A team of Chinese scientists has developed an artificial intelligence (AI) platform that doubles the early detection rate of esophageal cancer.
Cyst disease is often asymptomatic, but experienced doctors can detect tumors and precancerous lesions with endoscopy. The five-year survival rate exceeds 90%, and when tumors are treated early, clinical outcomes are significantly reduced when patients begin to develop symptoms.
The study, published Thursday (BJT) in the journal Science Translational Medicine, introduces an architecture designed with a deep learning algorithm trained on a database of more than 190,000 esophageal images collected from various clinics in China.
The clinical trial led by researchers from Taizhou Hospital in Zhejiang Province and Renmin Hospital of Wuhan University recruited more than 3,000 participants who underwent endoscopy. The AI system works on half of the volunteers.
According to the study, this platform doubled the incidence of high-risk esophageal lesions (1.8 percent) compared to the control group (0.9 percent).
Also, the real-time AI tool shows a high sensitivity and specificity of 89.7 percent and 98.5 percent, respectively, under practical conditions.
