Artificial intelligence better than doctors in finding cervical precursors


Artificial intelligence (AI) can be prepared to eliminate cancer of the cervixAccording to a study known on Thursday, it showed that computer algorithms can detect precancerous lesions much better than trained experts or conventional screening tests.

According to the World Health Organization (WHO), cervical cancer is the fourth most common cancer in women, with an estimated 570,000 new cases worldwide in 2018.

In recent years there have been major advances in detection and vaccination to prevent the spread of the human papillomavirus, which causes the majority of cases of cancer of this type, however, these findings have benefited mainly the women of the rich countries. .

About 266,000 women died of cervical cancer worldwide in 2012, 90% of them in low- and middle-income countries, according to WHO.

"Cervical cancer is now a disease of poor, low-resource," said study lead author Mark Schiffman., a physician in the Division of Epidemiology and Cancer Genetics at the National Cancer Institute near Washington who has sought a cure for the disease for 35 years.

"We are trying to find ways that are extremely cheap, extremely easy, but very accurate, to be able to attack cervical cancer with the vaccine and also a little later, through a simple technique that is based on cell phones or something style, "he told AFP.

Schiffman was part of a team that built an algorithm from a file with more than 60,000 cervical images collected in Costa Rica.

The images were taken using only a speculum, a small light and a camera, without the need for advanced images.

The study began in the 1990s and included more than 9,400 women who were followed up for up to 18 years.

The AI ​​technique, called an automated visual assessment, found precancerous cells with 91% accuracy, according to a report in the Journal of the National Cancer Institute.

In comparison, a review by a medical specialist found 69% of previous cancers, while conventional laboratory tests, such as the Papanicolaou, found 71%.

Among women aged 25 to 49 years, who face the highest risk of cervical cancer, the AI ​​algorithm was even more accurate because it found 97.7% of precancerous cells.

"He worked a lot better than humans seeing those same pictures, and he certainly did a lot better than I did," Schiffman said.

The goal is to implement the technology in the next three to five years, enroll more patients in clinical trials around the world and ultimately make them easily accessible anywhere.


Source link