JMIR Publications

ABSTRACT

Background:

Diabetic foot ulcers (DFUs) are a serious complication of diabetes that contribute substantially to morbidity, mortality, and healthcare burden. Accurate and timely wound assessment is essential for effective DFU management; however, conventional assessment methods remain limited by subjectivity, time constraints, and inter-observer variability.

Objective:

This scoping review aimed to map and synthesize artificial intelligence (AI)–based assessment models used in the evaluation of diabetic foot ulcers, with a focus on application domains, methodological approaches, and reported performance outcomes.

Methods:

A scoping review was conducted based on the five-stage framework proposed by Arksey and O’Malley and reported in accordance with the PRISMA extension for Scoping Reviews (PRISMA-ScR). Literature searches were performed in PubMed, ProQuest, and Scopus for studies published between 2014 and 2024. Study selection and data charting were conducted independently by two reviewers using predefined inclusion and exclusion criteria based on the Population – Concept – Context (PCC) framework. Extracted data were synthesized narratively and grouped according to major AI application domains.

Results:

A total of 76 records were identified, of which 16 studies met the eligibility criteria. The included studies primarily applied AI models for DFU image segmentation, risk prediction, classification, and wound monitoring. Convolutional neural networks were the most frequently used algorithms, with reported performance metrics including accuracy, Dice similarity coefficient, and area under the curve. Most studies relied on retrospective, single-center datasets and demonstrated limited external validation.

Conclusions:

AI-based assessment models show potential to support diabetic foot ulcer evaluation and monitoring. Nevertheless, heterogeneity in study design, limited dataset diversity, and challenges related to clinical integration and external validation remain significant. Future research should prioritize standardized datasets, prospective validation, and real-world implementation to support safe and effective clinical adoption.