The Application of Artificial Intelligence In The Assessment of Burn Injuries

Ching Clarise Lam¹, Ernest Tay¹, Cheng Hean Lo²,
¹Monash University, Melbourne, VIC, Australia
²Victorian Adult Burns Service, The Alfred, Melbourne, VIC, Australia

Abstract

Introduction
Burn injuries may be difficult to triage at point-of-care, expensive to assess, and inaccessible to most hospitals. The extensive use of Artificial Intelligence (AI) in burn depth classification has purportedly reduced diagnosis time and improved diagnostic accuracy. The aim of this systematic review is to update the current evidence and summarise the role of AI in burn depth classification.

Methods
A structured literature review was performed in Ovid MEDLINE and Ovid EMBASE. All published peer-reviewed randomised or non-randomised clinical studies, cohort studies, and prospective or retrospective studies involving AI in determining human burn wound depth were included.

Results / Discussion
In this paper, 36 out of 3061 identified articles were included. AI classification of burn depth provides superior diagnostic accuracy than burn specialists (60-80% accuracy). Most commonly used AI models are Convolutional Neural Networks (CNN), Support Vector Machine (SVM) and Artificial Neural Network (ANN). Their accuracies are 65-97%, 55-96%, and 82-91% respectively.

Currently, AI for burn depth assessment is used to triage surgical candidacy (especially in areas without specialist input) and predict healing time in larger burn units. However, with limited clinician familiarity in utilising AI, heterogeneity in training data and some resistance in clinical uptake, using AI is far from commonplace.

Conclusion
AI in assessing burns is still experimental but promising. To safely use this technology, burn units should train their AI models with human clinical data and body part specifications. Validation tools should be also developed to critically evaluate AI burns depth analysis in a clinical context.

Biography

Clarise is a final-year medical student at Monash University with solid enthusiasm for trauma, critical care and medical education. She is also interested in medical systems, leadership and AI. She can be found in boulder gyms or nature if not studying.

Ernest is currently a final year medical student at Monash University, with a keen interest in surgery and medical education. He is also heavily involved in student advocacy and clinical research. Outside of medical school, he is a student athlete and coach for Ultimate Frisbee, an avid hiker, and previously served as an Inspector in the police force.