Magnetic resonance imaging-derived bone marrow fat fraction and vertebral fractures in type 2 diabetes and osteoporosis

This is the authors' abstract. We don't add key points for this paper.

BACKGROUND Patients with type 2 diabetes mellitus (T2DM) and osteoporosis are at an elevated risk of fragility fractures.Magnetic resonance imaging (MRI)-derived bone marrow fat fraction (FF) may provide insights into skeletal quality beyond bone mineral density; however, its association with osteoporotic vertebral compression fractures (OVCFs) in this population remains unclear.AIM To investigate clinical and MRI-derived factors associated with prevalent OVCF in patients with T2DM complicated by osteoporosis and to develop a multivariable model for discriminating patients with and without OVCF.METHODS A retrospective analysis of clinical data pertaining to patients with diabete who underwent bone mineral density examination at Pingyang Hospital of Wenzhou Medical University from January 2023 to January 2025 was conducted.Of 216 patients screened, 134 patients with T2DM and osteoporosis met the eligibility criteria.Patients were categorized into fracture and non-fracture groups based on the presence of OVCFs.All included patients underwent spinal MRI fat quantification scanning to assess the bone marrow FF for each lumbar vertebra from L1 to L5, as well as the mean FF of these vertebrae.General and relevant clinical data were collected to identify factors associated with prevalent OVCF and construct a nomogram for discriminating prevalent OVCF status.A baseline clinical model incorporating age, fasting blood glucose (FBG), 25-hydroxyvitamin D [25(OH)D], and T-score was compared with an extended model additionally including L 2 FF.Model performance and the incremental after adjustment (P = 0.088).A nomogram incorporating clinical, biochemical, bone density-related, and MRI-derived parameters demonstrated high discrimination between patients with and without prevalent OVCF, although the additional contribution of L2 FF was limited and did not significantly improve model discrimination.External validation is required before clinical application, and the model should not be interpreted as predicting future fractures.