Impact of Prior Chimeric Antigen Receptor T-Cell Treatment on Graft-Vs-Host Disease and Non-Relapse Mortality after Allogeneic Hematopoietic Cell Transplantation: A Case-Control Propensity Matching Analysis

Chimeric antigen receptor T cell (CAR T) therapy has become a principal treatment modality in the management of relapsed/refractory hematologic malignancies in the past 7 years. Yet, in most cases, subsequent allogeneic hematopoietic cell transplantation (HCT) is warranted, either due to relapsed disease or as consolidation post-CAR T. Immune related toxicities, excess graft-vs-host disease (GvHD) and subsequent increase in non-relapse mortality (NRM) have been described after HCT with prior checkpoint inhibitor therapy or other immunotherapies (Haverkos et al, Blood 2017; Merryman et al, Blood 2017). Few retrospective reports have been published describing the feasibility and efficacy of HCT in patients previously treated with CAR T, but studies, to date, have not analyzed the incidence and severity of GvHD and NRM after HCT with prior CAR T compared to a similar population without prior CAR T treatment. Herein, we compared the cumulative incidence (CI) of GvHD and NRM in a matched case control study using propensity score to describe the impact of prior CAR T treatment on HCT outcomes.

We conducted a retrospective analysis of HCT outcomes in adult patients with any hematologic malignancy who underwent their 1st HCT at City of Hope from December 2017 until March 2024 after prior CAR T (the case group). Pediatric patients and those receiving a 2^nd HCT were excluded. This cohort was then matched with a control group of patients undergoing HCT during the same period but without prior CAR T at a ratio of up to 1:2 based on: diagnosis, donor type and age, conditioning, and graft source. The primary objective was incidence of NRM at day 100. Secondary objectives included incidence of day 100 acute GvHD, 1-year chronic GvHD, 1-year overall and disease-free survival (OS and DFS), CI of relapse (CIR), and engraftment. NRM, CIR, and GvHD were calculated using cumulative incidence curves and compared by the Gray's test, respectively. OS and DFS were calculated using KM curves and compared using the log-rank test.

In total, 161 patients were included: 65 in the case group and 96 in the control group. The most common CAR T products received were an investigational product (41.5%) followed by axicabtagene ciloleucel (26.2%); 60 (92.3%) received a CD19-targeting product. No systemic line of treatment was given between CAR T and HCT in 47.7% of patients. The median time from CAR T infusion to HCT was 6 months (range, 1-35). The two cohorts were well-balanced in terms of matching factors of diagnosis, donor type and age, conditioning, and graft source. There were no significant differences in other baseline factors other than disease status at transplant and disease-risk index (DRI) (P<0.001). The most common diagnosis was ALL (29 in the case group and 58 in the control group), followed by NHL (29 in the case group and 24 in the control group), MDS (6 in the case group and 12 in the control group), and BPDCN (1 in the case group and 2 in the control group). About half of patients in both case and control cohorts received myeloablative conditioning (55.4 vs 50%, respectively), most received peripheral blood stem cells as the graft source (93.8 vs 95.8%, respectively), and close to half received tacrolimus and sirolimus as GvHD prophylaxis (46 vs 44.4%, respectively).

Day-100 NRM in the case vs control groups were 6.2 vs 6.3% (P=0.62). Day-100 CI of grades II-IV and III-IV acute GVHD in the case vs control groups were 42 vs 35% (P=0.35) and 20% vs 16% (P=0.36), respectively. The 1-year CI of any and extensive chronic GvHD in the case vs control groups were 40% vs 45% (P=0.16) and 38% vs 42% (P=0.30), respectively. The 1-year CIR was significantly higher in the case vs control groups (27% vs 15%; P=0.027). There were no significant differences in neutrophil engraftment (P=0.57, median=16 days, 95%CI: 15-17) and platelet engraftment (P=0.22, median=19 days, 95%CI: 17-24) between the two cohorts. The 1-year OS and DFS in was significantly lower in the case vs control groups (62% vs 80%; P=0.018) and (53% vs 72%; P=0.018), but the significance was lost after adjustment for DRI in multivariable analysis.

In our analysis, unlike checkpoint inhibitors, receiving CAR T prior to HCT does not predict for a higher risk of NRM or GvHD, and thus, CAR T can be safely delivered in eligible patients without impacting HCT outcomes. Disease risk and status at the time of HCT remains the ultimate predictor of survival in this refractory population regardless of the type of therapy.