Cutting-edge artificial intelligence (AI) technology, embodied in a sophisticated large language model (LLM) known as GPT-4 (Generative Pre-Training–Model 4) from OpenAI, has demonstrated the ability to rival, and in certain instances, surpass human ophthalmologists in diagnosing and treating patients with glaucoma and retina diseases. Research conducted at the New York Eye and Ear Infirmary of Mount Sinai (NYEE) presents compelling evidence of AI's potential to provide decision-making support to medical professionals in the realm of ophthalmology.
Published in JAMA Ophthalmology, the study underscores the transformative impact of advanced AI tools trained on extensive datasets, text, and images. By leveraging its vast knowledge base, the AI system exhibits remarkable proficiency in diagnosing and managing cases related to glaucoma and retina disorders, offering valuable insights that align with or exceed those of seasoned ophthalmic specialists.
Dr. Andy Huang, lead author of the study and an ophthalmology resident at NYEE, remarks on the striking performance of GPT-4, noting its capacity to not only assist but also match or surpass the expertise of human specialists. The study involved a comparative analysis between responses generated by GPT-4 and those provided by 12 attending specialists and three senior trainees from the Department of Ophthalmology at the Icahn School of Medicine at Mount Sinai.
Through a meticulous evaluation of responses to standardized questions and patient cases, AI emerged as a formidable contender, demonstrating comparable or superior accuracy and thoroughness in medical advice and assessments, particularly excelling in glaucoma-related inquiries. Dr. Louis R. Pasquale, senior author of the study, highlights AI's proficiency in diagnosing and managing both glaucoma and retina patient cases, underscoring its potential as a valuable adjunct to clinical practice.
Dr. Huang envisions AI's integration into mainstream ophthalmic practice as a means to enhance diagnostic precision, alleviate clinician workload, and expedite patient access to expert guidance. By facilitating informed decision-making and streamlining treatment protocols, AI holds the promise of optimizing patient care and outcomes in the field of ophthalmology.
While further validation and refinement are warranted, the study heralds a promising future wherein AI-powered tools complement human expertise, ushering in a new era of precision medicine and enhanced healthcare delivery for individuals afflicted by vision-threatening conditions.
More: https://medicalxpress.com/news/2024-02-ai-outperforms-human-specialists-retina.html
