In recent years, significant advancements in artificial intelligence (AI) have brought speed and convenience to our lives. We can now clearly see the effects of AI in many areas of daily life, including the healthcare sector. According to recent reports, GPT-4, a large language model developed by OpenAI, has outperformed doctors in interpreting brain tumor reports more accurately. Here are the details…
GPT-4 Reads Preoperative MRI Reports Better
A study published in the European Journal of Radiology revealed that GPT-4 surpassed radiologists in identifying brain tumors using preoperative MRI reports. This research showed that the large language model made more accurate predictions than doctors.
As part of the study, 150 preoperative brain MRI reports were first translated from Japanese to English, and then GPT-4 was asked to provide diagnoses. The same reports were also reviewed by five different radiologists.
OpenAI announced the number of users for its popular chatbot. It is noticeable that there has been a serious increase.
GPT-4 achieved a 74% accuracy rate in its final diagnoses, while radiologists’ accuracy ranged between 65% and 79%. This study marks the first time that real clinical radiology reports were interpreted by a language model, as previous tests used mock reports.
Researchers highlighted the important role large language models play in diagnosing medical conditions, showcasing the potential of such technologies in healthcare. Additionally, this study further demonstrated the reliability of GPT-4.
What do you think about this? Feel free to share your thoughts in the comments section below.
