Cairo: Mai Kamal El-Din
A recent study conducted by researchers at Osaka Metropolitan University School of Medicine has found that ChatGPT, an AI-powered chat system, can diagnose brain tumors from MRI images with an accuracy similar to that of certified neuro-radiologists. The results of this study were published in the journal European Radiology, highlighting the increasing application of artificial intelligence technologies in the medical field.
ChatGPT (Generative Pre-trained Transformer) is a sophisticated chatbot developed by OpenAI, utilizing large language models that are enhanced with artificial intelligence capabilities. Researchers believe that these models are increasingly being employed across various scientific disciplines, including radiology. Such models, like ChatGPT, can assist in making medical diagnoses and providing alternative opinions.
Study Methodology
To assess the medical potential of ChatGPT, researchers provided the system with 150 reports on preoperative brain MRI scans. A comparison was made between the performance of ChatGPT and that of two certified neuro-radiologists and three general radiologists, all of whom evaluated the presence of brain tumors.
The results showed that ChatGPT achieved a diagnostic accuracy of 73%, while the neuro-radiologists and general radiologists had accuracies of 72% and 68%, respectively. It was noted that ChatGPT’s performance varied depending on the author of the clinical report. The accuracy reached 80% when the report was written by a neuro-radiologist and dropped to 60% when assessed by general radiologists.
Future Perspectives
These findings are significant as they indicate that artificial intelligence can serve as an effective tool to support doctors and alleviate their workload. Researchers plan to explore the use of large language models in other areas of diagnostic imaging in the future, which could enhance diagnostic accuracy and reduce the burden on medical professionals.
With rapid advancements in medical technology, integrating artificial intelligence into medical procedures has the potential to transform the practice of medicine, leading to more effective and accurate healthcare delivery for patients.
Scientists look forward to further research that could deepen their understanding of AI capabilities and how to apply them in various clinical environments. As this technology continues to advance, we can expect improvements in disease diagnosis and treatment, enhancing the overall effectiveness of the healthcare system.
