Table 2.
Summary table of machine learning applied to MTX delayed elimination or Adverse reactions
| Title | Authors | Screened Feature | Sample size | Algorithms | AUROC |
|---|---|---|---|---|---|
| Risk prediction for delayed clearance of high-dose methotrexate in pediatric hematological malignancies by machine learning [7] | Zhan M |
(1) Hematocrit, risk classification, dose, SLC19A1 rs2838958, sex, dose (2) SLC19A1 rs2838958, dose, sex |
205 | (1) C5.0 decision tree + SMOTE; (2) Nomogram |
(1) AUROC = 0.807 (95% CI 0.724–0.889) (2) AUROC = 0.690 (95% CI 0.594–0.787) |
| Predictive analysis of methotrexate elimination delay based on logistic regression model and ROC curve [13] | Wang Yang | SLCO1B1 T521C | 82 |
Logistic regression |
AUROC = 0.751 (0.627–0.875) |
| Plasma creatinine as predictor of delayed elimination of high-dose methotrexate in childhood acute lymphoblastic leukemia: A Danish population-based study [15] | Schmidt, D |
(1) Absolute increase in 36 h Plasma Creatinine (2) Relative increasein 36 h Plasma Creatinine (3) Infusion plasma MTX concentration |
218 |
Linear regression |
(1) AUROC = 0.930 (95%CI 0.910-0-960) (2) AUROC = 0.930 (95%CI 0.910-0-960) (3) AUROC = 0.810 (95%CI 0.750-0-860) |
| Risk factors for high-dose methotrexate-induced nephrotoxicity [22] | Shinichiro Kawaguchi | Urine pH at day 1 | 88 | Logistic regression |
0.750 (95% CI 0.573–0.927) |
| Predicting Hepatotoxicity Associated with Low-Dose Methotrexate Using Machine Learning [37] | Hu, Qiaozhi | BMI, age, number of drugs and comorbidities, doses of folic acid, antibiotic use, gender, immunosuppressive agents, Glucocorticoid use, First MTX use, Drinking, Type 2 diabetes, Chinese traditional medicine, Dose of folic acid, Infectious liver disease, history of kidney disease | 782 |
(1) XGBoost (2) AdaBoost (3) CatBoost (4) GBDT (5) LightGBM (6) TPOT (7) RF (8) ANN |
(1)0.94 (2)0.69 (3)0.91 (4)0.53 (5)0.87 (6)0.78 (7)0.97 (8)0.65 |
Abbreviations: SMOTE, Synthetic Minority Over-Sampling Technique; RF: Random Forest; XGBoost: eXtreme gradient boosting; AdaBoost: Adaptive boosting; LightGBM: Light gradient boosting machine; AUROC: Area under the receiver operating characteristic curve; CatBoost, Categorical boosting; GBDT, Gradient Boosting Decision Tree; TPOT, Tree-based Pipeline Optimization Tool; ANN, Artificial Neural Network; BMI, Body Mass Index