Spectral Analysis for Detecting Infections in Diabetic Foot Wounds: A Non-Invasive Diagnostic Approach

Diabetic foot ulcer infection is a major problem for people with diabetes and if untreated, can lead to amputation and even death. The current method of detecting the infection requires swab tests and wet lab; this typically involves contact with the wound and causes time delay due to the need for the wet lab. To overcome this, spectral analysis of the images has been proposed, which however requires hyperspectral imaging, that is both, expensive, and the analysis are slow. We propose the use of multi-spectral imaging analysis to detect the infection and have validated the method on real-world data. This technique has identified infection with sensitivity of 1, specificity 0.83 and accuracy of 0.71. The results show that this technique based on the spectral properties of the microbes present in the wound can be used to identify the infection in the wounds.Clinical Relevance- An estimated 167 million people will suffer from diabetic foot ulcers in their lives, with possible hospitalization and amputations. Early detection of infection is essential for the management of these wounds. Our work has the potential for providing real-time, non-contact assessment of the wounds, reducing hospitalization, and preventing amputations.