The impact of cannabis use on patient outcomes post immune checkpoint inhibitor (ICI) therapy in a longitudinal observational trial: The DiRECT Cohort.

12081 Background: Prior small retrospective studies suggest that cannabis might compromise the efficacy of ICIs for cancer therapy due to its immunosuppressive properties. Evidence from large prospective studies is lacking. Methods: We addressed this research gap in the DiRECT Cohort (URCC21038, NCT05364086), an ongoing, observational trial enrolling self-identified Black and White cancer patients (melanoma excluded) who are about to start ICI-containing therapy through the URCC NCORP Research Base nationwide clinical trial network. Longitudinal data on cannabis use are collected through self-administered surveys at baseline (A1), 6 months (A3), and annually (A4+), along with clinical data and patient-reported outcomes on physical and psychological symptoms. This interim analysis on cannabis use includes 1,666 patients enrolled between 04/01/2022 and 08/31/2024, with outcome data updated to 11/30/2024. Results: The mean age was 63.7 (±12.4) years; 408 (24.5%) patients were Black; 905 (54.3%) were women; the most common diagnosis was lung cancer (621 or 37.3%); and a majority had stage IV disease (906 or 54.4%). At A1, 284 (17.1%) patients reported cannabis use in any form, mostly orally (58.1%) or via inhalation (32.3%). Use rates remained stable at A3 (15.6%) and A4 (15.4%) (P = 0.47); yet patients reported fewer days of use in a month (P < 0.0001) but more times in a day (P = 0.07) than at A1. Cannabis users were younger and more likely to be Black, from states with permissive cannabis laws, and current or former cigarette smokers (P’s < 0.05), with no differences observed by gender, cancer type, or stage. After a median follow-up of 10.4 months (range: 0.03-31.0), 381 patients died, and 174 patients had progressed disease or entered hospice care. Median overall survival (OS) and event-free survival (EFS) were slightly higher in cannabis users than non-users at A1 (6.1 vs. 5.3 months, P = 0.07 and 6.0 vs. 5.4 months, P = 0.15, respectively). Cox hazards models adjusted for age, gender, cancer type, and stage revealed no significant association between cannabis use and OS (adjusted hazards ratio [aHR] = 0.82, 95% CI 0.61-1.10) or EFS (aHR = 0.92, 95% CI 0.72-1.17). Subgroup univariate analysis showed cannabis use was associated with longer OS and EFS within men (n = 743) and former smokers (n = 765) (P’s < 0.05), which became non-significant after adjusting for covariates in Cox models, except for OS within former smokers (aHR = 0.52, 95% CI 0.33-0.83). A significant interaction was noted between cannabis use and smoking status (P = 0.02). Conclusions: In this nationwide, diverse community-based oncology cohort, 17% of patients reported cannabis use concurrent with ICI therapy. Unlike previous retrospective studies, our prospective analysis finds no detrimental effect of cannabis on OS or EFS. Further analyses are ongoing to elucidate its impact on symptom management. Clinical trial information: NCT05364086 .