AI-enabled wound assessment in surgical practice: Current capabilities and future directions

Background: Chronic wounds, diabetic foot ulcers, pressure ulcers, and venous leg ulcers remain a persistent clinical challenge. Assessment depends heavily on subjective judgment, which drives inter-observer variability and inconsistent management. Artificial intelligence (AI) is addressing this gap by generating objective, reproducible, data-driven evaluations that complement clinician expertise.
Aim: This review synthesises current AI applications in chronic wound assessment and outlines practical pathways for wider clinical integration.
Current State: Smartphone-based imaging platforms enable standardised wound photography with automatic area and volume measurement. Deep learning models classify granulation, slough, and necrotic tissue at accuracy levels comparable to experienced clinicians. Sensor-based systems detect early infection signals through pH shifts, temperature gradients, and exudate composition. Predictive models estimate healing trajectories from patient demographics and comorbidities. AI-assisted decision support, linked to telemedicine infrastructure, extends specialist oversight to remote and resource-limited settings.
Future Directions: Biosensor-embedded dressings will enable continuous real-time monitoring of wound pH, oxygenation, temperature, and bacterial load, triggering early alerts before clinical deterioration is visible. Three-dimensional and four-dimensional imaging combined with patient-specific genomic and microbiome data will support personalised prognosis models and digital twin simulations for non-invasive treatment planning. AI-guided robotic debridement, integrated tele-wound clinics, and risk-stratification models for preventing chronic wound development represent further near-horizon applications.
Conclusion: AI has the potential to shift wound care from periodic, subjective review to continuous, standardised monitoring. Realising this potential requires prospective clinical validation, interoperability with existing health infrastructure, and coordinated input from clinicians, engineers, and regulators.