Artificial intelligence gives science and medicine a new impetus. It is a striking field of activity at a global level. They are already testing algorithms that reveal psychosis, various types of cancer and eye diseases, predict the risk of heart disease, and even help operators in emergency calls.
Now, a team of scientists headed by researchers from the National Institutes of Health of the United States (NIH) and Global Good have developed a computer algorithm that can analyze female cervical digital images and precisely identify pre-cancer changes that require medical attention. This Artificial Intelligence (AI) approach, called automated visual assessment, has the potential to revolutionize cervical cancer screening, especially in small resource settings.
To develop this method, researchers used a complete set of data to "train" a deep and automatic learning algorithm to recognize common patterns in medical images that can provide guidance.
The breakthrough was created jointly by researchers from the National Cancer Institute (NCI) with a global benefit and findings were published in the National Cancer Institute magazine.
"Our results show that a deep learning algorithm can use the images collected during routine cervical cancer screening to identify pre-cancerous changes that, if left untreated, can turn into cancer," said Mark Schiffmann of NCI and leading investigator. . "In fact, computer-aided image analysis was better able to identify pre-cancer cases than a human expert who reviewed the paprika smear under the microscope (cytology)," he added.
A very valuable tool
Automated visual assessment is easy to do. Health Visitors can use a cell phone or similar camera device for cervical cancer screening during one visit. This approach can be done with minimal training that makes it ideal for countries with limited health care resources, where cervical cancer is one of the main causes of women's illness and death.
To create an algorithm, the research team used more than 60,000 cervical images from the NCI photo file collected in a cervical cancer screening study in Costa Rica. More than 9,400 women took part in this study and continued for 18 years. The researchers got almost complete information on which cervical changes became cancer precursors and which were not. Photographs were digitized and then used to train this algorithm to distinguish neck conditions requiring treatment from those who do not.
Overall, the algorithm performed better than all standard screening tests to predict cancer diagnosed cases study in Costa Rica.
"When this algorithm is combined with vaccination against human papillomavirus (HPV), new HPV detection technologies and treatment improvements, cervical cancer may be controlled even in low-resource regions," said Maurizio Vecchione, Global Good Vice President.
Cervical or cervical cancer (CCU) is caused by HPV. Worldwide, CCU is the fourth most common cancer in women. There are about 530,000 new cases per year. Although it does not cause symptoms at the outset, it may later be a pain and other signs such as bleeding, reports the World Health Organization (WHO).
Read more about cervical cancer