While radiographers currently screen women for breast cancer with mammograms, alarming new research claims robots and artificial intelligence could actually be better at detecting and diagnosing the cancer than humans. Researchers from the University of California, Los Angeles (UCLA) have developed an artificial intelligence system that could help health professionals and pathologists read biopsies more accurately and allow them to better identify breast cancer.
The newly developed system helps interpret medical images used to diagnose breast cancer that can be extremely difficult for humans to classify, with research published in the JAMA Network Open claiming the new system makes a diagnosis that is almost as accurate – or even better – than experienced pathologists. Getting a correct diagnosis from the very beginning is vital so health professionals can guide patients to the most effective treatment.
Previous research shows diagnostic errors occur in as many as one in six women with a non-invasive type of breast cancer called ductal carcinoma in situ (DCIS) and that incorrect diagnoses are given in half of biopsy cases of breast atypia – abnormal cells that are associated with a higher risk of breast cancer. In Australia, women aged between 50 and 74 are encouraged to have a mammogram every two years as part of the Department of Health’s BreastScreen program.
If changes in the breast are detected through the initial screening, an ultrasound or biopsy may then be requested, while CT scans or MRI scans may be necessary if cancer is diagnosed.
“Medical images of breast biopsies contain a great deal of complex data and interpreting them can be very subjective,” the study’s senior author Joann Elmore said in a statement. “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 the same case a year later.”
Researchers said the artificial intelligence could provide more accurate readings consistently because it draws from a large data set and can recognise patterns in the samples that are associated with cancer, but aren’t always easy for humans to spot. The research team tested the system by feeding 240 biopsy images into a computer and trained it to detect patterns associated with an array of breast legions including benign, atypia, DCIS and invasive breast cancer. Three expert pathologists then made a formal diagnosis.
Researchers compared the readings from the system with diagnoses made by 87 practicing pathologists in the United States and found that while the artificial intelligence program came close to performing as well as humans when it came to differentiating cancer from non-cancer cases, the program actually outperformed humans when separating DCIS from atypia.
Despite being one of the greatest challenges in diagnosing breast cancer, the system correctly determined whether scans showed DCIS or atypia more than doctors could, with Elmore saying that the results are “very encouraging” and that the computer-based automated approach “shows great promise”.
Researchers are now working on training the system to diagnose melanoma skin cancer, but because the system is still in the testing phase, it’s important to pay attention to your breasts and keep up with screening visits. Self-examination at home is important and may pick up some changes, but breast screenings can detect cancer before you or even a health professional notices any changes.
Screenings are encouraged every two years, but checking the breasts between visits is important and any lumps, bumps or changes need to be addressed with a GP as soon as possible. Catching cancer early means treatment is more likely to be effective and less invasive, but also reduces the risk of cancer spreading to other parts of the body.
Important information: The information provided on this website is of a general nature and information purposes only. It does not take into account your personal health requirements or existing medical conditions. It is not personalised health advice and must not be relied upon as such. Before making any decisions about your health or changes to medication, diet and exercise routines you should determine whether the information is appropriate in terms of your particular circumstances and seek advice from a medical professional.