Data Mining in Bone Marrow Transplant (BMT) Medical Records
Keywords: Data mining, health records, machine learning, bone marrow transplant.
This project in in collaboration with the Medical Operations Research Laboratory (morLAB), University of Toronto and the Hematology-Oncology and Stem Cell Transplantation Research Center (HORSCT), Shariati Hospital and Tehran University of Medical Sciences.
Overview of Research
Bone marrow transplant (BMT) is a procedure that is common in the treatment of certain types of cancer, e.g., leukemia and lymphoma, and other diseases such as thalassemia. The procedure replaces faulty bone marrow with healthy stem cells. Patients undergoing the procedure face various risk factors such as infection, relapse of the disease, or rejection by the graft, each of which could cause serious complications or even death. It is hoped that analysing the data from past transplants and their outcomes would shed light on some of the underlying causes of success vs. failure and could potentially help in saving lives. With this goal, advanced machine learning techniques are being employed to perform a thorough statistical analysis and data mining in records from close to 2,000 patients undergoing BMT over the course of 19 at Shariati Hospital (Tehran University of Medical Sciences).
Research Team
Babak Taati, University of Toronto
Jasper Snoek, University of Toronto
Dionne Aleman, University of Toronto
Ardeshir Ghavamzadeh, Shariati Hospital
