Detection of Bone Fractures by AI Algorithms

Comparison of the Effectiveness of Recognition of Changes by a Doctor and AI Models

Abstract

Introduction. Artificial intelligence is playing an increasingly significant role in medicine and has the potential to assist not only specialists but also other medical professionals in detecting fractures based on X-ray and CT scans. Algorithms based on deep learning are particularly useful in radiology and orthopedics.

Purpose of the study. Summary of publicly available publications and studies comparing the effectiveness of AI and physicians in recognizing bone fractures.

Materials and methodology. A review of the literature available on PubMed, Google Scholar and Scopus was conducted.

Conclusions. The use of artificial intelligence algorithms can reduce the time required for fracture detection to as little as one minute. AI is better at identifying fractures that are difficult for humans to detect but struggles with more obvious and visible fractures. Artificial intelligence has the same or lower sensitivity and specificity in fracture detection as a specialist doctor but achieves higher accuracy compared to a resident doctor or a general practitioner.

Summary. Based on the analyzed studies, it can be observed that artificial intelligence  will prove useful as an aid; a suggestion in diagnosis for a young doctor, but does not replace specialist doctors, who achieve higher accuracy in detecting fractures on X-rays and CT scans. Furthermore, it is worth reconsidering the reanalysis of existing studies and comparing AI results not only with human knowledge but also with CT scans, as they are considered more reliable.