Loading...
By Kelly Young
Edited by David G. Fairchild, MD, MPH, and Lorenzo Di Francesco, MD, FACP, FHM
Some artificial intelligence computer programs outperformed pathologists in detecting lymph node metastases of breast cancer in a recent study. The findings appear in JAMA.
Researchers held a competition in which participants created computer algorithms to detect lymph node metastases. Their results were compared with findings from pathologists — one pathologist who took 30 hours to examine 129 whole-slide images, and 11 pathologists who had to perform the task in 120 minutes, which the authors say more accurately reflects clinical practice.
The top-performing programs all used deep learning-based systems that utilized AI to extract meaningful patterns. The top algorithm was more accurate than the pathologists with a time constraint and comparable to the pathologist with no time limit.
An editorialist concludes: "AI may be just what pathology has been waiting for. While still requiring evaluation within a normal surgical pathology workflow, deep learning has the opportunity to assist pathologists by improving the efficiency of their work, standardizing quality, and providing better prognostication. Like electron microscopy, immunohistochemistry, and molecular diagnostics ahead of AI, there is little risk of pathologists being replaced."
In a separate JAMA study, researchers report that another deep-learning system had high sensitivity and specificity in detecting diabetic retinopathy and other eye diseases.
Comment
LINK(S):
JAMA article on lymph node metastases (Free abstract)
JAMA editorial (Subscription required)
JAMA article on diabetic retinopathy (Free abstract)
Background: Physician's First Watch coverage of machine learning to detect diabetic retinopathy (Free)