Automated Prediction of Diabetes-Related Foot Ulcer Healing at Initial Assessment
Authors' conclusion
This study demonstrated early feasibility of using TherMUL to assist clinicians in predicting 12-week DFU healing outcomes at first clinical presentation, with moderate discriminative performance. This non-contact, point-of-care approach has the potential to support earlier identification of DFUs unlikely to heal and enable more timely, targeted wound management strategies.
Study context
- Population
- Adults with neuropathic DFUs recruited from 2 specialist foot clinics in Melbourne, AustraliaClassifier trained using the images from 44 healed and 24 unhealed DFUs · healed DFUs (training) 44 · unhealed DFUs (training) 24Events: 24 unhealed DFUs
- Predictors
- TherMUL: texture features of thermal images taken at the initial visit, extracted using grey-level co-occurrence matrix (GLCM) analysis, with a support vector machine (SVM) classifier
- Predicted outcome
- Wound-healing status at 12 weeks
Methodological detailsthermal image texture analysis, support vector machine classifier, preliminary feasibility assessment on a separate unseen dataset
- Thermal image texture analysis
- Support vector machine classifier
- Preliminary feasibility assessment on a separate unseen dataset
Model performance
As reported in the abstract
Area under the curve
Mean (AUC)
0.635
Sensitivity for predicting unhealed DFUs
75.8%
Specificity for identifying healed DFUs
53%
Further findings
Predictions matching 12-week podiatry outcomes in the preliminary evaluation
- Predictions correctly matched the 12-week podiatry outcomes
- 5 of 7
- Preliminary feasibility assessment on a separate unseen dataset
Original publication
Automated Prediction of Diabetes-Related Foot Ulcer Healing at Initial Assessment. J Diabetes Sci Technol. Published 8 November 2026. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →