UCLA researchers have developed an artificial intelligence system which can diagnose breast cancer more accurately. The study was published in JAMA Network Open, and its authors are Ezgi Mercan of Seattle's Children Hospital, Sachin Mehta and Linda Shapiro of the University of Washington, Jamen Bartlett of Southern Ohio Pathology Consultants and Donald Weaver of the University of Vermont. The medical images which sometimes could be difficult to read for the naked human eye could be read almost accurately with the help of the new system.
Dr. Joann Elmore who is the study’s senior author and a professor of medicine at the David Geffen School of Medicine at UCLA, said, “It is critical to get a correct diagnosis from the beginning so that we can guide patients to the most effective treatments. Ezgi Mercan of Seattle's Children Hospital is the study's first author.
Back in 2015, Elmore showed the differences between the opinions of pathologists on their interpretation of breast cancer. The study also revealed that diagnostic errors occurred in about one out of every six women who had ductal carcinoma in situ (a noninvasive type of breast cancer). These diagnoses were given in about half of the biopsy of cases of breast atypia.
Elmore further added, “Medical images of breast biopsies contain a great deal of complex data and interpreting them can be very subjective. Distinguishing breast atypia from ductal carcinoma in situ is important clinically but very challenging for pathologists. Sometimes, doctors do not even agree with their previous diagnosis when they are shown in the same case a year later.”
The scientific data could be extracted more accurately with the help of artificial intelligence because it draws the data from a large data set, and recognizes the patterns in the samples associated with cancer. For the system to recognize patterns, a computer was fed 240 breast biopsy images. The correct diagnoses based on the images were conducted by a consensus among three expert pathologists.
The reading done by the computer was compared to the diagnoses made by 87 pathologists. The AI system turned out to be better than the doctors in identifying and differentiating the DCIS from atypia. The system had sensitivity between 0.88 and 0.89, while the pathologists' average sensitivity was 0.70.
Commenting the results Elmore said, “These results are very encouraging. There is low accuracy among practicing pathologists in the U.S. when it comes to the diagnosis of atypia and ductal carcinoma in situ, and the computer-based automated approach shows great promise.”
The researchers are now working on training the system to diagnose melanoma.