Abstract / Summary
Therapy-related myeloid neoplasms (t-MN) after chimeric antigen receptor (CAR) T-cell therapy occur with short latency, but whether they carry an independent survival disadvantage is unclear. We assembled a multicenter retrospective cohort of 399 adults with t-MN after CAR T-cell therapy (n=56), autologous hematopoietic cell transplantation (auto-HCT; n=98), or standard chemotherapy/radiation (n=245), and constructed a nearest-neighbor latency-matched cohort (n=220). Median latency was 15.2 months for post-CAR-T versus 38.0 and 30.0 months for post-auto-HCT and standard therapy (p<0.001). In the matched cohort, median overall survival (OS) was 9.1 months for post-CAR-T versus 16.9 and 18.7 months for post-auto-HCT and standard therapy (p=0.013). In multivariable Cox models, the hazard ratio for overall mortality was 1.98 (95% CI 1.33-2.94) for post-CAR-T and 0.99 (95% CI 0.74-1.32) for post-auto-HCT, each relative to standard therapy; The post-CAR-T estimate was 1.85 (95% CI 1.21-2.82) after adjustment for TP53 status and 2.15 (95% CI 1.23-3.76) in an era-restricted cohort with near-complete molecular data. TP53 variant allele frequency independently predicted OS (HR 1.14 per 10% increment, 95% CI 1.03-1.26). The hazard ratio for allogeneic HCT, modeled as a time-varying covariate, was 0.74 (95% CI 0.52-1.06). Post-CAR-T t-MN carries an independent survival disadvantage not explained by TP53 clone burden; allogeneic HCT remains the only potentially curative option.