Thursday, December 5 | 9:30 a.m.-9:40 a.m. | R3-SSMK11-1 | Room E450A
In this study, researchers from France found that an AI algorithm could find incidental vertebral fractures in cancer patients that weren’t initially detected.
Researchers from Gustav Roussy in France sought to compare the performance of a commercial AI algorithm with radiological assessment for detecting vertebral fractures in cancer patients, who face an increased risk for these fractures due to depletion of serum estrogen levels. They retrospectively applied CINA-VCF software (Avicenna.ai) to chest, abdomen, and pelvis CT exams acquired from September to December 2023 in adult patients being treated for stage IV cancer.
Next, two interventional radiologists with expertise in cancer treatment assessed the AI findings and determined if they would have treated severe fractures with vertebroplasty. Of the 1,501 patients included in the study, the AI software detected vertebral fractures in 501, including 436 true positives.
Furthermore, 81% of the AI-detected fractures were not described in the initial radiological report. These included 10 severe fractures, nine of which required an interventional treatment, according to the researchers.
“AI could assist radiologists to detect more incidental vertebral fractures in adult cancer patients, optimizing timely treatment and reducing associated morbidity and economic burden,” presenter El Mehdi Mniai, MD, and co-authors wrote. “Automated assessment of L1 trabecular attenuation offers potential for osteoporosis screening, facilitating early identification of at-risk individuals.”
Get all the details by sitting in on this presentation.