Predicting delayed methotrexate elimination in pediatric acute lymphoblastic leukemia patients: an innovative web-based machine learning tool developed through a multicenter, retrospective analysis

BACKGROUND: High-dose methotrexate (HD-MTX) is a potent chemotherapeutic agent used to treat pediatric acute lymphoblastic leukemia (ALL). HD-MTX is known for cause delayed elimination and drug-related adverse events. Therefore, close monitoring of delayed MTX elimination in ALL patients is essential. OBJECTIVE: This study aimed to identify the risk factors associated with delayed MTX elimination and to develop a predictive tool for its occurrence. METHODS: Patients who received MTX chemotherapy during hospitalization were selected for inclusion in our study. Univariate and least absolute shrinkage and selection operator (LASSO) methods were used to screen for relevant features. Then four machine learning (ML) algorithms were used to construct prediction model in different sampling method. Furthermore, the performance of the model was evaluated using several indicators. Finally, the optimal model was deployed on a web page to create a visual prediction tool. RESULTS: The study included 329 patients with delayed MTX elimination and 1400 patients without delayed MTX elimination who met the inclusion criteria. Univariate and LASSO regression analysis identified eleven predictors, including age, weight, creatinine, uric acid, total bilirubin, albumin, white blood cell count, hemoglobin, prothrombin time, immunological classification, and co-medication with omeprazole. The XGBoost algorithm with SMOTE exhibited AUROC of 0.897, AUPR of 0.729, sensitivity of 0.808, specificity of 0.847, outperforming the other models. And had AUROC of 0.788 in external validation. CONCLUSION: The XGBoost algorithm provides superior performance in predicting the delayed elimination of MTX. We have created a prediction tool to assist medical professionals in predicting MTX metabolic delay. SUPPLEMENTARY INFORMATION: The online version contains supplementary material available at 10.1186/s12911-023-02248-7.