China unveils AI system for early stomach cancer detection

Chinese researchers have developed an advanced artificial intelligence (AI) model capable of detecting early-stage gastric cancer using non-contrast computed tomography (CT) scans.

This is reported by Xinhua News Agency, a partner of TV BRICS.

The AI system is named GRAPE (Gastric Cancer Risk Assessment Procedure with Artificial Intelligence).

It enables detection of stomach tumours in the early stages, relying solely on non-invasive, widely available CT imaging, previously considered ineffective for gastric cancer diagnosis.

The model was trained on nearly 100,000 CT scans sourced from 20 medical institutions across China.

It demonstrated an 85.1 per cent sensitivity in identifying early gastric tumours, surpassing the performance of experienced human radiologists.

The AI uses a novel image registration technique, aligning contrast-enhanced CT data with non-contrast scans, allowing the model to identify tumours with high precision.

Clinical trials conducted in two hospitals showed the model raised detection rates to 24.5 per cent, including cases in asymptomatic individuals.

In a retrospective study, GRAPE identified signs of cancer in CT scans taken two to ten months before formal diagnosis.

The model has already been deployed at hospitals in Zhejiang and Anhui provinces, with further implementation planned both within China and abroad.

China’s broader strategic push into AI-driven medical diagnostics has yielded several national-level breakthroughs.

Researchers also recently developed an AI tool for cervical cancer screening, achieving accuracy on par with, or superior to, human experts, the source claimed.