Survival rates for myelodysplastic syndromes (MDS) varies widely – from months to decades – highlighting the need for prognostic systems that incorporate clinical, pathologic, and molecular data to predict survival more accurately in this patient group.
In a study published in the Journal of Clinical Oncology, researchers sought to develop a prognostic model for patients with MDS receiving a variety of MDS-targeted treatments. First, they enrolled 1,471 patients with comprehensively annotated clinical and molecular data in a training cohort and analyzed data using machine learning techniques. They then built a prognostic model using a random survival algorithm, which was then validated in external cohorts.
In the training cohort, the median age was 71 years, and the most commonly mutated genes were SF3B1, TET2, and ASXL1. “The algorithm identified chromosomal karyotype, platelet, hemoglobin levels, bone marrow blast percentage, age, other clinical variables, seven discrete gene mutations, and mutation number as having prognostic impact on overall and leukemia-free survivals,” the authors reported.
The model was then validated in an independent external cohort of 465 patients, which included patients with MDS treated in a prospective clinical trial, patients with paired samples at different time points during the disease course, and patients who underwent hematopoietic stem cell transplantation.
The model developed by the researchers also appeared to outperform established prognostic models in MDS. “The new model was dynamic, predicting survival and leukemia transformation probabilities at different time points that are unique for a given patient, and can upstage and downstage patients into more appropriate risk categories,” they concluded.