RSNA Margulis Prize Recognizes AI Research in X-Ray Imaging

November 11, 2022 — The 2022 Alexander R. Margulis Award for Scientific Excellence from the Radiological Society of North America (RSNA) will be awarded to Ali Guermazi, MD, Ph.D., M.Sc., for the radiology paper entitled “Improving Radiographic Performance and Efficiency of Fracture Recognition Using Artificial Intelligence.

Named in honor of Alexander R. Margulis, MD, a distinguished researcher and inspiring visionary in the field of radiology, this annual award recognizes the best original scientific paper published in RSNA’s flagship journal, Radiology.

“This year’s Margulis Award recognizes the growing importance of artificial intelligence in our field. The authors studied fracture detection by 24 radiologists and clinicians with and without AI,” said senior editor David A. Bluemke, MD, Ph.D. “10% better fracture detection with AI, while reducing the time of radiologists. This study validates the steady increase in the use of AI tools that are becoming a common part of many clinical practices, especially in musculoskeletal radiology.

In the Food and Drug Administration (FDA) registration study, researchers retrospectively analyzed 480 X-ray exams from various US hospitals.

“AI can be a powerful tool to help radiologists and other physicians improve diagnostic performance and increase efficiency, while potentially improving the patient experience at the time of the hospital or clinic visit. clinic,” said Dr. Guermazi, director of the Quantitative Imaging Center, professor of radiology. and Medicine, and Associate Dean of the Office of Diversity at Boston University’s Chobanian & Avedisian School of Medicine and Chief of Radiology at the VA Boston Healthcare System.

The researchers included x-rays of the limbs, pelvis, spine and rib cage. The examination group included adults over the age of 21 with evidence of trauma and fracture prevalence of 50%. There were 240 patients with a total of 350 fractures and 240 patients without fractures.

The studies were reviewed twice by 24 American board-certified readers from six different specialties, including radiology, orthopedic surgery, rheumatology, emergency medicine (including physicians and physician assistants), and family medicine. .

According to Dr. Guermazi, the readings were taken with and without commercially developed software using an algorithm trained on precisely annotated X-ray images from multiple institutions, acquired on a wide variety of systems. The readers had a period of one month between the two analyses.

“The results of the study showed an absolute gain in sensitivity in detecting fractures of 10.4% with the aid of the software, with the software showing a sensitivity of 75.2% compared to 64.8% without the aid. software,” Dr. Guermazi said. The results also revealed an absolute gain in specificity – from 90.6% to 95.6% – for detecting fractures with the assistance of software.

Without being surprised by the sensitivity of the algorithm, Dr. Guermazi did not expect a gain in specificity.

“Computer-aided detection systems can be easily sensitive but usually result in a significant loss of specificity. Here, the algorithm also helped reduce false positive rates,” he said. “The time saving was a nice surprise, given that the algorithm brings additional information to look at in addition to the native images. It was not clear that the algorithm would speed up the interpretation time.

Dr. Guermazi noted that one of the biggest challenges the team faced during the study was training 24 readers from diverse backgrounds to read with AI. Despite this challenge, he said readers thought using the AI ​​algorithm was simple, user-friendly and extremely intuitive.

Clinical validation studies are underway focusing on specific parts of the body where, according to Dr. Guermazi, the gold standard is being established using CT/MRI to assess the algorithm’s ability to detect radiographically visible and occult lesions.

“Ultimately, I think my fellow radiologists will join in seeing AI as a friend rather than a foe,” Dr. Guermazi said. “As it becomes clearer that it can beat the human eye on certain specific and repetitive or tedious tasks, AI will be seen as an excellent complement to heavy clinical workflow.”

The Margulis Prize will be presented at the 108th RSNA Scientific and Annual Meeting (RSNA 2022) in Chicago, November 27-December 27. 1.

For more information: www.rsna.org

Find more RSNA22 coverage here

Comments are closed.