From Notes to Data: NLP and Transformers in Radiology

Dr. Linda Chu speaks with Dr. Michael Trambert and Florent Saint-Clair about the unique workflow challenges in mammography and breast imaging. They explore clinical hurdles, the promise and pitfalls of AI, and how interoperability solutions can streamline efficiency and improve patient care. Sponsored by Dicom Systems. 

Dr. Linda Chiu unpacks how natural language processing and large language models are transforming radiology, with insights from a review by Dr. Felipe Kitamura and colleagues. From tokenization to transformers, she explores both the promise and challenges of applying these powerful AI tools to clinical practice. Texts Are More than Notes, They Are Data: A Glimpse intoHow Machines Understand Text. Kitamura et al. Radiology 2025; 316(2):e243217.

Dr. Refky Nicola speaks with Dr. Bari Dane, Associate Professor of Radiology at NYU, about the new SAR consensus recommendations for defining small bowel Crohn’s disease strictures on CT and MR enterography. They explore updated imaging criteria, the challenges of distinguishing inflammation from fibrosis, and emerging roles for ultrasound and AI in diagnosis and treatment response. SAR Consensus Recommendations for Defining Small BowelCrohn Disease Strictures at CT and MR Enterography. Dane et al. Radiology 2025; 316(1):e243123.

Dr. Lauren Kim speaks with Dr. Anna Rita Larici about the international consensus guidelines for imaging and managing post-COVID-19 residual lung abnormalities. They discuss the importance of standardized definitions, CT imaging protocols, and unified terminology to guide diagnosis and follow-up care for patients with long COVID. Best Practice: International Multisociety Consensus Statement forPost–COVID-19 Residual Abnormalities on Chest CT Scans. Yoon et al. Radiology 2025; 316(1):e243374.

Dr. Linda Chu speaks with Dr. Andrew Trout, Professor of Radiology and Director of Clinical Research at Cincinnati Children’s, and Dr. Erin Angel, Vice President of Research and Scientific Affairs at GE HealthCare. They discuss the unique challenges of pediatric imaging and how collaboration and technology are advancing care for young patients while improving imaging for all. Sponsored by GE HealthCare.

Dr. Linda Chu and Dr. Sid Dogra speak with Jessie Gommers and Dr. Ioannis Sechopoulos about their study on how AI decision support influences radiologist performance and visual search in screening mammography. They explore key findings on sensitivity, specificity, reading time, and the future role of visual tracking in optimizing human–AI collaboration. Influence of AI Decision Support on Radiologists’ Performance and Visual Search in Screening Mammography. Gommers and Verboom et al. Radiology 2025; 316(1):e243688.  

Dr. Refky Nicola speaks with Dr. Perry Pickhardt about his study comparing CT colonography and multi-targeted stool DNA testing for colorectal cancer screening. They explore differences in sensitivity, specificity, cost-effectiveness, and strategies to optimize detection while minimizing invasiveness. CT Colonography versus Multitarget Stool DNA Test forColorectal Cancer Screening: A Cost-Effectiveness Analysis. Pickhardt et al. Radiology 2025; 315(3):e243775.

In this episode, we explore how nonprofit organizations like LBDA are working alongside industry partners to shape the evolving landscape of dementia biomarkers—bridging scientific innovation with real-world care. Featuring an engaging conversation with Dr. Sudhir Sivakumaran, Dr. Kathleen Poston, and Dr. Dustin Dunham on clinical utility, patient-centered research, and the road to broader adoption of biomarkers in Lewy body dementia, Alzheimer’s disease and related disorders. This episode is sponsored by GE HealthCare

In this episode, Dr. Linda Chu explores a major Radiology consensus statement on optimizing CT angiography for suspected pulmonary embolism. The discussion covers advanced imaging techniques, key considerations for special populations, and standardized reporting practices to improve diagnostic clarity and patient outcomes. Optimal Approach to Performing and Reporting ComputedTomography Angiography for Suspected Acute PulmonaryEmbolism: A Clinical Consensus Statement of the ESC Working Groupon Pulmonary Circulation & Right Ventricular Function, the FleischnerSociety, the Association for Acute Cardiovascular Care (ACVC) andthe European Association of Cardiovascular Imaging (EACVI) of theESC, Endorsed by European Respiratory Society (ERS), Asian Societyof Thoracic Radiology (ASTR), European Society of Thoracic Imaging(ESTI), and Society of Thoracic Radiology (STR). Radiology 2025; 315(3):e243833.

Dr. Linda Chu speaks with Dr. Ramin Khorasani about targeted interventions that reduced ambiguous radiologist recommendations for additional imaging while dramatically improving the clarity and follow-through of actionable recommendations. They explore how a structured system of care, closed-loop communication, and leadership engagement can advance high-value, patient-centered care in radiology. This episode is sponsored by Mayo Clinic.  Cumulative Effect of Targeted Interventions on Radiologist Recommendations for Additional Imaging. Abbasi et al. Radiology 2025; 315(3):e243750.