Detecting breast cancer cells with artificial intelligence produces international award


    The joint project between Southwest Research Institute and UT Health San Antonio trained an algorithm to identify cancel cells in slides. The original slide is on the left with the computer-generated model on the right. (SBG photo)

    SAN ANTONIO - Humans get bored. Computers don't.

    That's the essence of a time-saving breakthrough that only happened after two local companies- Southwest Research Institute and UT Health San Antonio - combined forces.

    They recently captured first-place honors in a major international breast cancer cell detection competition.

    "This is one of the ironies in life that sometimes you don't know who the other players are," said Dr. Bradley Brimhall, a pathologist at UT Health. "You don''t know the talent in your own back yard."

    By training a computer algorithm to analyze cell images and determine which were cancerous. their joint project beat 100 others to win the Breast Path Q: Cancer Celluarity Challenge.

    The joint project between Southwest Research Institute and UT Health San Antonio trained an algorithm to identify cancel cells in slides. The original slide is on the left with the computer-generated model on the right. (SBG photo)

    "It was really exciting to see that we got first place. and I'm not surprised because our engineers are the best," said Hakima Ibaroudene, engineer and challenge leader at SwRI. "Our group has decades of experience in using computer vision for autonomous vehicles and industrial robotics."

    The key was to take that knowledge and apply it to breast cancer cell research. Enter UT Health, which taught Southwest's engineers what to look for.

    "What we were able to do is train our algorithm, to teach the algorithm how to recognize certain images." Ibaroudene said. " UT Health taught us to recognize the images and where the cancer is and we were able to feed that to the algorithm. "

    Hakima Ibaroudene is an engineer at Southwest Research Institute and project leader. (SBG photo)

    Instead of doctors spending long hours scanning slides of potentially cancerous cells, a machine can now perform that same task.

    "If you are taking mundane tasks off of a physician's plate of work to do, that's going to be popular with everybody." Brimhall said. "You're dealing with tremendous amounts of information.

    "We're looking for needles in a haystack. We can't be vigilant 100 percent of the time. But a machine can be hyper-vigilant all the time.. The computer has an infinite attention span. It doesn't get bored."

    The trained algorithm could also put San Antonio on the global map when the winners present their findings in San Diego next week.

    "This is where things are going in medicine i believe." Brimhall said, mentioning collaboration with not just engineers but also businesses.

    Dr. Bradley Brimhall is a pathologist at UT Health San Antonio. (SBG photo)

    Moving forward, the same technology might be used to find leukemia or prostate cancer cells.

    News In Photos

      Loading ...