Prediction models incorporating second metacarpal cortical index for osteoporosis in rheumatoid arthritis: Externally validated machine learning models developed using data from the KURAMA cohort.

OBJECTIVE: Osteoporosis and osteopenia are significant concerns in rheumatoid arthritis (RA), predisposing patients to fragility fractures. While dual-energy X-ray absorptiometry (DXA) is the gold standard for bone mineral density (BMD) assessment, simpler screening tools are needed. This study aims to assess the correlation between the second metacarpal cortical index (2MCI) and BMD in RA patients, and to evaluate machine learning (ML) models utilizing 2MCI and clinical parameters for predicting osteoporosis/osteopenia presence and BMD.

METHODS: Data from the KURAMA cohort (n = 302) and an external validation cohort (n = 32) were analyzed. BMD in the hip and forearm was obtained using DXA and 2MCI was calculated from plain hand X-ray. ML models were trained to predict osteoporosis/osteopenia presence and BMD using 2MCI and clinical variables and validated using external cohort.

RESULTS: 2MCI correlated significantly with hip and forearm BMD. ML models incorporating 2MCI and other clinical parameters showed good performance in predicting osteoporosis/osteopenia presence and BMD. External validation demonstrated the generalizability of the models.

CONCLUSION: ML models utilizing 2MCI and clinical parameters show promise for osteoporosis screening in RA patients.