Checkpoint inhibitors exposure does not affect survival and safety in PMBCL patients receiving CAR T-cells therapy: An analysis of the italian CART-SIE study

Abstract

Abstract Treatment of relapsed/refractory primary mediastinal B-cell lymphoma (R/R PMBCL) patients has been revolutionized by the advent of checkpoint inhibitors (CPIs) and CAR T-cells therapy. Most patients receiving CAR T-cells have been treated with CPIs, either as part of salvage treatment, as bridging treatment or after CAR T-cells failure. This study aims to evaluate whether exposure to CPIs influences efficacy and toxicity profile of CAR T cells therapy. CAR-T SIE is a prospective multicenter observational study enrolling patients treated with antiCD19 CAR T-cells in 23 Italian centers. Among the 1309 patients (pts) enrolled in the study, 93 with R/R PMBCL who received axicel (n=90) or lisocel (n=3) were eligible for this analysis. All pts provided written informed consent for the study. Fifty-five pts received CPIs either as part of salvage therapy (n=20), as bridging therapy (n=26) or both (n=9). An additional 11 pts received CPIs following CAR T-cell failure, whereas 38 pts never received CPIs. No significant differences in clinical characteristics were observed between CPI exposed and CPI-naïve pts: median age was 37 and 34 years, respectively; 49% and 45% of patients were female; 29% and 39% presented stage III or IV; 31% and 40% had extranodal disease, and 29% and 40% had received more than two prior lines of therapy. After a median follow-up of 23.7 months (range, 12.2 – 39), 2-years OS and PFS for the entire cohort were 81.9% and 61.8%, respectively. 2-years PFS was similar between patients never exposed to CPIs (62.3%) and those who received CPIs as salvage or bridging (61.3%) (HR1.12, 95%CI 0.56-2.26). Similarly, no significant differences in 2-years OS were observed, with rates of 89% and 77.5% in pts never exposed and exposed to CPIs, respectively (HR 1.96, 95%CI 0.62 – 6.16). No significant differences in PFS were observed between pts who received CPIs exclusively as bridging therapy and those pts treated with CPI both during salvage and bridging therapy (HR 0.33; 95%CI 0.11-1-02). PFS was similar between pts who received CPI in second line versus third and further line (HR 0.44, 95%CI 0.11-1.79). The type of CPI used, Nivolumab or Pembrolizumab, had no impact on either OS or PFS. Overall response rate at 30 days after CAR T infusion was comparable between CPI exposed and CPI-naive pts (87% and 81%, respectively); the proportion of patients achieving a metabolic CR at 30 days was similar between the two groups (47% and 52% respectively). Following CAR T-cell failure, 11 patients received CPIs, achieving an ORR of 60%, with 1 partial response (PR) and 5 complete responses (CR). In comparison, 8 patients were treated with alternative therapies, resulting in an ORR of 40% (1 PR and 1 CR). No significant difference in ORR and OS was observed between patients rescued with CPI and those receiving other treatments, but numbers are very small and no conclusion can be drawn. Regarding safety, CRS was observed in 85% of CPI-exposed patients and 95% of CPI-naïve patients (p = ns), predominantly of grade1-2 (87% and 91% of pts respectively). Immune effector cell-associated neurotoxicity syndrome (ICANS) was reported in 38% CPI-exposed patients and 35% CPI-naïve patients, severe ICANS (grade > 3) occurred in 48% patients treated with CPIs and in 38% of patients who were not (p=ns). In this large real-world cohort of patients with R/R PMBCL treated with CAR T-cell therapy, prior exposure to CPI did not appear to impact CAR T-cell outcomes, including response rates, PET scan results, and survival. Moreover, CPI exposure was not associated with an increased risk of CRS or ICANS. Findings of this study suggest that the use of CPI is safe and may be beneficial as part of salvage therapy to improve disease control before infusion or to rescue patients following CAR T-cell failure. Due to the limited number of patients experiencing CAR T-cell failure, the optimal salvage strategy in this setting remains undetermined.