Impact of graft versus host disease prophylaxis with post transplant cyclophosphamide on the prognostic value of HCT-CI

Abstract

Abstract Background: The hematopoietic cell transplantation specific comorbidity index (HCT-CI) is a crucial validated prognostic tool used to make baseline assessment for allogeneic hematopoietic cell transplantation (HCT). Since the results of BMT CTN 1703, the adoption of post-transplant cyclophosphamide (PTCY) in reduced intensity conditioning (RIC) matched related (MRD) and unrelated donors (MUD) transplants has transformed the landscape of graft versus host disease (GVHD) prevention. Based on the CIBMTR database, PTCY is increasingly being used in myeloablative (MAC) transplants (65% MAC MUD and 43% MAC MRD). High-dose cyclophosphamide has a distinct risk profile, including bladder and cardiac toxicity, which could influence traditionally used HCT-CI-based prognostication. Furthermore, it remains unclear how the individual comorbidities within the HCT-CI impact outcomes based on the utilization of PTCY. To that end, we assess the prognostic value of HCT-CI and evaluate the impact of HCT-CI on prognosis between patients undergoing PTCY and non-PTCY based GVHD prophylaxis. Methods: Patients with a diagnosis of MDS (myelodysplastic syndrome) or AML (acute myeloid leukemia) who underwent HCT between 2018 and 2022 at DFCI were identified. All comorbidities constituting the HCT-CI, HCT protocols, and clinical outcomes were retrieved from an institutional database and validated on chart review by two independent reviewers. We used multivariable Fine and Gray regression analysis using a bootstrap validation method for NRM to identify an optimized subset of HCT-CI. We utilized the Akaike information index (AIC) for the assessment of model fit and C-index for predictive ability of models. Results. Five hundred eighty-six patients were included, of whom 184/586 (31%) received PTCY-based GVHD prophylaxis. Baseline comorbidities as captured by the HCT-CI were balanced between patients in non-PTCY and PTCY-based GVHD prophylaxis subgroups, except for the renal comorbidity, which was more common in the PTCY subgroup but was only present in 8 patients in the entire cohort. Higher HCT-CI score was associated with an increased cumulative incidence of NRM in univariable analysis (3-year NRM: 22% (≥ 5) vs 12% (2-4) vs 4.8% (0-1), p=0.0002) and multivariable analysis (sub distribution HR 4.28 for HCT-CI≥ 5, p=0.0003, 2.61 for 2-4, p=0.013). A subset of the comorbidities: diabetes, infection, peptic ulcer disease, renal, rheumatologic, and severe pulmonary comorbidities, had the similar impact of NRM as the total HCT-CI score, with C-index 0.773 (95% CI 0.722, 0.824) for the subset vs 0.768 (95% CI 0.716, 0.82) (p=0.78) for the total. The presence of any of the six comorbidities was associated with high NRM (3-year estimate: 23% vs 6.6% (p<0.0001); this was true regardless of the use of PTCY. Severe pulmonary comorbidities had a strong independent association with NRM in both PTCY and non-PTCY subgroups. Infection, renal, and rheumatologic comorbidities had a strong association with NRM only in the non-PTCY subgroup. Cardiac comorbidities and diabetes had a strong association with NRM only in the PTCY subgroup. The addition of cardiac comorbidities further improved the C-index to 0.848 (95% CI 0.757, 0.94) in the PTCY group, but this had no effect in patients receiving non-PTCY-based GVHD prophylaxis. Conclusion. We demonstrate that in this single-center study, HCT-CI retains prognostic significance for patients receiving PTCY-based GVHD prophylaxis. A subset of six comorbidities had similar prognostic power as HCT-CI, but cardiac comorbidities weren't part of this subset; adding them to the subset improved predictability in the PTCY group but not in the non-PTCY group. Most importantly, cardiac comorbidities and diabetes showed a strong association with NRM only in the PTCY group. While our study supports the prognostic power of HCT-CI in patients using PTCY, it also highlights the need to better account for the increased risk of NRM with cardiac comorbidities and diabetes. Future studies could focus on elucidating the links between cardiac comorbidities and NRM when using PTCY, and reducing the risk of NRM in these patients, potentially by using a lower dose of cyclophosphamide.