DOCSF 2026: Lightning Theater

Recorded live at DOCSF 2026 in San Francisco, six five-minute talks from the DOCSF Science Lightning Theater test AI and digital tools against real clinical questions, from patient education to imaging to discharge planning. Moderated by Fabrizio Billi of UCLA's Biomedical Innovation and Motion Intelligence Group.

You'll hear from:

• Usman Ali: whether ChatGPT, Gemini and Copilot answer shoulder surgery questions as accurately in Spanish as in English, and the one question that tripped them up.
• Shragvi Balaji (Baylor College of Medicine): a model built on more than 5,000 ACL reconstructions that updates its recovery prediction at each visit, and what it reveals about early non-responders.
• Alidad Ghiassi, MD (HIA Technologies): AI-generated avatar videos that turn consent documents into interactive education, and the gap between what patients think they know and what they do.
• George Rublev, MD (Yale School of Medicine): whether semaglutide before total knee replacement reduces opioid use afterward, in a matched national cohort.
• Sahil Sethi (University of Chicago Pritzker School of Medicine): an AI model trained on the first public shoulder MRI dataset that helps surgeons catch Bankart lesions they missed.
• Brian Shear, MD (University of Maryland Medical Center): how the mobility data an iPhone collects passively can predict whether a fracture patient goes home or to rehab.