FocalNet, AI System from UCLA as Accurate as a Radiologist

MRI machine and screens

UCLA Researchers introduce results from their new artificial intelligence (AI) system and find that it performs as well as experienced radiologists in detecting prostate cancer.

Back in the Spring UCLA researchers developed an AI system to help radiologists improve their ability to diagnose cancer. Called FocalNet, it helps identify and predict the aggressiveness of the disease-evaluating magnetic resonance imaging, or MRI, cans, and it doe so with nearly the same level of accuracy as experienced radiologists. In tests, FocalNet was 80.5 percent accurate in reading MRIs, while radiologists with at least a decade of experience were 83.9 percent accurate.


FocalNet is an AI neural network that uses an algorithm that comprises more than a million trained variables; it was developed by UCLA researchers. The team trained the system by having it analyze MRI scans of 417 men with prostate cancer; scans were fed into the system so that it could learn to assess and classify tumors in a consistent way and have it compare the results with actual pathology specimen. Researchers compared the AI system results with readings by UCLA radiologists who had more than 10 years of experience.


This research, led by UCLA, suggests that an AI system could save time and potentially provide diagnostic guidance to less-experienced radiologists.

Lead Research/Investigators

Kyung Sung, assistant professor of radiology

Dr. Steven Raman, UCLA clinical professional of radiology

Dr. Dieter Enzmann, chair of radiology at UCLA

Ruiming Cao, UCLA Graduate Student