Artificial intelligence in traumatology and orthopedics. Reality, fantasy or false hopes?

Abstract

Background. In recent years, the topic of artificial intelligence (AI) in medicine has been actively discussed not just as a promising solution but the one that can help to improve some results. A significant growth of interest in AI systems all over the world began in the early-mid 2010s, which allowed us to consider the practical application of such systems. The aim of the study is to analyze all the software products (SP) registered in our country as a medical device, including those with AI technology, and to evaluate their applicability in traumatology and orthopedics. Methods. The study included all the SP having a registration certificate of a medical device according to the OKPD2 code 58.29.XX.XXX (services for publishing other software). In the state register of medical devices and organizations (individual entrepreneurs), which is engaged in the production and manufacturing of medical devices, we found 111 registered SP according to the inclusion criterion, as at February 14, 2024. Results. We proposed to categorize all registered SP as follows: systems working with the DICOM standard images (47 pcs, 42%), laboratory data (20 pcs, 18%), microscopy images (7 pcs, 6%), photographic images (5 pcs, 5%), medical information systems (4 pcs, 4%), text data mining systems (3 pcs, 3%), clinical decision support systems (3 pcs, 3%), Holter ECG analysis (2 pcs, 2%), other systems (16 pcs, 14%). Systems applicable to traumatology and orthopedics accounted for 4 pcs (4%). Conclusions. Unfortunately, the real-world applicability of existing solutions in the field of traumatology and orthopedics can be regarded as minimal in comparison with pulmonology, oncology, and laboratory diagnostics, where AI programs have already achieved significant success.