MRI-Derived Vertebral Bone Marrow Fat Fraction for Osteoporosis Prediction in Type 2 Diabetes: Evaluation of Inter-Device Consistency and Clinical Risk Predictors

Introduction : This study investigates the association between lumbar spine bone marrow fat fraction (FF) on MRI and clinical risk factors in patients with type 2 diabetes mellitus (T2DM). Additionally, it evaluates the inter-device consistency of FF measurements and determines the predictive value of FF for osteoporosis. Materials and method : A total of 109 T2DM patients were enrolled, with lumbar spine FF quantified using T1-VIBE-DIXON sequences on a Siemens 3.0T MRI and mDIXON-Quant sequences on a Philips 3.0T MRI. Inter-device agreement was assessed. Bone mineral density (BMD) was assessed using dual-energy X-ray absorptiometry (DXA), with participants stratified into normal, osteopenic, or osteoporotic categories based on T-score thresholds. ROC analysis was conducted to establish the optimal FF cutoff for osteoporosis prediction, while linear regression identified clinical factors associated with FF, including gender, age, BMI, diabetes duration, diabetic peripheral neuropathy (DPN) and biochemical parameters. Results : FF measurements demonstrated strong inter-device agreement, with no significant bias between scanners (P > 0.05). ROC analysis determined an FF threshold of 61.4% for osteoporosis prediction (AUC = 0.87, 95% CI: 0.73-0.95, sensitivity: 72.5%, specificity: 87.1%) and 55.2% for osteopenia (AUC = 0.77, 95% CI: 0.67–0.88, sensitivity: 78.9%, specificity: 64.5%). Regression analysis identified female gender (B = 7.13, P < 0.001), advanced age (B = 0.38, P < 0.001), LDL-C > 2.6 mmol/L (B = 309, P =0.02), and DPN (B = 3.03, P =0.02) as independent predictors of increased FF. Conclusion : Lumbar spine FF emerges as a reliable biomarker for osteoporosis risk in T2DM patients, demonstrating robust inter-device comparability. Identifying key clinical risk factors enhances osteoporosis risk stratification, supporting MRI-based marrow fat assessment as an adjunct to DXA for early diagnosis and personalized bone health management.