EBMT NEWSLETTER | September 2015 | Volume 48 - Issue 3

EBMT
Important dates
Machine learning is a field of artificial intelligence dealing with the construction and study of systems that can learn from data, rather than follow explicitly programmed instructions. It is commonly applied in "big data" settings –such as, web applications and computerized vision. Machine learning is part of the data mining approach for knowledge discovery.
 
Led by Dr. Roni Shouval, Prof. Arnon Nagler, Dr. Myriam Labopin, and Prof. Mohamad Mohty on behalf of the EBMT’s ALWP, investigators have developed a machine learning based prediction model for mortality following allogeneic hematopoietic stem cell transplantations (HSCT) in acute leukemia patients. The study, entitled “Prediction of Allogeneic Hematopoietic Stem-Cell Transplantation Mortality 100 Days After Transplantation Using a Machine Learning Algorithm: A European Group for Blood and Marrow Transplantation Acute Leukemia Working Party Retrospective Data Mining Study”, was recently published in the Journal of Clinical Oncology http://www.ncbi.nlm.nih.gov/pubmed/26240227.
 
More than 28,000 patients were included, and served for model development and validation. The prediction model is based on the alternating decision tree algorithm, and shows competence in prediction of short and long term mortality. It reveals interactions between predictors, and is available for free online use http://bioinfo.lnx.biu.ac.il/~bondi/web1.html.
 
In another study using machine learning algorithms, Dr. Shouval and colleagues from the ALWP ran repetitive computerized simulations (in-silico modeling) to explore the boundaries of prediction of HSCT associated mortality. By applying this novel computational approach, authors empirically demonstrate the sample size required to develop robust prediction models, rank variables according to their prognostic contribution and guide algorithm selection for improved prediction. This work was awarded with the Best Science Award at the last EBMT annual meeting in Istanbul.
 
Dr. Shouval estimates that with the advent of large registries, incorporating clinical, genomic and biological data, machine learning will enter routine use in medical prognostic research. The recent publications by the ALWP pave the way for applying a similar methodology in additional aspects of hematopoietic stem cell transplantation.
Bookmark and Share