Comprehensive Summary
Javanmardi et al. analyzed the rising role of image recognition artificial intelligence in diagnosing rare bone diseases. Specifically, the researchers sought to gauge global clinician interest in potentially utilizing AI-based clinician assistants to improve standard imaging processes of the genotype-phenotype correlations with these rare bone diseases. Some of the AI-based image recognition softwares are Next Generation Phenotyping, which is an application of advanced computer vision techniques to obtain medical image data of people with specific genetic conditions, and Bone2Gene AI, which will eventually be designed to distinguish unique imaging patterns linked to different, various rare bone diseases. To do this, the researchers sent out an online survey from March to September 2024, disseminating it via scientific meetings, newsletters, emails, and phone calls; the researchers collected responses from 103 healthcare providers across 27 countries. 89% of the respondents were physicians with 81% of those physicians working at academic medical centers. While specialties were quite diverse and spread out, most of the medical specialties included pediatrics, endocrinology, and medical genetics. The results of the survey revealed that 91% of the survey participants believed that the artificial intelligence image recognition was highly significant in diagnosing rare bone diseases. While around half of the participants did express a degree of concern over the possibility of errors with the artificial intelligence, 81% of the participants still reported that they were likely to integrate the image recognition into the diagnostic processes of their work.
Outcomes and Implications
While medical rare bone disease image recognition artificial intelligence software is still developing (such as in the case of Bone2Gene AI), feedback from numerous healthcare providers has shown that many of the survey participants would be open and willing to use AI-based clinician assistants. Despite the concern of errors and artificial intelligence hallucination, many of the participants were still willing to incorporate new AI image recognition into their work. More research into this medical imaging software will be necessary to ensure fewer mistakes or errors, but this imaging software has proven to be quite significant in diagnosing rare bone diseases as of now. Transparent development and collaboration between clinicians and the AI developers will be necessary to enforce effective and safe clinical implementation of the software.