Patient‐based prediction algorithm of relapse after allo‐HSCT for acute Leukemia and its usefulness in the decision‐making process using a machine learning approach

Abstract

Although allogeneic hematopoietic stem cell transplantation (allo‐HSCT) is a curative therapy for high‐risk acute leukemia (AL), some patients still relapse. Since patients simultaneously have many prognostic factors, difficulties are associated with the construction of a patient‐based prediction algorithm of relapse. The alternating decision tree (ADTree) is a successful classification method that combines decision trees with the predictive accuracy of boosting. It is a component of machine learning (ML) and has the capacity to simultaneously analyze multiple factors. Using ADTree, we attempted to construct a prediction model of leukemia relapse within 1 year of transplantation. With the model of training data (n = 148), prediction accuracy, the AUC of ROC, and the κ‐statistic value were 78.4%, 0.746, and 0.508, respectively. The false positive rate (FPR) of the relapse prediction was as low as 0.134. In an evaluation of the model with validation data (n = 69), prediction accuracy, AUC, and FPR of the relapse prediction were similar at 71.0%, 0.667, and 0.216, respectively. These results suggest that the model is generalized and highly accurate. Furthermore, the output of ADTree may visualize the branch point of treatment. For example, the selection of donor types resulted in different relapse predictions. Therefore, clinicians may change treatment options by referring to the model, thereby improving outcomes. The present results indicate that ML, such as ADTree, will contribute to the decision‐making process in the diversified allo‐HSCT field and be useful for preventing the relapse of leukemia.

A Systematic Review of Machine Learning Techniques in Hematopoietic Stem Cell Transplantation (HSCT)

“…Importantly, due to dynamic characteristics unique to each study, it remains challenging to identify an optimal ML technique that can be applied robustly across all conditions. However, our results suggest that ADT technique could be useful in the field of HSCT due to their interpretability which is crucial in the clinical settings as shown in primary studies [17,24,25,27]. The findings were generalizable, robust, and clinically relevant.…”

“…However, the requirement of discretizing input variables could be a major drawback. ADT technique was used in four of the reviewed studies where it was used primarily to predict survival/death and relapse post-HSCT [17,24,25,27]. ADT outperformed RF model in one study [16,20,24,26,35,36,40].…”

“…Best performance was achieved, while predicting likelihood of death and relapse for the next 0-3 months. Another study [17] analyzed data of 217 patients with acute leukemia to predict relapse within one year of HSCT using ADT. However, major clinical and demographic factors considered here for building predictive model were age, diagnosis (e.g., AML), rDRI (refined disease risk index) [43], donor type, graft source, use of total body irradiation, and conditioning regimen.…”

Interactive web application for plotting personalized prognosis prediction curves in allogeneic hematopoietic cell transplantation using machine learning

“…Supporting this idea, Fuse et al also showed that the predictive performance of relapse in a single institute where the predictive model was developed using ADTree was better than that in another institute using the identical predictive model (AUC; 0.75 vs 0.67). 26 We have published the source code of our web application on the Internet. Each transplant center can also develop a web application using an in-house prognosis prediction model with its past patient data.…”

Application of machine learning in the management of acute myeloid leukemia: current practice and future prospects

“…80 Relapse after transplantation can be estimated by using alternating decision trees. 81 ML can also be used to predict development of acute GVHD after allogeneic transplantation 82 and stratify outcomes in chronic GVHD, revealing novel groups at risk based on clinical phenotypes more accurately than current approaches based on cumulative severity. 83…”