AI boosts cervical cancer diagnoses

Inside a facility in Wuhan, Hubei province, tens of thousands of cervical smears will be checked each day for abnormal cells that may develop into cancer.

Such a workload is unimaginable in any major hospital as a pathologist is able to read no more than 100 samples a day.

But nowadays, advances in artificial intelligence have made it possible to accelerate cervical cancer screening and expand the service to women from urban and less developed, rural regions.

At the core of the facility's screening capability is an AI-assisted platform developed by Landing Med, a domestic company founded in 2000.

After a cervical smear is collected, "it can be scanned and uploaded to the online platform within three minutes. Then the online system can return a diagnosis within five minutes and is capable of reading up to 50,000 smears a day," said Sun Xiaorong, the company's chairman.

Hubei initiated a free cervical cancer screening program in 2022, and the AI-powered system had carried out nearly 5.27 million diagnoses by the end of last year. It is expected to cover all of the 12.6 million women aged 35 to 64 in the province by the end of this year.

Cervical cancer is caused by persistent infection with the human papillomavirus and is the second most common type of cancer for women in China.

In addition to expanding HPV vaccination among girls, rolling out screening to detect precancerous lesions is seen as the key to reining in the disease.

However, a lack of trained practitioners to read and flag precancerous signs, especially in the country's vast countryside, has been a major barrier.

Sun, a doctor and a researcher focusing on early cancer detection technologies, returned from Canada in the early 2000s to her hometown of Wuhan, where the local government's early efforts in cervical cancer screening made the city an ideal testing ground for AI technology.

Over the past two decades, Sun said that the company has accumulated over 10 million samples and relevant data to train and develop its AI system.

"The system's efficiency and accuracy are both high and it can help address a shortage of pathologists and resolve the problem of uneven skills of grassroots medical workers," she said.

Landing Med said the AI diagnosis system has been introduced to 31 provincial-level regions across China, including in remote frontiers in the Xizang autonomous region and towns in the mountains of Yunnan province.