Researchers develop new AI system to predict breast cancer risk faster, could cut biopsies
Researchers at Houston Methodist, including those of Indian origin, have developed an artificial intelligence (AI) software that can help doctors to quickly and accurately predict breast cancer risk.
Significantly, researchers claim that the new technique could also potentially lead to a reduction in unnecessary biopsies.
The computer software intuitively translates patient charts into diagnostic information at 30 times human speed and with 99 per cent accuracy.
Researchers used the AI software to evaluate mammograms and pathology reports of 500 breast cancer patients.
"This software intelligently reviews millions of records in a short amount of time, enabling us to determine breast cancer risk more efficiently using a patient's mammogram," said Stephen Wong, one of the researchers from Houston Methodist Research Institute in the US.
"This has the potential to decrease unnecessary biopsies," Wong noted.
The software scanned patient charts, collected diagnostic features and correlated mammogram findings with breast cancer subtype.
Clinicians used results, like the expression of tumour proteins, to accurately predict each patient's probability of breast cancer diagnosis.
Manual review of 50 charts took two clinicians 50-70 hours. AI reviewed 500 charts in a few hours, saving over 500 physician hours.
Currently, when mammograms fall into the suspicious category, a broad range of three to 95 per cent cancer risk, patients are recommended for biopsies.
Over 1.6 million breast biopsies are performed annually in the US, and about 20 per cent are unnecessarily performed due to false-positive mammogram results of cancer free breasts, estimates the American Cancer Society.
The team hopes this artificial intelligence software will help physicians better define the per cent risk requiring a biopsy, equipping doctors with a tool to decrease unnecessary breast biopsies.
The study is reported in a paper published online in the journal Cancer.
(With agency inputs)