Early Identification of Low Bone Density Risk Using a Radiofrequency Echographic Multi Spectrometry-Based Prediction Model

Osteoporosis is a major public health problem characterized by reduced bone mineral density (BMD) and increased fracture risk. Early identification of individuals with low bone density remains essential for prevention. This study aimed to evaluate a multivariable logistic regression-based model integrating clinical and Radiofrequency Echographic Multi Spectrometry (REMS)-derived parameters for the detection of low bone density in women. A total of 324 women undergoing REMS assessment of the lumbar spine and hip were included. Clinical variables (age, body mass index [BMI], menopausal status, and lifestyle factors) and REMS-derived measurements were analyzed. Binary logistic regression was used to identify independent factors associated with low BMD (T-score < −1 SD). Model performance was assessed using odds ratios (ORs), omnibus chi-square testing, pseudo-R2 statistics, and classification accuracy. The Youden index was applied to determine optimal cut-off values. Age, menopausal status, BMI, and basal metabolic rate (BMR) were identified as independent factors associated with low lumbar spine and femoral neck BMD. Increasing age and menopause were associated with higher odds of low bone density, whereas higher BMI and BMR were inversely associated with low bone density. In conclusion, a logistic regression model combining clinical and REMS-derived parameters demonstrated the ability to identify women with low bone density, supporting improved individualized risk stratification in clinical practice.