Artificial Intelligence Can Now Diagnose Breast Cancer
Breast Cancer, an unfortunately common form of cancer that affects over 300,000 people per year worldwide, may be mitigated by the increasing strength of artificial intelligence. Specifically, researchers discovered an artificial intelligence system that aids pathologists to analyze biopsies with greater accuracy, and help them better detect and identify breast cancer. This process is a challenge for human doctors, but fortunately, artificial intelligence’s emergence may eliminate this challenge entirely.
This new system, which was published in the journal ‘JAMA Network Open,’ has shown its merits by viewing the same medical images viewed by doctors in the process of identifying breast cancer, but being noticeably more accurate in their conclusions as to whether or not the individual in the photo is at risk of this cancer. This is a profound similarity with previous research on this subject performed by Google. Google AI recently claimed a 99% accuracy in metastatic breast cancer detection - a form of detection that lies in viewing an individual’s lymph nodes to determine a possibility of breast cancer. However, this differs from google’s attempt in that it can objectively aid doctors in becoming more accurate overall.
While the researchers’ system is far from autonomous, it may become a tool to help doctors be more accurate, said Linda Shapiro, a computer science professor at the University of Washington. “We’re billing it as something like computer-aided diagnostics, where it can make suggestions and show pathologists what it’s thinking and why,” she said.
On the JAMA journal, the algorithms were mentioned in a comparison with human pathologists. More so than just detecting breast cancer, researchers wanted to know if the artificial intelligence program could perform basic medical acts. And this task was to identify ductal carcinoma in situ from atypia — two conditions that are fairly commonly diagnosed incorrectly. The results were impressive: accuracy rates were 89 percent for the system and 70 percent for the pathologists. The computer-based approach had similar accuracy to the group of 87 doctors in differentiating cancer from non-cancer tissues.
Despite this indisputable success, there exist a handful of hurdles between the researchers and the reliability of AI in this process. Because unfortunately, the data necessary for cancer diagnostics is not at the level it could be. To create the data-sets necessary for artificial intelligence to function properly, however, researchers at universities in the United States relied on recognized Doctor Jamen Bartlett, who labeled dozens of complex images over the course of nine months.
Fortunately, the conclusions and results indicate positivity for the future of the medical sector. With the low accuracy currently held by pathologists in these predictions, artficial intelligence can make a triumphant and heroic role in these tasks. It is not to say that the pathologists, doctors and medical staff will be replaced entirely by artificial intelligence, however, but rather complimented heavily to make up for their failures.