Application of machine learning on health examination data for predicting the decrease of bone mineral density

Abstract

BACKGROUND: Timely identification and preventative strategies for diminished bone density can markedly enhance patients' quality of life and reduce economic burdens. This study intended to create machine learning algorithms that precisely forecast the probability of bone mineral density loss. METHODS: The study comprised people aged 40 years and above who received health examinations at an affiliated institution from January 2022 to January 2024. Five machine learning algorithms were employed to forecast the risk of osteoporosis: k-nearest neighbor (KNN), random forest (RF), support vector machine (SVM), artificial neural network (ANN), and logistic regression (LR). The efficacy of these algorithms was assessed according to accuracy, sensitivity, specificity, and area under the receiver operating characteristic curve (AUROC). RESULTS: This study comprised 11,132 patients, of whom 3,568 exhibited diminished bone density. The original dataset comprised 17 variables, and after data screening, 13 variables were incorporated into the machine learning model. The AUROC scores for ANN, KNN, LR, RF, and SVM were 0.882, 0.906, 0.684, 0.918, and 0.896 for males, and 0.881, 0.843, 0.784, 0.922, and 0.872 for females, respectively. The accuracies of ANN, KNN, LR, RF, and SVM were 0.83, 0.86, 0.75, 0.88, and 0.82 for males, and 0.81, 0.77, 0.74, 0.85, and 0.79 for females. CONCLUSION: Herein, we created five machine learning algorithms to precisely predict bone density reduction. The RF model had superior performance in both male and female cohorts, attaining the highest AUROC. Implementing machine learning models in clinical implementation can improve the prevention, identification, and early intervention of bone density deterioration.

Ähnliche Arbeiten