Effect of Radiation Therapy on Outcomes after CAR T-Cell Therapy for Non-Hodgkin Lymphoma

Abstract

<h3>Purpose/Objective(s)</h3> Patients (pts) who receive CD19 CAR T-cell therapy (CART) for non-Hodgkin's Lymphoma (NHL) with high-risk features, including bulky disease (dz) and increased inflammatory markers, are at higher risk of CART failure. The impact of radiation therapy (RT) on post-CART outcomes is poorly understood and is thus assessed here. <h3>Materials/Methods</h3> We retrospectively analyzed 133 pts who received CART for NHL from 01/2018 – 07/2021 at a single academic center. We compared pts who received RT within 100 days prior to CART infusion, either as bridging or as salvage prior to leukapheresis (RT cohort), and those who did not (no-RT cohort). Tumor ≥ 5cm defined bulk. Progression free survival (PFS) and overall survival (OS) were analyzed by Kaplan-Meier method and Cox regression modeling. Associations between RT and covariates were evaluated by Fisher's exact tests. <h3>Results</h3> 15 pts received RT, 11 as bridging and 4 prior to leukapheresis due to symptoms/bulk. Median time from RT to CART infusion was 12 days. There were no significant differences in baseline pt and dz characteristics between the 2 cohorts. In RT vs no-RT pts, bulk was present in 73% vs 54% (p=0.27) and ferritin was increased (> 290 ng/ml) in 93.3% vs 68.4% (p=0.065). More RT pts received axicabtagene ciloleucel (axi-cel) at 66.7% vs 48.3% (p=0.022). In all pts, use of systemic or any bridging therapy was associated with worse OS (HR 2.8, p<.01; 3, p<.001) and PFS (HR 1.9, p<.01; HR 2.1 p<.01) while RT bridging was not (HR 1.5, p=0.48; HR 1, p=0.9). Bulky dz, higher ECOG, and increased inflammatory markers were also associated with inferior PFS/OS. The RT cohort had a best overall response rate (ORR) of 84.6% vs 60.7% in no-RT pts (p=0.13). Median f/up was 17.2 months (mo). For the RT and no-RT cohorts, median PFS was 22 mo and 7.2 mo; median OS was not reached for RT and was 39 mo for no-RT. PFS/OS were not statistically different. There were no differences in cytokine release syndrome (CRS). RT pts had more grade ≥ 3 neurotoxicity at 53.3% vs. 18.6% (p=0.006). At day 90 post-CART, RT pts had less neutropenia at 7.7% vs. 38.3% (p=0.03) and less grade ≥3 neutropenia at 0% vs. 22.5% (p=0.04). For pts w/ bulky dz, ORR was 77.8% with RT vs. 52.5% w/o RT (p=0.28). Median RT dose for all RT pts was 29.4 Gy (range 22.5-60) delivered in median of 10 fractions (range: 9-24). RT pts w/ bulky dz received a median of 25 Gy (range: 22.5-60) also delivered in median of 10 fractions (range: 9-24). In pts w/ bulk, relapse occurred locally in 57.1% of RT pts vs. 88.6% of no-RT pts (p=0.06). <h3>Conclusion</h3> Pts who received RT had similar outcomes to those who did not and trended towards decreased local relapse. RT pts had more neurotoxicity; this may have been confounded by increased use of axi-cel and trends towards increased inflammatory markers and increased bulky dz. Reduced neutropenia in RT pts was a novel finding. Our data support further multi-center studies to investigate the role of RT in mitigating CART failure in NHL pts with high-risk, bulky dz.