Alibaba’s ’Grape’ AI Now Beats Radiologists at Spotting Early-Stage Stomach Cancer

Alibaba just dropped a diagnostic bomb—their new 'Grape' AI detects early stomach cancer faster and more accurately than human radiologists. No hype, just hard numbers.

How? The system crunches imaging data with machine learning precision that leaves traditional methods in the dust. Forget second opinions—this is algorithmic certainty.

And before you ask—no, this won’t be monetized via an NFT collection. Yet.

Game-changer or overreach? Either way, healthcare’s AI arms race just got a Chinese turbocharge.

TLDRs;

• Alibaba’s new AI tool, Grape, detects stomach cancer from CT scans with higher accuracy than radiologists.
• The model achieved 85.1% sensitivity and 96.8% specificity in clinical tests with Zhejiang Cancer Hospital.
• Grape will be used in cancer screening programs across Zhejiang and Anhui provinces.
• Despite strong technical results, success depends on patient trust and integration into clinical workflows.

Alibaba Group has introduced a powerful artificial intelligence tool called Grape, aimed at detecting stomach cancer in its early stages.

Developed in collaboration with Zhejiang Cancer Hospital, Grape leverages DEEP learning to analyze 3D CT scans with a level of precision that surpasses most radiologists. The model achieved 85.1% sensitivity and 96.8% specificity in recent trials, making it one of the most promising medical AI applications in China to date.

Grape is designed specifically for gastric cancer, which remains one of the deadliest forms of cancer in East Asia due to its late-stage diagnosis. The tool is the latest innovation from Alibaba’s Damo Academy, a research division responsible for several AI-powered medical tools including a pancreatic cancer detection model that recently received fast-track status from the US FDA. Building on this momentum, Grape is expected to play a key role in early screening campaigns throughout the provinces of Zhejiang and Anhui.

AI Accuracy Surpasses Human Diagnosticians 

In real-world tests, Grape demonstrated superior diagnostic capabilities when compared to human radiologists, particularly in identifying early-stage tumors that are often difficult to detect. Its high specificity means fewer false positives, while its strong sensitivity ensures a higher detection rate for actual cancer cases.

These figures suggest that Grape could not only reduce misdiagnoses but also ease the burden on overworked healthcare providers in China’s public hospital system.

Despite the technical breakthrough, experts note that clinical adoption will require more than just high performance. Dr. Cheng Xiangdong from Zhejiang Cancer Hospital emphasized that fewer than one-third of patients currently follow through with endoscopic screenings, largely due to discomfort, cost, and fear. Grape’s success, therefore, will depend heavily on whether patients feel comfortable trusting an AI-based diagnosis over traditional procedures.

Healthcare Rollout Will Require Cultural Shift 

Beyond patient hesitancy, another challenge lies in persuading healthcare professionals to embrace AI assistance. Doctors often express concerns about losing autonomy or encountering workflow disruptions when new technologies are introduced. Alibaba will need to focus on building trust within the medical community through transparent training, ease-of-use design, and seamless integration into existing diagnostic systems.

Studies on AI adoption in healthcare show that involving end-users during development can significantly improve trust and usability. It remains unclear whether frontline radiologists were consulted during Grape’s design phase, although the model underwent rigorous validation with Zhejiang Cancer Hospital. As Alibaba prepares for regional rollout, the company is expected to provide structured onboarding support for medical staff and institutions.

Future Plans Target Broader Cancer Detection

The Grape model is not a standalone effort but part of a broader push by Alibaba to use artificial intelligence in early cancer detection. The team behind Grape has already hinted at expanding its capabilities to screen for other gastrointestinal cancers in future iterations. This aligns with global trends in AI-driven preventive healthcare, where early diagnosis can dramatically improve survival rates and reduce treatment costs.

If Grape succeeds in real-world deployments, it could mark a turning point in China’s public health approach to cancer. However, the road ahead will be shaped as much by social and institutional readiness as by technological excellence.