The relationship of social support and psychosocial complexity with chimeric antigen receptor T-cell therapy outcomes

Abstract Background: Chimeric antigen receptor T-cell therapy (CAR-T) is a transformative therapy for relapsed/refractory hematologic malignancies with unique toxicities and complex logistics, including requiring a caregiver. Moreover, utilization of CAR-T is low with multiple barriers to access. Yet, no data exists examining CAR-T outcomes among those with limited social support and/or increased psychosocial complexity. Methods: We identified all adults undergoing CAR-T at Massachusetts General Hospital (MGH) from 6/2016-1/2025 using our institutional database. We excluded patients receiving a first-in-human CAR-T product or allogeneic products. Each patient underwent a systematic social work assessment to describe their psychosocial status prior to CAR-T. We retrospectively reviewed the social work assessment and any additional documentation in the electronic health record to identify extent of social support and psychosocial complexity. We generated a social support score incorporating the existence, availability, type of caregiver and whether they resided with the patient (SSS, 0-4, with higher scores equaling higher support) and a psychosocial complexity score incorporating the social support factors plus financial concerns, language barrier, substance use disorder history, unstable housing, transportation concerns, and other psychosocial concerns (PCS, 0-10, with higher scores equaling higher complexity). We utilized the median split to define limited social support and increased psychosocial complexity. We assessed the relationship between limited social support and increased psychosocial complexity with CAR-T toxicity (intensive care unit [ICU] admission, cytokine release syndrome [CRS] [yes vs no and grade 3+], immune cell-associated neurotoxicity syndrome [ICANS] [yes vs no and grade 3+] and length of stay [LOS]) using Fisher's exact test for categorical variables and t-tests for LOS. We also evaluated the association of SSS and PCS with overall survival (OS) using multivariable Cox regression controlling for age, diagnosis, Eastern Cooperative Oncology Group performance status (ECOG PS), and location of CAR-T infusion (outpatient vs inpatient). Results: We identified 449 patients with a systematic social work assessment. The median age was 68 years (range: 23-94), and the majority were White race (84%), non-Hispanic ethnicity (87%), married/partnered (73%), and ECOG PS 0-1 (83%). The most common diagnoses were non-Hodgkin lymphoma (71%) and multiple myeloma (28%), and the majority of CAR-T was infused inpatient (67%). A plurality of patients received liso-cel (29%), followed by tisa-cel (19%), axi-cel (18%), and ide-cel (13%). On assessment, 5% of patients could not identify a caregiver, 13% had caregiver availability concerns, 4% had unstable housing, and 5% had substance use disorders. Overall, 24% of patients had financial concerns, 7% had transportation concerns, and 7% had other psychosocial concerns noted by social work assessment. The median SSS was 4, with 41% having a score below the median, and the median PCS was 1, with 38% having a score above the median. SSS below the median (compared to above median) was not associated with ICU admission (5% vs. 7%, p=0.69), CRS (76% vs. 77%, p=0.65), grade 3+ CRS (3% vs. 2%, p=0.76), ICANS (32% vs 39%, p=0.16), grade 3+ ICANS (14% vs. 17%, p=0.36), or LOS (17 days vs 17 days, p=0.87). PCS above the median (compared to below median) was not associated with ICU admission (6% vs. 6%, p=0.84), CRS (78% vs. 77%, p=0.73), grade 3+ CRS (3% vs. 2%, p=0.76), ICANS (33% vs 38%, p=0.27), grade 3+ ICANS (14% vs. 17%, p=0.51), or LOS (17 days vs 17 days, p=0.79). Neither SSS (HR 1.02, p=0.76) nor PCS (HR 0.98, p=0.71) were associated with OS. Conclusions: In a systematic social support analysis, a significant percentage of patients noted caregiver availability issues and/or financial concerns. However, neither limited social support nor increased psychosocial complexity were associated with increased CAR-T toxicity or worse survival, underscoring that psychosocial barriers should not limit access to CAR-T.