NIH researchers develop AI technique to detect cervical precancer

11 January 2019 (Last Updated December 23rd, 2019 10:23)

A research team led by the National Institutes of Health (NIH), in alliance with Global Good, has developed an artificial intelligence (AI)-based approach to identify cervical precancer, even in low-resource settings.

NIH researchers develop AI technique to detect cervical precancer
Medical images from the National Cancer Institute (NCI) were used to train the algorithm. Credit: rawpixel.

A research team led by the National Institutes of Health (NIH), in alliance with Global Good, has developed an artificial intelligence (AI)-based approach to identify cervical precancer, even in low-resource settings.

Dubbed automated visual evaluation, the new technique involves a computer algorithm designed to analyse digital images of the cervix and accurately detect precancerous changes requiring medical attention.

The researchers used comprehensive datasets to ‘train’ the algorithm to recognise patterns in complex visual inputs such as medical images.

“When tested, the computer algorithm was found to be better than a human expert reviewing Pap tests at detecting cervical precancer.”

More than 60,000 cervical images from the National Cancer Institute (NCI) were used to train the algorithm to differentiate between cervical conditions that need treatment and those that do not.

Health workers will only require a camera device such as a cell phone for cervical screening. In addition, the new method can be performed with minimal training, allowing it to be used even in regions with limited health care resources.

When tested, the computer algorithm was found to be better than a human expert reviewing Pap tests at detecting cervical precancer.

Study senior author Mark Schiffman said: “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.

“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 team intends to use various imaging methods to further train the algorithm on representative images of cervical precancers and normal cervical tissue obtained from women across the world.

Global Good executive vice-president Maurizio Vecchione said: “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 setting.”