大阪都立大学的研究人员发现,ChatGPT 在 106 例肌肉骨骼成像病例中与放射科住院医师的准确性相匹配,但不如获得委员会认证的放射科医生。 Osaka Metropolitan University researchers found ChatGPT matched radiology residents' accuracy in 106 musculoskeletal imaging cases, but was inferior to board-certified radiologists.
大阪大都会大学研究人员比较了查特格伯特和放射学家对106个肌肉骨骼成像病例的诊断准确性。 Osaka Metropolitan University researchers compared ChatGPT and radiologists' diagnostic accuracy in 106 musculoskeletal imaging cases. 虽然GPT-4的性能优于GPT-4V,并且与放射科医生的准确性相匹配,但低于董事会认证的放射科医生. GPT-4 outperformed GPT-4V and matched radiology residents' accuracy, but was inferior to board-certified radiologists. 这项研究强调了在医学诊断中基因改变的人工智能的潜力和局限性,建议在广泛临床采用之前进行进一步研究。 The study highlights the potential and limitations of generative AI in medical diagnostics, suggesting further research before widespread clinical adoption.