The aim of this study was to develop and validate a clinical and transplant-specific prognostic score using data from a large cohort of patients with myelodysplastic syndromes reported to the European Society for Blood and Marrow Transplantation registry. A Cox model was fitted to detect clinical and transplant-related variables prognostic of outcome. Then, cross-validation was performed to evaluate the validity and consistency of the model. Seven independent risk factors for survival were identified: age ≥50 years, matched unrelated donor, Karnofsky Performance Status <90%, very poor cytogenetics or monosomal karyotype, positive cytomegalovirus status of the recipient, blood blasts >1%, and platelet count ≤50 × 109/L prior to transplantation. Incorporating these factors into a four-level risk score yielded hazard ratios for death, with low-risk (score of 0-1) as reference, of 2.02 (95% CI: 1.41-2.90) for the intermediate-risk group (score of 2-3), 3.49 (95% CI: 2.45-4.97) for the high-risk group (score of 4-5), and 5.90 (95% CI: 4.01-8.67) for the very high-risk group (score of >5). The score was predictive of survival, relapse-free survival, relapse, and non-relapse mortality (P<0.001, respectively). Cross-validation yielded significant and reproducible improvement in prognostic ability with C-statistics being 0.609 (95% CI: 0.588-0.629) versus 0.555 for the Gruppo Italiano Trapianto di Midollo Osseo registry and 0.579 for the Center for Blood and Marrow Transplant Research registry. Prediction was even further augmented after applying a nomogram using age and platelets as continuous variables showing C-statistics of 0.628 (95% CI: 0.616-0.637). In conclusion, compared to existing prognostic systems, this proposed transplant-specific risk score offers improved performance with respect to post-transplant risk stratification in myelodysplastic syndromes.
