MRI Radiomics Dataset for Osteoporosis Prediction Based on L1–L4 Vertebrae

This dataset contains MRI radiomics features extracted from L1–L4 vertebrae for osteoporosis prediction. The dataset includes vertebra-level radiomics features, clinical variables, osteoporosis outcome labels, and a subset used for intraclass correlation coefficient (ICC) analysis to assess feature reproducibility. Radiomics features were extracted from manually delineated regions of interest using PyRadiomics v3.0.1. MRI images were acquired using a United Imaging u790 3.0T superconducting MR system, and bone mineral density was measured using a GE Prodigy Advance DXA system. The dataset is intended to support reproducible research on osteoporosis prediction using vertebral MRI radiomics.