MRI-Based Model Improves Cardiovascular Risk Prediction in Diabetes patients: Study
A new MRI-based predictive model may help identify patients with diabetes who are at higher risk of adverse cardiovascular events, according to a study published in Radiology. The model demonstrated good predictive accuracy and outperformed established cardiovascular risk models, suggesting that cardiac MRI could enhance risk stratification and support more personalized prevention and management strategies in people with diabetes.
Patients with diabetes mellitus (DM) are at increased risk of adverse cardiovascular outcomes. Current risk scores for DM rely solely on clinical risk factors and ignore parameters that directly reflect cardiac structure and function such as imaging biomarkers. A study was done to develop a cardiac MRI-based predictive model for cardiovascular outcomes among participants with type 2 DM and evaluate the model in comparison with established clinical risk models.
This study prospectively and retrospectively enrolled participants with DM who underwent cardiac MRI between January 2016 and December 2023, comprising a training set, internal test set, and external test set, in which the risk model was developed and evaluated. The primary outcome was heart failure hospitalization or cardiovascular death. Multivariable Cox regression analysis was performed to develop the risk model.
Among 1388 participants with DM (mean age, 57 years ± 12.3 [SD]; 955 men), 145 of 810 participants in the training set experienced the primary outcome during a median follow-up of 37.6 months (IQR, 26.5–58.6 months). The MRI-based risk model demonstrated good discrimination (C index, 0.73) and acceptable calibration. On the basis of eight identified risk predictors, an integer-based SCAN-MRI (sex, coronary artery disease, age, atrial fibrillation, N-terminal pro-B-type natriuretic peptide, and MRI variables) risk score was created to predict 3-year outcome incidence.
Compared with established WATCH-DM (weight [body mass index], age, hypertension, creatinine, high-density lipoprotein cholesterol, diabetes control [fasting plasma glucose], electrocardiography QRS duration, myocardial infarction, and coronary artery bypass grafting; area under the receiver operating characteristic curve [AUC], 0.66) and Thrombolysis in Myocardial Infarction Risk Score for Heart Failure in Diabetes risk models (AUC, 0.65), the SCAN-MRI risk model showed better predictive performance (AUC, 0.76; both P < .001).
Adding cardiac MRI markers into these risk models improved the discriminative ability to predict adverse outcomes (AUC, WATCH-DM: 0.66 to 0.74 [P < .001]; Thrombolysis in Myocardial Infarction Risk Score for Heart Failure in Diabetes: 0.65 to 0.73 [P < .001]). In the external test set, the risk model showed good performance in predicting adverse outcomes (C index, 0.71).
A cardiac MRI-based multivariable risk model, integrating clinical factors and MRI parameters, demonstrated good discrimination and better performance than current established risk models in predicting adverse outcomes in participants with DM.
Reference:
Yang, W., Lian, X., Teng, F., Guo, T., Song, Y., Chen, J., Zhou, Y., et al., & Lu, M. (2026). SCAN-MRI: Cardiac MRI–integrated risk score for predicting cardiovascular events in type 2 diabetes mellitus.
