Background: Mixed results surround the accuracy of commonly used prognostic risk scores to predict overall survival (OS) and non-relapse mortality (NRM) in allogeneic hematopoietic stem cell transplant (allo-HCT) recipients. We hypothesize that a simple prognostic score performs better than conventional scoring systems.
Patients and methods: OS risk factors, HCT-CI, age-HCT-CI, and augmented-HCT-CI were studied in 299 patients who underwent allo-HCT for myeloid and lymphoid malignancies. A scoring system was developed based on results and validated in a different cohort of 455 patients.
Results: Two-year OS was 51% (95% confidence interval (CI) 0.45–0.56); 2-year NRM was 34% (95% CI 0.29–0.39). HCT-CI and associated scores were grouped into 0–2 and ≥3. Age and HLA mismatch status were the only risk factors to affect OS in multivariate analysis (p = 0.02 and 0.05, respectively). HCT-CI and associated scores were not informative for OS prediction. The weighted scoring system assigned 0 to 2 points for age < 50, 50–64, or ≥65, respectively, and 0–1 points for no HLA mismatch versus any mismatch (except HLA-DQ). Distinct 2-year OS (62%, 53%, and 38% [p = <0.001]) and NRM (24%, 34%, and 43% [p = 0.02]) groups were characterized. The scoring system was validated in a second independent cohort with similar results on OS and NRM (p < 0.001).
Conclusions: A simple scoring system based on recipient's age and mismatch status accurately predict OS and NRM in two distinct cohorts of allo-HCT patients. Its simplicity makes it a helpful tool to aid clinicians and patients in clinical decision-making.