Dynamic Mortality Risk Prediction in Myelodysplastic Syndromes Using Longitudinal Clinical Data

Patients with myelodysplastic syndromes (MDS) exhibit diverse disease trajectories necessitating different clinical approaches ranging from watch-and-wait strategies to hematopoietic stem cell transplantation. Existing risk scores like the IPSS-R or EASIX provide static risk stratification at diagnosis but do not capture evolving disease dynamics. We address this problem by introducing a dynamic, data-driven approach to predict short-term mortality risks repeatedly, across the patient’s disease course. We develop a machine learning model based on gradient-boosted decision trees to estimate one-year mortality risks from both longitudinal parameters from blood values and diagnosis-based features. We train the model on a dataset of patients from the MDS Registry Düsseldorf (n=1024) and validate on patients from University Hospitals Heidelberg (n=286) and Mannheim (n=31). Validations on independent cohorts achieved AUROC scores of around 0.8 and better predictive performance for one-year survival compared to a diagnosis-only baseline model. The model accurately predicted mortality risks as early as within the first 90 days of diagnosis. Feature importance analysis revealed clinically plausible feature-label relations, supporting interpretability. This dynamic risk model enables continuous, individualized assessment of one-year mortality risk in MDS patients, offering a supplement to static scores used at diagnosis. Our results highlight the utility and importance of including longitudinal parameters in risk assessment analysis.