(ECNS) --A hospital in eastern China announced the launch of the world's first AI model for gastric cancer screening based on medical imaging, DAMO GRAPE, according to its press conference held in Hangzhou City, Zhejiang Province, on Wednesday.
This AI model is the first to use plain CT scans to detect early-stage gastric cancer lesions. In a large-scale clinical study involving nearly 100,000 participants across 20 medical centers in China, the model significantly improved the detection rate of gastric cancer. The research results were published in the prestigious journal Nature Medicine.
The AI model achieved impressive accuracy, demonstrating a sensitivity of 85.1% and a specificity of 96.8%. In other words, if 100 people had gastric cancer, the AI would correctly detect about 85 of them, while potentially missing around 15 cases (false negatives). Then, out of every 100 healthy individuals tested, approximately 97 would be correctly labeled as healthy, while about 3 would be falsely identified as having cancer (false positives).
Hu Can, a PhD candidate specializing in gastrointestinal surgery at Zhejiang Cancer Hospital and the paper's lead author, said the AI model could potentially fill a major gap in the early detection of gastric cancer through imaging. During their research, analysis of CT scans from 11 patients prior to their gastric cancer diagnoses and found that this AI model could detect the cancer 2 to 10 months earlier than traditional methods used in hospitals.
This model, for the first time, makes image-based screening for gastric cancer possible, said Cheng Xiangdong, the hospital's Party secretary and vice president of the Chinese Anti-Cancer Association, who is also the paper's corresponding author, adding that they already deployed the AI model in several high-incidence areas in China and plan to expand its use abroad.
(By Gong Weiwei)
京公网安备 11010202009201号