AUTHORS
Gazeau Nicolas, Rémi Tilmont2, David Beauvais3, Nicolas Gower, MD4, Remi Pescarmona5, Julie Demaret, MD6, Benjamin Podvin, PhD7, Nicolas Jacquet-Francillon8, Maxence Gambiez9, Maryam Idlhaj10, Ibrahim Yakoub-Agha, MD, PhD11, Hervé Ghesquières, MD, PhD12, Emmanuel Bachy I, MD, PhD13, Pierre Sesques, PhD14
AFFILIATIONS
1 CHU Lille, Lille, France
2 CHU Lille, Lille, France
3 CHU Lille, Lille, France
4 CHU Lille, Lille, France
5 CHU Lyon, Lyon, France
6 CHU Lille, Lille, France
7 CHU Lille, Lille, France
8 CHU Lyon, Lyon, France
9 CHU Lille, Lille, France
10 CHU Lyon, Lyon, France
11 CHU Lille, Lille, France
12 CHU Lyon, Lyon, France
13 CHU Lyon, Lyon, France
14 CHU Lyon, Lyon, France
ABSTRACT
Background: Outcomes and management of patients (pts) with partial response or stable disease (PR/SD) at month 1 (M1) after CD19-directed CAR T cells for large B-cell lymphoma (LBCL) remain poorly defined. We assessed a multiparametric routine clinical, biological, and imaging parameters to predict subsequent complete response (CR) conversion and guide risk-adapted surveillance.
Methods: This retrospective two-center study included LBCL pts treated with axi-cel, liso-cel, or tisa-cel with PR/SD at M1. A multimodal strategy was applied, integrating clinical, biological, and imaging parameters collected before and after infusion. CAR T cell expansion, quantified by flow cytometry with cross-validated across centers, was analyzed as a continuous variable. Response at M1 was assessed per Lugano 2014 criteria. CR conversion predictors were evaluated using Wilcoxon/Fisher tests and logistic regression. OS and PFS were compared with a control cohort in CR at M1.
Results: Overall, 112 PR/SD pts at M1 after CAR T cell infusion for LBCL were included (91 PR, 21 SD). Median age was 60 and 66% were male. Pts received axi-cel (63.2%), tisa-cel (19.8%), or liso-cel (17%). Histology was LBCL in 75.5%, transformed LBCL in 16.0%, and PMBL in 5.7% of pts. Median of lines of therapy (LOT) was two. At lymphodepletion (LD), disease status was PD in 67.0%, PR in 16.0%, and SD in 14.2%. Median CAR T cell peak at day 7-10 post-infusion was 84.5 cells/mm3 (range: 0-3594). CRS occurred in 95 pts (85%) (Grade ≥3, 5%), while ICANS occurred in 53 pts (47%) (Grade ≥3, 9%).
Among evaluable pts at M1, 49 (45%) subsequently converted to CR, with comparable conversion rates between initial PR (46%) and initial SD (38%) pts. In univariate analysis, conversion to CR was associated with more favorable disease status at LD (PR versus SD/PD p=0.016), higher CAR T peak (median 140 vs 68 cells/mm3; p=0.01), lower pre-CAR SUVmax (median 18 vs 23; p=0.003), and lower M1 SUVmax (median 6 vs 10.9; p<0.001). No significant association was observed with sex, histology, LOT, LDH, CRP, ferritin, CAR-HEMATOTOX score at LD or at M1, post-CAR consolidation treatment, CRS or ICANS grade or duration, neutropenia phenotype according to ICAHT.
In multivariate analysis, PR at LD (OR 6, 95% CI 1.52-26.3, p=0.011), CAR T peak (OR 1.0026, 95% CI 1.00-1.00, p=0.0237), and M1 SUVmax (OR 0.92, 95% CI 0.86-0.98, p=0.0149) remained independently associated with conversion to CR.
In survival analyses including 292 CR, 91 PR, and 21 SD pts at M1, CR was associated with superior OS (p=0.003) and PFS (p<0.0001) than PR/SD.
Median OS and PFS in CR pts were 92 months and 36.6 months, compared with 35.4 months and 4.1 months in PR/SD pts, respectively. Outcomes did not differ significantly between PR and SD pts, for either OS (median 51.1 vs 21.1 months, p=0.31) or PFS (4.2 vs 3.0 months, p=0.46).
Among PR/SD patients, stratification by the median CAR T peak (84.5 cells/mm3) distinguished high from low expanders. High expanders showed a trend toward longer OS (51.1 vs 21.3 months, p=0.31) and significantly longer PFS (8.8 vs 3.0 months, p=0.04) compared with low expanders.
Based on the multivariable model, we derived a 3 point prognostic score assigning 1 point for each favorable factor: at LD, CAR T peak above the median, and post-CAR SUVmax below the median. This score identified 4 groups with stepwise outcome differences. Median OS was 10.6, 37.2, 51.1 months, and not reached for scores 0, 1, 2, and 3, respectively (p=0.0019), while median PFS was 2.7, 3.6, 25.8, and not reached, respectively (p<0.001). After a median follow-up of 2 years, 2-year OS rates were 32.0%, 53.5%, 67.3%, and 100.0%, and 2-year PFS rates were 12.5%, 27.2%, 52.9%, and 66.7% for score groups 0, 1, 2, and 3, respectively.
Conclusions: To our knowledge, this is the first study to use an integrated multimodal approach to characterize outcomes in pts with PR/SD after CD19 directed CAR T therapy for LBCL. Conversion was independently associated with pre-LD disease status, higher CAR T expansion, and lower M1 SUVmax. The resulting three-parameter score identified a progressive improvement in survival across increasing score groups, with pts scoring 2 or 3 achieving outcomes approaching those observed in pts in CR. These routine parameters may refine early risk stratification and management of PR/SD pts.
External validation is ongoing and will be presented at the meeting.
DISCLOSURES
Gazeau Nicolas: No relevant conflicts of interest to declare.