AI helps detect pre-cancerous colorectal growths

A new diagnostic system uses machine learning to differentiate between benign and malignant polyps in less than one second.

Developed by a team of researchers from Showa University in Yokohama, Japan, the artificial intelligence (AI) diagnostic system determines whether or not a growth is malignant. Using its knowledge database of more than 30,000 images, the AI takes less than a second to diagnose a patient’s polyps. Colorectal cancer is the second deadliest cancer (lung cancer is the first), so early detection is a must for survival.

In the first study of its kind, the AI system correctly identified polyps 86 percent of the time. With colorectal growths fairly common, particularly in people over the age of 50, rapid identification of pre-cancerous growths could reduce some of the burden on health systems strained by lack of funding and staffing combined with ageing populations. The scientists behind the discovery are planning a multi-site study as well as applying for regulatory approval for clinical use of the system.

Using technology to promote and support health and wellbeing has become fairly common, as many people track their diet and fitness with wearables and apps. Using AI and other innovations as a means of predictive healthcare, like a vest that alerts wearers to a forthcoming epileptic seizure; and retrofitted defibrillators that predict heart failures before they occur; are exciting developments. What other aspects of healthcare could use predictive tech to help minimize injury?