AI Outperforms Physicians In Medical Image Interpretation

A study evaluated how a reasoning model developed by OpenAI performs using messy electronic health records. The study is described as one of the largest conducted to date. The research indicates the technology's effectiveness in real-world diagnostic settings despite messy electronic health records. David Rich, chief clinical officer at New York's Mount Sinai health system, was not involved in the research. Researchers evaluated an OpenAI reasoning model's ability to diagnose patients and make clinical decisions. The model outperformed a large physician baseline when processing complex case reports and clinical vignettes. Researcher Rajman Rai noted the model excelled at parsing electronic health records for specific cases, such as a patient with lupus. The study was part of a series of experiments regarding AI in medical decision-making. Evidence shows AI can interpret medical images better than expert physicians. This capability applies to X-rays, mammograms, CAT scans, PET scans, retina scans, cardiograms, and pathology slides. The development follows several years of progress in the field. The technology is described as a step forward for medical imaging.