 2 years, PFS was 45.7% (95% CI: 24.2–64.9; Figure 1A) and OS was 49.7% (95% CI: 25.0–70.3; Figure 1B) with 11/23 (48%) patients still in CR at last follow-up (Figure 1C). Factors associated with PFS and OS in univariable analysis include LDH, response to BT and number of prior therapeutic lines (Table S3). Consistent with previous reports, high LDH was associated with inferior PFS (hazard ratio [HR] = 2.08, 95% CI: 1.30–3.33; p = 0.002) and OS (HR = 2.47, 95% CI: 1.39–4.40; p = 0.002) (Figure 1D). CR or PR response to BT was associated with superior OS (HR = 0.2, 95% CI: 0.04–0.97; p = 0.045) and a similar but non-significant association with PFS (HR = 0.38, 95% CI: 0.11–1.27; p = 0.12) (Figure 1E). More treatment lines for RT (HR = 1.99 for each additional line, 95% CI: 1.09–363; p = 0.025), prior BTKi (HR = 4.62, 95% CI: 1.23–17.31; p = 0.023) and prior BCL2i (HR = 4.51, 95% CI: 1.33–15.28; p = 0.023) were all associated with inferior PFS. del(17p) and TP53 mutation were associated with significantly worse PFS (HR = 4.17, 95% CI: 1.11–15.64; p = 0.034) but not OS (HR = 3.62, 95% CI: 0.94–13.98; p = 0.062). Only one known IGHV-unmutated patient was infused with CAR-T, so impact on PFS and OS could not be assessed. Currently 1/7 TP53-mutated and 1/8 IGHV-unmutated patients are alive and in remission at 12.7 and 22.3 months of follow-up respectively. Including non-responders, 11/23 (48%) patients relapsed at a median of 2.1 months post CAR-T (IQR: 0.9–7.4; range: 0.6–14.3): 9/11 with LBCL, 1/11 with CLL/LBCL and 1/11 unknown. Seven (64%) patients received salvage therapy (Table S4). Ten (43%) patients died at a median of 5.4 months (IQR: 1.3–16.6) post CAR-T due to PD in 8/10, grade 5 CRS in 1/10 and pneumonia in 1/10. In summary, chemo-refractory RT carries a dismal prognosis, but CAR-T therapy holds substantial promise. Here, we demonstrate that CAR-T for RT is feasible. Despite CLL/RT T-cell ‘fitness’ concerns, CAR-T manufacture was successful in 26/27 (96%) patients. Furthermore, 23/27 (85%) leukapheresed patients were inf

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.