AI system can detect bowel cancer in less than 1 second

Researchers at Showa University in Yokohama, Japan, said the artificial intelligence (AI) system only needs to look at endoscopic images of tumors.

AI will diagnose based on endoscopic images. Photo: GUILLAUME SOUVANT/AFP/GETTY IMAGES

According to ZDNet, the system is designed to detect both duodenal, colon, and rectal cancers and was trained with more than 30,000 extensive endoscopy images.

After analyzing 250 patients and 300 intestinal tumors, AI was able to predict malignant tumors in less than 1 second with 94% accuracy.

“The most notable breakthrough in this system is that AI allows immediate radiological biopsies of rectal polyps during colonoscopy,” said Dr. Yuichi Mori, who led the research team.

Patients don't have to wait days for biopsy results from the lab to know if they have cancer. This could prevent unnecessary surgery for some patients."

These discoveries were presented at the United European Gastroenterology Week conference in Barcelona, ​​Spain. However, the method for detecting and diagnosing these tumors has not yet been approved.

“We believe these results are acceptable for clinical diagnosis and our immediate goal is to achieve approval,” he added.

It is not yet clear whether the system can be similarly applied to other types of cancer using imaging and other scanning methods.

Using AI to detect cancer isn’t an entirely new idea, however. Earlier this year, a team from the University of Warwick created an image repository of tumors and immune cells based on thousands of human cell types. The team has since begun work with the UK’s National Health Service on detecting cancer in poultry and lung cancer.

A group of scientists in Australia from the Commonwealth Scientific and Industrial Research Organization recently announced an AI research project to monitor the process of creating new blood vessels to detect the development of cancer cells.

According to TTO