Radiology Advances Podcast | RSNA

Episode 22: Can LLM-generated summaries help patients understand lung cancer screening reports?

This episode discusses a study from the University of California, San Francisco in the United States that tested whether GPT-4o-generated patient-friendly summaries improve comprehension of lung cancer screening CT reports. In a within-subjects survey of 1,815 adults across Lung-RADS 1, 2S, and 4B vignettes, the summaries significantly improved objective comprehension and reduced anxiety for all three report types. Largest gains were in participants with low self-rated English and health literacy. These findings support using LLM summariesas a potential health-equity tool, while highlighting the unmet patient need for personalized next-steps guidance. Self-reported comprehension of large language model-generated summaries of lung cancer screening reports: a vignette survey. Serna et al. Radiology Advances, 2026, 3, umag008. [https://doi.org/10.1093/radadv/umag008]

Episode 21: Can AI catch cardiomegaly on chest CTs ordered for other reasons?

This episode explores a study from the University of Texas Southwestern Medical Center and MD Anderson Cancer Center in the United States that clinically validates an FDA-cleared AI tool for measuring total cardiac volume on non-contrast, non-gated chest CT. Across 307 patients with paired echocardiography, the AI discriminated normal from abnormal cardiac volume with an AUC of 0.81 in men and 0.77 in women, and far outperformed routine radiologist sensitivity for cardiomegaly. The tool offers a tunable, reproducible opportunistic screening layer on chest CT's already being performed. Radiology Advances, 2026, 3, umag013. Fan et al. [https://doi.org/10.1093/radadv/umag013]

Episode 20: Minimum Data for Maximum Accuracy

This episode explores a study from the Emory Sports Performance and Research Center and the University of Lausanne that determined how few annotated MRI exams are needed to train a reliable deep learning model for thigh muscle segmentation. Using the nnU-Net framework with incrementally larger training sets, the researchers found that just 20 high-quality annotated subjects produced clinically acceptable segmentation across 14 thigh muscles, with biomarker agreement virtually indistinguishable from expert manual segmentation. All tools and trained models have been made openly available. Optimizing MRI annotation workflows for high-accuracy deep learning thigh muscle segmentation in athletes. Slutsky-Ganesh et al. Radiology Advances, 2026, 3, umag005 [https://doi.org/10.1093/radadv/umag005]

Episode 19: Leveraging Federated Learning to Supplement an AI Learning Dataset

This episode discusses a study from UCLA in the United States that used federated learning to train a deep learning model for automatic segmentation and quantification of visceral and subcutaneous abdominal fat in children using free-breathing 3D MRI. By leveraging a larger adult dataset alongside a small pediatric cohort, the model achieved strong agreement with expert manual segmentation in under three seconds per patient. Cross-cohort federated learning for pediatric abdominal adipose tissue segmentation and quantification using free-breathing 3D MRI. Zhang et al. Radiology Advances, 2026, 3, umag002 [https://doi.org/10.1093/radadv/umag002]

Episode 18: Ferumoxytol MRI to detect slow gastrointestinal bleeding

This episode reviews a proof-of-concept study from Mayo Clinic Minnesota on the use of ferumoxytol-enhanced MRI for detecting gastrointestinal bleeding after a comprehensive conventional workup has been negative. We examine how this blood pool agent's prolonged intravascular half-life addresses the diagnostic challenge of slow and intermittent GI bleeding, and discuss the clinical implications for patient management. Feasibility of ferumoxytol-enhanced MRI for detection of gastrointestinal bleeding when conventional evaluation is negative. Wells et al. Radiology Advances, 2026, 3, umaf043. [https://doi.org/10.1093/radadv/umaf043]