The application of thermography as a screening tool to detect type 2 diabetic foot disease

Abstract

Aim: To determine whether thermography can be used as a screening modality to detect T2DM complications by measuring the cutaneous temperature of specific areas of the foot and using the logistic regression curves and equations provided in literature. Method and Methodology: This was a quantitative, postpositivist and non-experimental correlational research to investigate the sensitivity, specificity, positive-predictive value (PPV) and negative predictive value (NPV) of thermography in identifying diabetic foot complications in participants with Type 2 Diabetes Mellitus. For each participant, a thermogram of the plantar aspect of the foot was taken using FLIR T630. FLIR ResearchIR programme generated the cutaneous temperature on eight regions of interest (ROIs). The probability of complications for each foot was calculated using equations previously published in literature based on the cutaneous temperature. Another podiatrist, blinded to the results of the thermography analysis, performed Ankle-Brachial Pressure Index, Spectral Doppler Waveform Analysis (for detection of peripheral arterial disease), and 10g Monofilament and 128Hz Tuning Fork to diagnose neuropathy or neuroischaemia. The probability of diabetic foot complications was then compared with the actual presence or absence of diabetic foot complications. The sensitivity, specificity, positive-predictive and negative predictive values were calculated for four different probability cut-off points (0.4, 0.5, 0.6, 0.7) on eight different regions of interest. Results: Using a 0.5 probability cut-off point, it was found that the sensitivity of detecting complications was high, ranging from 93% to 97% in the toes and 93% to 100% in the forefoot. However, specificity was lower, ranging from 54% to 69% in the forefoot and 69% to 80% in the toes. The positive predictive value varied from 0.71 to 0.79 in the toes and from 0.64 to 0.71 in the forefoot, while the negative predictive value ranged from 0.92 to 0.96 in the toes and 0.92 to 1 in the forefoot. Increasing the probability cut-off point yielded higher specificity and PPV but lower sensitivity and NPV. Importantly, all participants with PAD, neuropathy, and neuroischemia were identified using the 1st MTH ROI. Conclusion: Results indicate that thermography can effectively detect chronic diabetic foot complications with high sensitivity and specificity. In a clinical setting, this technology can be a helpful tool for reducing waiting times and resources for diabetic foot screening and detecting T2DM foot complications faster and more conveniently.