Abstract Background:Chimeric antigen receptor T-cell therapy (CAR T) has improved outcomes in relapsed/refractory diffuse large B-cell lymphoma (DLBCL), mantle cell lymphoma (MCL), and multiple myeloma (MM). Despite its benefits, CAR T carries risks such as cytokine release syndrome (CRS) and immune effector cell-associated neurotoxicity syndrome (ICANS), which can increase healthcare resource utilization (HCRU), including hospitalizations and intensive care unit (ICU) stays. As CAR T use expands in the US, understanding real-world (rw) toxicity management and HCRU is essential. This study aimed to assess the prevalence and treatment of CRS/ICANS, as well as associated HCRU, in patients (pts) with DLBCL, MCL, and MM treated within a large, multisite oncology network.Methods:This retrospective cohort study used the Flatiron Health Research Database, a longitudinal, EHR-derived pt-level dataset from routine clinical care across US academic and community oncology practices. The dataset included structured and unstructured EHR data curated via technology-enabled abstraction, including human review, NLP, and machine learning. Pts diagnosed with DLBCL, MCL, or MM on or after January 1, 2017 were included. A subset receiving commercial (non-trial) CAR T was further curated to generate disease-specific registries enriched with CAR T-relevant variables. CAR T administration was evaluated by infusion setting (inpatient or outpatient) and practice type (academic or community). The rw adverse events of CRS and ICANS were characterized by prevalence, grade, time to onset, duration, resolution, treatment, infusion setting, and practice type. HCRU included hospitalization rates, median length of stay (LOS), and ICU admissions. Results: A total of 2,520 received CAR T: 1,600 with DLBCL, 187 with MCL, and 733 with MM. Median age was 61 to 63 yrs and mostly males (58%-73%). Most pts received CAR T in the inpatient setting (72%-80%) and at academic sites (84%-90%). Any grade CRS occurred in 66% pts with DLBCL, 76% with MCL, and 77% with MM; of which 5.2%, 7.0%, and 2.8% were grade 3, respectively. In DLBCL, those with CRS grade 3 significantly decreased from 9.6% in 2017-2020 to 3.1% in 2021-2024. CRS resolution rates were 74% in DLBCL, 78% in MCL and 85% in MM. CRS rates across the setting of CAR T infusion were: inpatient (72%-80%) and outpatient (53%-87%), and academic (68%-77%) and community (62%-79%) sites. Median time to CRS onset from CAR T was 3 (DLBCL), 4 (MCL), and 3 (MM) days, with duration of 5, 5, and 3 days, respectively. Any grade ICANS occurred in 36% (DLBCL), 52% (MCL), and 18% (MM), of which 35%, 42%, and 17% were grade 3, respectively. In DLBCL, pts with ICANS grade 3 significantly decreased from 40% in 2017-2020 to 32% in 2021-2024. ICANS resolution rate was 70% in DLBCL, 72% in MCL, and 82% in MM. Median time to ICANS onset was 6 (DLBCL), 7 (MCL), and 5 (MM) days, with duration of 6, 6.5, and 2 days, respectively. Pts received the following CRS/ICANS treatment(s) across diseases: tocilizumab (64%-82%), dexamethasone (40%-89%), anakinra (5.5%-22%), methylprednisolone (1.9%-14%), siltuximab (0.9%-4.1%). Hospitalization rate after CAR T was 51% of DLBCL, 69% of MCL, 56% of MM. Median LOS was 5 days in DLBCL and MM, 7.5 days in MCL. ICU stays ranged from 5.2% in MM to 14% in MCL. For grade 3 CRS, 44%-70% were hospitalized, median LOS was 4-8.5 days and 19%-50% had ICU stay. For ICANS grade 3: 59%-82% hospitalized, with median LOS of 5-11.5 days and 18%-32% had ICU stay. Conclusions: In this rw study of over 2,500 pts with DLBCL, MCL, and MM treated with CAR T, the prevalence and severity of CRS/ICANS were generally consistent with clinical trials, with a notable decline in grade 3 events over time in DLCBL. Majority of pts were treated in the inpatient setting and at academic centers, though a meaningful proportion received CAR T in the outpatient setting (8-18%) or at community practices (5-13%), reflecting increased diversity in care delivery. Across disease cohorts, pts with grade 3 CRS had more than double ICU-level care compared to pts with lower-grade events, reflecting the substantial HCRU burden associated with severe toxicities. While most CAR T therapy was delivered in academic centers, the occurrence and management of CRS and ICANS across both academic and community sites support the feasibility of delivering and managing CAR T therapy safely in varied rw practice settings.
Chen et al. (Mon,) studied this question.