AI algorithm helps to spot overlooked fractures

Monday, November 27 | 1:40 p.m.-1:50 p.m. | M6-SSNPM01-2 | Room E351

In this talk, researchers will highlight the potential for AI software in detecting fractures that are often missed on radiographs.

Researchers led by presenter Vasantha Venugopal, MD, of AI firm Carpl.AI assessed the performance of commercial AI software Rayvolve (AZmed) in a study involving 459 cases from three institutions.

The cases, which included scaphoid, radial head, femoral neck, tibial plateau, Lisfranc, or avulsion fractures of the medial and lateral malleoli, were retrospectively evaluated by Rayvolve and two musculoskeletal radiologists.

In cases with mismatched results between the software and the radiologists, the AI software produced sensitivity of 83.9% and specificity of 93.5%. Meanwhile, the first radiologist had a sensitivity of 71.7% and a specificity of 92.1% on these cases. The second radiologist had sensitivity of 80.7% and a specificity of 80.9%.

“The AI system outperformed the radiologists in detecting commonly missed fractures, showcasing its potential for improving fracture detection in clinical practice,” the authors wrote. “The statistical analysis suggests that combining the strengths of AI and radiologists through a synergistic model could lead to even better diagnostic performance.”

The study highlights the benefits of incorporating AI into fracture detection workflows and makes the case for a synergistic approach, according to the researchers.

“In this model, radiologists and AI systems collaborate in a consensus review process, reevaluating cases where disagreement occurs,” the authors wrote. “This collaboration between AI and radiologists can lead to better patient outcomes by reducing the rate of missed fractures and improving overall diagnostic accuracy.”

Attend this session on Monday to learn more.

Page 1 of 2
Next Page