It has been announced that the AI-powered medical diagnostic system developed by Microsoft, MAI-DxO, achieved a higher accuracy rate compared to previous diagnostic results of physicians in an evaluation conducted using past medical cases.
According to information shared by the company, 304 "complex cases" published in The New England Journal of Medicine were examined by AI agents in the study. The performance of MAI-DxO was compared with the diagnostic results of doctors in similar past cases.
According to the findings, the system provided approximately 2.3 times higher diagnostic accuracy when evaluating cases compared to the past results of doctors who utilized open sources such as books and the internet. When compared to the results of doctors attempting to make a diagnosis without using external sources, it was stated that the success gap increased to approximately 3.7 times.
The details of the research were published in the journal Science. However, it was emphasized that the study was based on the evaluation of past medical cases in a simulation environment rather than real-time clinical practice, and that real physicians did not participate directly in the comparison process.
For this reason, although the results point to the potential of artificial intelligence in medical diagnosis, it is stated that the impact of such systems on clinical use needs to be evaluated through broader studies that include real patient conditions.
News Source: 12punto
