Sechenov University scientists tested AI to detect cervical cancer

Russian scientists have tested artificial intelligence (AI) for the early detection of cervical cancer: the neural network showed an accuracy of 93% when analyzing colposcopic images. The study was conducted by specialists from the Department of Obstetrics and Gynecology No. 1 at Sechenov University in collaboration with the Sechenov Center for Motherhood and Childhood. This was reported to Izvestia on July 15.

Each patient underwent an extended colposcopy, cytological examination, testing for human papillomavirus (HPV), which is considered the main cause of cervical cancer, as well as other necessary examinations. The resulting images were uploaded to an artificial intelligence program created by the Russian company Vix Point AI. At the same time, the algorithm did not know any preliminary diagnoses.

The conclusions of the neural network were compared with the opinion of a medical council of three obstetricians and gynecologists. The coincidence of the results was 93%. In other cases, the algorithm revealed less pronounced changes than the experts. After a biopsy and histological examination, all diagnoses were confirmed.

According to Tea Djibladze, Professor of the Department of Obstetrics and Gynecology No. 1 at Sechenov University, this is the first study in Russia in which artificial intelligence was used to analyze colposcopic images.

The injected spot: AI will detect skin pigmentation disorder at the earliest stage.

The Russian vitiligo diagnostic tool has already been tested in India.

"The study showed that using an AI program is a promising method that can become part of a comprehensive diagnosis and screening that allows identifying risk groups. This will allow women to be referred in time for in—depth examination and to diagnose changes at the earliest stages," the specialist emphasized.

Experts stressed that in order to improve the accuracy of the algorithm, it needs to be continued to be trained. In the future, it is planned to upload not only colposcopy images, but also cytology results, HPV testing and other clinical data to the system. Such an integrated approach will make it possible to detect precancerous changes even more effectively and reduce the number of cases of late diagnosis of cervical cancer.

Scientists noted that artificial intelligence is already actively used in obstetrics and gynecology around the world. Neural networks help diagnose ovarian and endometrial tumors, predict the risk of preeclampsia, premature birth, assess the likelihood of successful natural childbirth after cesarean section, and even predict the development of postpartum depression.

At the same time, experts added that artificial intelligence should not replace a doctor. His task is to help make decisions, especially in regions where access to highly specialized specialists is limited. The final diagnosis and the choice of treatment tactics are still left to medical specialists.

Russian President Vladimir Putin said at the plenary session of the Eurasian Economic Forum in Astana on May 28 that Moscow is now an example of the digitalization of medicine, since while a patient is traveling in an ambulance, doctors can issue his "dossier." By the time you arrive at the hospital, all the necessary information is already available. Putin stressed that such technologies are fundamentally changing the approach to providing medical care, although previously it seemed impossible.