United States: According to new research, artificial intelligence could be used as an assistant to support Radiologists in checking the accuracy of their reports by acting as the responsible backup editor.
Study Details
The April 16 issue of the journal Radiology mentioned that the OpenAI GPT-4 model had about the same error detection accuracy as a group of radiologists benchmarked against data incorporated into radiology reports.
Promising Results
“This efficiency in detecting errors may hint at a future where AI can help optimize the workflow within radiology departments, ensuring that reports are both accurate and promptly available, thus enhancing the radiology department’s capacity to deliver timely and reliable diagnostics,” said lead researcher Dr. Roman Gertz, a radiology resident at University Hospital of Cologne in Germany, as reported by HealthDay.
According to Gertz, this is the first study that directly contrasts the performance of humans to that of GPT-4 in the context of radiology reports while also taking into account the factors of accuracy, speed, and cost-effectiveness.
Research Methodology
The study was conducted with the participation of Gertz and his colleagues to gather 200 radiology reports based on X-ray, CT scan, and MRI imaging between June and December 2023 at a single hospital.
In this study, the researchers intentionally inserted 150 errors in 100 of the reports; there were omissions and spelling errors, some passages became confusing, and other mistakes generally.
The researchers subsequently assigned both the GPT-4 and the group of six radiologists, including two senior radiologists, two attending physicians, and two residents, the same task of reading the reports and pointing out the mistakes.
AI vs. Human Performance
Consequently, researchers noted that humans had an error detection rate of 83%, compared with 89% for senior radiologists and 80% for attending radiologists and radiology residents.
Two weeks were not enough to show any statistical significance between human and GPT-4 performance.
AI automation has reduced the processing time for each report even less than the fastest human reader, and the average correction cost of each report was lower than that of the most cost-efficient radiologist.
Future Implications
The study reveals that AI can play a significant role in improving patient care, as it can help improve the accuracy of radiology reports and the speed with which they can be returned, Gertz said.
“The study addresses critical health care challenges such as the increasing demand for radiology services and the pressure to reduce operational costs,” Gertz said in a journal news release, as reported by HealthDay.
“Ultimately, our research provides a concrete example of how AI, specifically through applications like GPT-4, can revolutionize health care by boosting efficiency, minimizing errors, and ensuring broader access to reliable, affordable diagnostic services — fundamental steps toward improving patient care outcomes,” Gertz added.