Multi-Radiomics Features of Super-Resolution MRI for Predicting Minor Amputation in Wagner 3 Diabetic Foot Ulcers

Area of Science:

• Medical Imaging
• Radiology
• Diabetic Foot Complications

Background:

• Diabetic foot ulcers (DFUs) are a major cause of morbidity and amputation.
• Wagner grade 3 DFUs carry a high risk of minor amputation.
• Predictive models for amputation are crucial for timely intervention.

Purpose of the Study:

• To evaluate the effectiveness of multi-radiomics features (RFs) from super-resolution MRI in predicting minor amputation.
• To assess the predictive performance of radiomics models for Wagner 3 DFUs.

Main Methods:

• Retrospective analysis of 275 patients with Wagner 3 DFUs.
• Deep-transfer learning used to generate super-resolution proton density weighted images (SRPDWI).
• Radiomics features extracted from soft tissues and phalanges; random forest models developed for prediction.

Main Results:

• The integrated radiomics model achieved an AUC of 0.83, sensitivity of 0.71, and specificity of 0.79 in the test set.
• Models showed excellent performance in the training set (AUC 0.91-0.99).
• Clinical indicators did not improve the predictive performance of the radiomics model.

Conclusions:

• Multi-radiomics features from SRPDWI show potential for predicting minor amputation risk in Wagner 3 DFUs.
• Radiomics analysis offers a non-invasive tool for risk stratification.
• Further validation is warranted to integrate this into clinical practice.