Document Type : Original Article

Authors

1 Department of Health Information Management, School of Health Management and Information Sciences, Iran University of Medical Sciences, Tehran, Iran

2 Hematopoietc Stem Cell Research Center, Shahid Beheshti University of Medical Sciences, Tehran, Iran

Abstract

Background: Allogenic hematopoietic stem cell transplantation is considered as an effective treatment for patients with acute myeloid leukemia. However, complications of transplantation, like aGVHD, affect the efficiency of allogenic hematopoietic stem cell transplantation. The present study aimed to implement different models of data mining (DT) (single and ensemble) for prediction of allogenic hematopoietic stem cell transplantation in patients with acute myeloid leukemia (transplantation against host disease).
Method: We conducted this developmental study on 94 patients with 34 attributes in Taleghani Hospital, Tehran, Iran, during 2009–2017. In this practical study, the data were analyzed via DT algorithms, including decision tree, random forest and gradient boosting (ensemble learning), artificial neural network (Single Learning), and support vector machine. Some criteria, like specificity, accuracy, F-measure, AUC (area under curve), and sensitivity, were reported in order to evaluate DT algorithms.
Results: There were 34 transplantation-related variables; some predictors, such as liver, hemoglobin, and donor blood group, were found to be the most important ones. To predict aGVHD, the two selected algorithms included the most appropriate DM models, ANN and SVM classifiers, with ROC of 100.
Conclusion: This study indicated that DT algorithms could be successfully used for approving the efficiency of the models predicting allogenic hematopoietic stem cell transplantation.

Keywords

This article has been accepted for publication and undergone full peer review but has not been through the copyediting, typesetting, pagination, and proofreading process, which may lead to differences between this version and the Version of Record. Please cite this article as doi: 10.30476/mejc.2022.94116.1715