A group of Japanese researchers have recently used an artificial intelligence (AI) system to detect early stage stomach cancer at the same efficacy of veteran physicians. Researchers tested the system with 100 images of normal stomach tissue, and 100 images of cancerous stomach tissue, and found that it was capable of detecting cancer from an endoscopic image in a matter of 0.004 seconds. Riken and the National Cancer Center report that the system correctly diagnosed cancer with an accuracy rate of 80 and recognized normal stomach tissue 95% of the time. The institute claims that these rates are on par with those of highly trained doctors, and plan to implement this AI into clinical settings in the form of a supplemental diagnostic device.
Early Detection Challenge, AI Benefits?
Early detection of stomach cancer is extremely challenging, being that the infected tissue often resembles symptoms of general inflammation, however it is imperative that the cancer is detected as early as possible to yield long-term survival.
It took the A.I. algorithm only 0.004 seconds to judge whether an endoscopic image showed early-stage cancer; correctly detected cancer in 80 percent of cancer images, while the accuracy rate was 95 percent for normal tissue. https://t.co/K7MtIPTYPz
— Berci Meskó, MD, PhD (@Berci) July 22, 2018
AI Impact on Stomach Cancer
This system has the potential to make a profound impact on healthcare in Japan, where incidence of stomach cancer is exceptionally high. The National Cancer Center projected there to be 132,800 incidences of stomach cancer in Japan in 2017, with 47,400 cases being fatal, making stomach cancer the second most prevalent form of the disease in Japan, and the third most fatal. A study conducted in 2006 found that the gastritis prevalence was 47% in UK citizens, and 60% in Japanese citizens. From this data the researchers concluded that the unusually high rate of gastritis in Japan accounts for the prevalence of stomach cancer among their population.
Other companies are beginning to use AI in diagnosis of various diseases, with companies like Lunit developing AI capable of lung and breast cancer detection, and Ultromics using AI to diagnose lung cancer and heart disease. Acquired by Google in 2014, DeepMind has been working with National Health Services to use AI in diagnosing eye diseases as well.
Check out a recent article on an exciting new company in the healthcare space.