AIMI Research Meeting: Working with and Teaching Nediser, the A.I. Radiology Resident - Bao Do, MD

Drs. Do and Fang are developing Nesider, an A.I. "radiology resident", that is integrated into workflow at Palo Alto VA to pre-dictate MSK x-rays. Nediser started residency training 1/2021 and has been trained with 168,000 and interpreted 37,000+ clinical images. Nediser uses basic NLP to populate comparison dates and vision to select normal templates, detect hardware, evaluate patella alignment and leg length and angle discrepancy, and measure Cobb angles. The purpose of this talk is to share our experiences in developing Nediser, workflow and safety challenges of using AI, and how Nediser is being continuously trained as it embarks its "residency" along levels of radiology autonomy. Nediser will never replace radiologists, but it reduces dictation effort and occasionally dictates a few studies in a row without disengagements (= no human intervention).

 

Bao Do is a Clinical Associate Professor of Radiology at Stanford/VA Palo Alto and Chief of Radiology Informatics for the Sierra Pacific Network VAs (VISN 21). Dr. Do works with Dr. Payam Massaband and Dr. Charles Fang to build the radiology department of the future at VA Palo Alto, leveraging informatics and A.I. to automate processes such as study protocolling and image interpretation.