Duration of lymphopenia after CAR-T cell therapy is associated with increased risk of non-melanoma skin cancers and decreased survival

Abstract Introduction: Although CD19-directed chimeric antigen receptor T-cell therapies (CAR-T) have improved the prognosis for patients (pts) with relapsed/refractory (r/r) B-cell lymphomas (BCL), only about 30% of pts achieve durable disease-free survival and late effects of this therapy are still being defined. Immunocompromised pts have high risk of developing non-melanoma skin cancers (NMSC), including cutaneous basal cell and squamous cell carcinomas (Hall et al, Am Soc Clin Oncol Educ Book. 2020). Thus, we examined the incidence of NMSC post CAR-T in r/r BCL pts as a potential marker of ongoing immunosuppression. Methods In this single-center, retrospective study, we reviewed medical records of BCL pts who received CAR-T between May 2018 - October 2024 at the Hospital of the University of Pennsylvania. Eligible pts had no antecedent history of NMSC and routine immune system studies (e. g. , absolute lymphocyte count ALC, CD3 and CD4 T cell counts) and clinical follow-up (f/u) available. ALC recovery (ALC-R) post CAR-T was defined as ALC ≥ 1x10³ cells/µL; recovery period was defined as time from CAR-T infusion to last f/u, progressive disease, or death. Kaplan-Meier and log-rank tests were used for survival analyses. Cox proportional hazard models and linear regression were used for univariate and multivariate analyses. Results Of 369 pts screened, 284 were eligible for analysis (22 excluded due to death before CAR-T, 63 due to prior NMSC). Of 284 pts, 21 (7. 4%) developed NMSC post CAR-T. For 21 pts with NMSC post CAR-T, median age was 65 years (yrs; IQR 61-71) ; 16 pts (76. 2%) were male; 19 (90%) had large B-cell (LBCL), 1 (5%) follicular (FL), 1 (5%) mantle cell (MCL) lymphoma; 12 pts (57. 1%) received tisacel, 5 (23. 8%) axicel, 3 (14. 3%) lisocel, 1 (4. 8%) brexucel. For 263 pts without NMSC post-CART, median age was 61 yrs (IQR 52-68) ; 166 pts (63%) were male; 210 (80%) had LBCL, 32 (12%) FL, 21 (8%) MCL; 141 (53. 6%) pts received tisacel, 67 (25. 5%) axicel, 34 (12. 9%) lisocel, 21 (8. 0%) brexucel. Of 154 pts (54. 2%) with ALC-R, 8 pts (5. 2%) had NMSC; ALC did not recover (ALC-NR) in 130 pts (45. 8%) and 13 pts (10%) had NMSC. Median time to ALC-R was 28 days (IQR 9-189) ; median time for ALC-NR pts was 91 days (IQR 50-240). For ALC-R vs ALC-NR: median age was 60 yrs (IQR 51-68) vs 62 yrs (IQR 57-69) ; 93 (60. 4%) vs 89 (68. 5%) were males; CAR-T received tisacel 94 (61. 0%) vs 59 (45. 4%), axicel 31 (20. 2%) vs 41 (31. 5%), lisocel 16 (10. 4%) vs 21 (16. 2%), brexucel 13 (8. 4%) vs 9 (6. 9%) ; LBCL diagnoses (dx) were 119 (77. 3%) vs 110 (84. 6%). Bendamustine was lymphodepletion (LD) for 223 pts (78. 5%) (NMSC, n=15 71. 4%; no NMSC, n=208 79. 1%; ALC-R, n=121 78. 6%; ALC-NR, n=102 78. 5%). Best complete response rates were numerically higher in pts with NMSC (16/21 pts 76. 2%) than pts without NMSC (147/263 pts 55. 9%) and in pts with ALC-R (102/154 66. 2%) than ALC-NR (61/130 47. 0%). For pts with ALC-R, 43 pts (27. 9%) died of lymphoma (32/43 72. 1%), infection (6/43 14. 0%), cardiac event (1/43 2. 3%) secondary malignancy (1/43 2. 3%), and unknown (4/43 9. 3%). For pts with ALC-NR, 50 pts (38. 5%) died of lymphoma (39/50 78%), infection (6/50 12%), secondary malignancy (1/50 2%), and unknown (4/50 8%). Analysis for risk of developing NMSC post CAR, including age, dx, gender, race, LD, CAR-T product, and baseline lymphocyte counts, showed ALC-R vs ALC-NR and time to ALC recovery were significant (P0. 05 and P0. 001, respectively) ; pts with ALC-NR had an increased risk of NMSC (p=0. 05). We evaluated whether pts with ALC-NR (n=130, 45. 8%) had an increased risk of other secondary malignancies (n=11, 3. 9%) ; although not statistically significant, a trend was noted (p=0. 1). Further, we examined ALC-R and survival. Median f/u for ALC-R and ALC-NR pts was 751 and 558 days, respectively. Pts with ALC-NR had median PFS 162 days (95% CI, 115-272) and median overall survival (OS) not reached (95% CI, 796-not reached) ; pts with ALC-R had median PFS 690 days (95% CI, 413-1066) and median OS was not estimable (n=154, events=43). Both PFS and OS were statistically significant between groups with improved outcomes for ALC-R pts (HR 0. 48, p0. 001 for PFS and HR 0. 43, p0. 001 for OS). Discussion While infection is the primary concern for most clinicians in the immediate post CAR-T period, our retrospective study highlights additional risks related to prolonged lymphopenia, including decreased PFS and OS and increased risk of NMSC.