The National Institute of Health and Global Good has developed an algorithm that can analyze images of a woman’s cervix. It can also identify precancerous changes that require medical attention. This artificial intelligence (AI) approach, known as automated visual evaluation, has the power to change cervical cancer screening.
The researchers have used comprehensive datasets to train a machine so that it can recognize the patterns. Investigators created the approach at the National Cancer Institute (NCI) and Global Good.
The team from the National Cancer Institute said, “Our findings show a deep learning algorithm and can use images during cervical cancer screening so that precancerous changes can be identified and can be treated. We realized that computer analysis of the images was better at identifying precancer than a human expert reviewer.
On the same lines, Dr. Vikas Goswami, a renowned oncologist, says, “The new method is beneficial and less expensive than other methods. The healthcare workers can use a phone or similar camera device for cervical screening and treatment during a visit. Besides, the approach can also be performed with minimum training and can be ideal for countries who have limited health care resources.”
The research team used more than 60,000 cervical images during a cervical cancer screening study that was carried out in Costa Rica to create the algorithm. There were more than 9400 women who participated in the study with a follow up that lasted up to 18 years.
Due to the different studies, the researchers gained complete information on which cervical changes became precancers and which did not. The photos were used to train a deep algorithm so that it can distinguish cervical conditions that require treatment and do not require treatment.
When the algorithm is combined with HPV vaccination, emerging technologies in HPV and improvements in treatment can be controlled. The researchers plan to train the algorithm on representative images of cervical precancers and cervical tissue from women, using a camera and imaging options.
This is important since there are a lot of variations in the appearance of the cervix in several geographic regions. The goal of the project is to create the best algorithm for everyday use.