How AI Improves Early GI Cancer Screening: A Clinical Perspective

Gastric and colorectal cancer remain among the most common cancers in China, and early detection is the single most effective way to improve outcomes. Endoscopy is the gold standard for diagnosis, but lesion detection depends heavily on the operator's experience — and during a busy endoscopy list, subtle early-stage lesions can be missed.

That is where AI-assisted endoscopy software such as yoyolink comes in.

The challenge in early GI cancer screening

In tertiary hospitals, high case volumes put pressure on physicians' time: capturing images, writing reports and compiling quality-control data all compete with the examination itself. In primary-level hospitals, endoscopists often have less experience, and the detection rate of early-stage cancer and flat or subtle lesions tends to be lower.

Large-scale screening programs — including government livelihood initiatives and routine screening in high-incidence regions — amplify both problems: more examinations, more documentation, and an urgent need for consistent quality.

How AI helps

yoyolink acts as a real-time "second pair of eyes" during gastroscopy and colonoscopy:

• Real-time lesion prompts: self-developed deep learning models highlight suspected lesions such as early-stage cancer, polyps and ulcers as the scope moves;
• Anatomical site tracking: automatic annotation of the duodenal bulb, gastric fundus, colon and other sites helps ensure complete coverage, reducing missed regions;
• Automatic documentation: screenshots, procedure timing, quality scoring and structured image-text reports remove repetitive paperwork;
• QC statistics: completeness, Boston Bowel Preparation scores and detection indicators are aggregated into charts automatically, supporting continuous improvement.

Evidence from clinical validation

yoyolink is co-developed with several renowned Class-A tertiary hospitals in China and holds an NMPA medical device registration certificate. More than 1000 real clinical cases have been collected across centers in multicenter clinical validation, and the system has been deployed at multiple Shanghai hospitals for endoscopy centers and early cancer screening.

A model that scales from tertiary to primary care

For tertiary hospitals, yoyolink improves efficiency. For primary-level hospitals and screening programs, it raises the detection rate of early-stage cancer and other lesions — exactly where the need is greatest. Because image-text information is fully traceable and QC data are standardized, the same AI can support medical consortia (医共体/医联体), paired-assistance programs and government screening initiatives.

Learn more

• Clinical applications — scenarios, audiences and deployment roadmap
• Product overview — how the software works
• Contact us — request clinical validation materials

From detection to follow-up

Early cancer screening matters only if findings are acted on quickly. yoyolink's patent CN112233752A/B adds criticality assessment, report print priority and neural-network physician matching after lesion detection, so high-risk findings move faster into follow-up and specialist pathways. Data security is reinforced by patent CN112714233B (block-based encrypted image transmission), supporting hospital-local storage and compliant deployment in screening networks.

Patents & IP

Basemed Medical protects its technology with granted invention patents:

• "Information Processing Method and Device for Endoscopic Examination Report" (CN112233752A/B, co-filed with Zhejiang University): intelligently processes endoscopy reports — detects lesions, assesses criticality, prioritizes report printing, and matches patients with suitable physicians via neural networks. View | PDF
• "Method and System for Intelligent Transmission of Endoscopic Images Based on Block Decoding" (CN112714233B): transmits endoscopic images in blocks with multi-layer key information and quality detection, improving transmission quality, security and efficiency. View | PDF

See the full portfolio: About · Patents & IP