Machine Learning:Driven Intelligent Analysis of Bone Mineral Density to predict the risk of osteoporosis (Preprint)

• DOI: 10.2196/preprints.77972
• Publication Date: 2025-06-03T14:40:17Z
• Authors: Liu Bo; Hongli Chang; Peipei Li; Hongguang Chang; Shenting Yang; Wubing He
• Keywords: Preprint

<sec> <title>BACKGROUND</title> With the global population aging, osteopenia and osteoporosis have emerged as significant public health issues. Osteopenia is a precursor stage of osteoporosis, markedly increasing risk fractures. Traditional assessments for primarily rely on bone mineral density (BMD) testing, often overlooking other potential factors, such inflammation hormonal levels. Therefore, identifying new factors developing personalized assessment models essential. </sec> <title>OBJECTIVE</title> This study aims to develop machine learning model integrating systemic immune-inflammatory index (SII), clinical parameters, biochemical markers improve early detection osteopenia/osteoporosis explore inflammatory-BMD correlations. It evaluates RUSBoosted Trees Logistic Regression algorithms enhancing DXA-based screening accuracy while supporting prevention strategies through multidimensional data-driven prediction. <title>METHODS</title> employs various techniques classification regression predictions BMD, aiming enhance capabilities by analyzing individual physical examination data inflammatory markers. Data were sourced from National Health Nutrition Examination Survey (NHANES), utilizing Trees, Coarse Tree, binary tasks evaluating performance. <title>RESULTS</title> Classification analyses BMD femoral neck, trochanteric region, spine indicated that ranged 75% 91%, demonstrating good performance in tasks. Notably, excelled handling imbalanced datasets. <title>CONCLUSIONS</title> integrates with biomarkers, immune-inflammation an intelligent predicting BMD. approach enhances supports treatment strategies. Future research should further relationship between SII advance precision medicine.