Aziz Nazha, MD of Cleveland Clinic, Cleveland, OH highlights the importance of developing cost-effective strategies for the treatment of myelodysplastic syndromes (MDS) at the American Society of Hematology (ASH) Congress, 2016 in San Diego, CA. He explains that only about 30-40% of MDS patients respond to therapy involving hypomethylating agents (azacitidine and decitabine) and as a result, it is crucial to identify which patients will respond positively to ensure they do not undergo long-term therapy without any benefit. His team conducted a study, with a cohort of 433 patients, to identify the recurrent somatic mutation that makes some MDS patients more responsive to the therapy. While prior studies associated the TP53 and TET2 mutations to response to therapy, Nazha’s team did not find this in their study. They did, however, identify the NF1 mutation as being associated with response. Unfortunately, they were not able to build a realistic model to for use in the clinic using this information and therefore tried to apply machine learning algorithms to identify patients who will respond to the therapy. Using this approach, they achieved an accuracy in predicting response of 69% for low-risk MDS patients and approximately 73% for high-risk MDS patients. To enhance the model, it is important to identify the important factors that are involved in the decision-making process of the algorithm. The factors predictive of response were hemoglobin and platelet count, and age (over 69). The aim is to use this information to build additional computational algorithms and develop a more reliable model that can be used in the clinic.