Researchers Develop Cancer-risk Detecting AI Software To Revolutionize Early Breast Cancer Prevention

While more people are engaging in various researches on ways to detect cancer as early as possible, technological advancements are not far behind. A highly potential program is currently being studied intensively for this top purpose.

One of two physically healthy women are set to endure those needed screening tests yearly, or find out that they have cancer, PDDNet relays. However, studies have further noted on these tests giving out false-positive results.

To combat the unnecessary anxieties a woman faces while waiting on such results, the Houston Methodist Research Institute has shared information that artificial intelligence can already extend such advancement in the field of cancer studies. Although the AI has yet to be named, the cancer-detecting software has been found to review patients' charts and translate them into relevant diagnostic data at a 99 percent accuracy rate.

This is another great breakthrough in cancer studies, as manual physician evaluations of patient charts could take up to more than 60 hours. The AI software can actually performed the same review proceedings faster.

Study chairman Dr. Stephen T. Wong of the Department of Systems Medicine and Bioengineering said that the AI's precision in reviewing a number of these records in a shorter time would allow them to predict breast cancer risks effectively via mammogram screening, avoiding unnecessary biopsies, according to Breast Cancer News.

Through Dr. Jenny C. Chang of the Cancer Center at the Houston Methodist Research Institute, the AI software has been proven efficient in evaluating up to more than 500 breast cancer patients' data. This led the researchers to determine how its NLP (natural language processing) algorithm, or the AI software, is to get key traits for those patients meeting the criteria.

Through the AI software, the team was also able to determine the degree of probability for patients at greatest risk for breast cancer. Check out how it works in the video below: