AI-Powered Thermography for Diabetic Foot Risk Stratification: Multicenter Cross-Sectional Study.

Area of Science:

• Medical Technology
• Artificial Intelligence in Healthcare
• Diabetology

Background:

• Diabetic foot complications pose a significant health burden, especially in the Middle East and North Africa.
• Existing screening methods for diabetic foot complications face limitations in objectivity, invasiveness, and scalability.
• Novel approaches are crucial for effective early detection and management of diabetic foot risks.

Purpose of the Study:

• To assess the efficacy of the AI-powered Thermal Foot Scan (TFScan) system.
• To evaluate TFScan's ability to identify patients at high risk of diabetic foot complications using noninvasive thermography.
• To analyze temperature patterns and asymmetries for risk stratification.

Main Methods:

• A multicenter, cross-sectional study involving 1120 individuals with diabetes in Saudi Arabia.
• Utilized a smartphone-compatible infrared camera for thermal imaging.
• Applied AI algorithms to analyze foot angiosomal temperature patterns and asymmetries, stratifying risk into four categories.
• Correlated TFScan classifications with clinical risk factors, neuropathy symptoms, and thermal abnormalities.

Main Results:

• 9.3% of participants were classified as moderate or high risk, showing significantly higher prevalence of diabetic complications.
• The high-risk group had increased rates of peripheral artery disease (20.2%), cardiovascular disease (57.7%), neuropathy (11.5%), and foot deformities (14.4%).
• Significant thermal abnormalities, including temperature asymmetries (≥2.2 °C), were concentrated in moderate- and high-risk groups, indicating perfusion deficits and inflammatory states.

Conclusions:

• The TFScan system successfully stratified diabetic patients into clinically relevant risk categories.
• Moderate- and high-risk groups demonstrated a greater burden of vascular, neuropathic, and thermal abnormalities.
• AI-enhanced thermography shows promise as a scalable, objective tool for proactive diabetic foot management, warranting further longitudinal validation.