Artificial intelligence could diagnose breast cancer better than doctors after being trained to read MRI scans
- Researchers at UCLA trained its AI system using 240 breast biopsy images
- Its reading were compared against diagnoses made by 87 pathologists
- The machine was on par with doctors on all, but more accurate at deciphering between two types in particular
- DCIS and atypical hyperplasia look similar but they are very different
- DCIS requires radiation therapy and hormonal therapy and sometimes a mastectomy
- Atypical hyperplasia is a precancerous lesion that should be removed but should not be followed by other treatment
A computer could be better than a doctor at diagnosing certain types of breast cancer, new research suggests.
Researchers at the University of California, Los Angeles, trained an artificial intelligence system using 240 biopsy images of different breast cancer types, and tested it against 87 pathologists.
The machine performed more or less as well as doctors at detecting and classifying all of the breast cancers.
However, it was better at making one crucial distinction: telling the difference between DCIS (ductal carcinoma in situ) and atypical hyperplasia, two very different types of breast cancer, that require different treatments, but look similar.
AI may be better than a pathologist at diagnosing some cancers, UCLA research suggests
‘Medical images of breast biopsies contain a great deal of complex data and interpreting them can be very subjective,’ said Dr Joann Elmore, lead author of the study published in the JAMA Network Open journal.
‘Distinguishing breast atypia from ductal carcinoma in situ (DCIS) 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.’
People diagnosed with DCIS are recommended a mastectomy, or a lumpectomy followed by radiation therapy.
Atypical hyperplasia, however, is a precancerous condition that affects the breast. Doctors would recommend removing the cells, and perhaps prescribe the hormone therapy tamoxifen, but would not order a mastectomy.
‘These results are very encouraging,’ added Dr Elmore, a professor at the University of California, Los Angeles.
‘It is critical to get a correct diagnosis from the beginning so that we can guide patients to the most effective treatments.’
With further improvements, researchers hope AI could be a vital tool in aiding pathologists, and are looking at how it could be used to diagnose melanoma next.