Advertisement

AI identifies precancerous cervix better than human experts, report says

By Tauren Dyson

Jan. 10 (UPI) -- A new artificial intelligence method could be key for future diagnosis of cervical cancer.

The algorithm, developed by researchers at the National Cancer Institute, was 91 percent accurate in diagnosing cancer in women during a study in Costa Rica, according to findings published Thursday in the Journal of the National Cancer Institute.

Advertisement

"Our findings show that a deep learning algorithm can use images collected during routine cervical cancer screening to identify precancerous changes that, if left untreated, may develop into cancer," Mark Schiffman, a researcher at the NCI's Division of Cancer Epidemiology and Genetics and study senior author, said in a news release. "In fact, the computer analysis of the images was better at identifying precancer than a human expert reviewer of Pap tests under the microscope (cytology)."

The researchers created the algorithm using more than 60,000 cervical images from an NCI archive of photos collected as part of a cervical cancer screening study conducted in Costa Rica during the 1990s. They digitized the photos and trained the deep learning algorithm to differentiate cervical diagnosis that required treatment from those that didn't require any.

Advertisement

The deep learning algorithm diagnosis was compared to images from patients that had already been diagnosed during the Costa Rica study.

The current human eye method to diagnose cervical precancer relies on a health worker to dilute acetic acid to the cervix and then examine it for "aceto whitening," an indicator of possible disease. This naked eye method remains popular due to how inexpensive and convenient it is. But it needs tweaking due to its inaccuracies, researchers say.

The new method allows health workers to examine a cervix with a cell phone or other camera device.

According to the National Institutes of Health, cervical cancer is the third most common cancer for women globally.

"When this algorithm is combined with advances in HPV vaccination, emerging HPV detection technologies, and improvements in treatment, it is conceivable that cervical cancer could be brought under control, even in low-resource settings," said Maurizio Vecchione, executive vice president of Global Good.

Latest Headlines