Artificial intelligence in foot and ankle surgery: Current applications, limitations and future directions.

Abstract

Background: Artificial intelligence (AI) is increasingly studied in orthopedics for image interpretation, automated measurement, risk prediction, rehabilitation monitoring, and patient communication. Foot and ankle care is well suited to these applications because decisions often depend on imaging measurements, deformity assessment, wound monitoring, and longitudinal recovery data.

Purpose: To summarize current and emerging AI applications in foot and ankle care, distinguish clinically supported uses from experimental concepts, and identify principles likely to remain relevant as individual models evolve.

Study design: Current concept review.

Methods: A targeted narrative search of PubMed/MEDLINE, recent systematic reviews, reference lists, and relevant United Kingdom guidance was performed for publications available through August 10, 2026. Peer-reviewed studies addressing a defined foot or ankle clinical task and reporting clinically relevant validation were prioritized.

Results: The strongest evidence is in image-based applications, including ankle fracture detection and classification, automated deformity measurements, and three-dimensional analysis of weight-bearing computed tomography. Image-based models can assess diabetic foot ulcers, and wearable sensors combined with machine learning show promise for estimating Achilles tendon loading. Perioperative prediction models remain largely retrospective. Large language models may assist with patient information but are not sufficiently reliable for unsupervised diagnosis or triage. The proposed Foot and Ankle AI Assistant remains an unvalidated research concept.

Conclusions: AI may support selected aspects of foot and ankle care, particularly imaging and automated measurement, but most applications still require external and prospective validation before routine clinical use. AI should support, rather than replace, clinical judgment and surgeon-led decision-making.