Exploratory risk score incorporating donor type, MF-HCT high-risk mutations, and day 100 chimerism for outcome stratification in myelofibrosis patients undergoing allogeneic hematopoietic cell transplantation

Abstract Background Allogeneic hematopoietic cell transplantation (HCT) is curative for myelofibrosis (MF), but relapse and graft-versus-host disease-free/relapse-free survival (GRFS) remain suboptimal. While MIPSS70 informs baseline risk, it does not account for transplant-specific factors such as donor type, high-risk mutations, and early chimerism, whose additive prognostic value remains unclear. Methods We retrospectively studied 192 MF patients undergoing allogeneic HCT at City of Hope (2012–2025; cutoff April 2025). MF-HCT high-risk mutations (HRM: TP53, ZRSR2, BCOR, NPM1) were defined empirically based on endpoint-driven associations, with BCOR and NPM1 retained despite borderline significance. Mixed day 100 (D100) chimerism was defined as <95% donor cells. Cox models adjusted for MIPSS70 examined associations with overall survival (OS), relapse-free survival (RFS), and GRFS. A simple exploratory risk score assigned 1 point for HRM, mixed chimerism, and non-MRD donor. Weighted scoring was not applied. Landmark and time-dependent analyses of D100 chimerism were attempted, but due to sparse events, post-D100 findings were interpreted descriptively. Multivariable models adjusted for age, sex, MIPSS risk, and other transplant factors, which were not independently predictive and were excluded from the final exploratory risk score. Associations with outcomes were assessed using Cox proportional hazards models, and Kaplan-Meier (KM) analysis evaluated risk separation. Results Median age was 61 years (range 35–74); 58% were male. MIPSS categories were intermediate (32%), high (48%), and very high (20%). Driver mutations included JAK2 (62%), CALR (25%), MPL (7%), and triple-negative (6%). Donor types included MRD (33%), MUD (42%), haploidentical (15%), and MMUD (10%); peripheral blood was the graft source in 82%. At a median follow-up of 5.0 years, there were 111 GRFS events, 78 deaths, and 50 relapses. HRM predicted inferior GRFS (HR 2.9, p=0.0008) and RFS (HR 2.7, p=0.009), independent of MIPSS70. Mixed D100 chimerism strongly predicted relapse (HR 5.8, p<0.001) and GRFS failure (HR 4.2, p<0.001) but not OS. Donor type had no significant effect. The exploratory score stratified 3-year GRFS (>60% for 0–1 factors vs <20% for ≥2 factors). Landmark Cox models post-D100 were unstable due to few events; crude 3-year GRFS rates were lower in HRM+ (71% vs 93%), while mixed chimerism was too rare for reliable estimates. Conclusions MF-HCT HRM and mixed D100 chimerism were strong predictors of relapse and GRFS, independent of MIPSS70. The exploratory risk score refines post-transplant risk assessment and complements existing prognostic models but requires validation in multicenter cohorts. Limited post-D100 events highlight the need for larger datasets to clarify the prognostic impact of early chimerism dynamics.