Alibaba Group Holding Ltd.’s research institute has unveiled an artificial intelligence model that can detect colorectal cancer from routine CT scans with nearly 87 percent accuracy, a development that could challenge existing screening methods and bolster the company’s position in the growing health-tech market.
The model, called DAMO COCA, achieved 86.6% sensitivity and 99.8% specificity in a clinical study involving over 27,000 non-contrast CT scans, according to research from Alibaba DAMO Academy and Guangdong Provincial People’s Hospital published in Annals of Oncology.
The AI system successfully identified five cases of colorectal cancer that were previously missed in real-world clinical data. The method is the first of its kind to enable opportunistic screening for the disease without requiring patients to undergo invasive bowel preparation.
This marks Alibaba’s third major AI cancer screening model, following earlier releases for pancreatic and gastric cancers. The "CT + AI" initiative signals a strategic push into the multi-billion dollar medical imaging market, potentially creating a new revenue stream and setting a higher bar for competitors in AI-driven diagnostics.
Colorectal cancer is the second most common cause of cancer death in the United States, yet screening rates remain low, particularly in under-resourced communities. Traditional methods range from invasive colonoscopies to stool-based tests like the fecal immunochemical test (FIT) and the newer FIT-DNA test. While mail-in kits have improved participation, a recent study published in JAMA Internal Medicine found that even with outreach, completion rates were only 28% for FIT-DNA and 23% for FIT, with colonoscopy follow-up rates for abnormal results at a low 36%.
Alibaba's DAMO COCA proposes a different path. By analyzing existing, non-contrast CT scans—often performed for other medical reasons—the AI can opportunistically screen for cancer without additional procedures or patient compliance hurdles. The two-stage deep learning framework is designed to interpret complex intestinal structures, an area where human radiologists can miss lesions.
The development places Alibaba in direct competition with established medical device and diagnostic companies. While stool-based tests are less invasive than colonoscopies, they have lower adherence and require significant logistical support for patient outreach and follow-up, costs often borne by community health centers or test manufacturers. An AI that piggybacks on existing scans could offer a more cost-effective and scalable solution.
For investors, DAMO COCA represents a key proof point for Alibaba's long-term R&D strategy beyond its core e-commerce and cloud businesses. While the company has not disclosed a monetization strategy, the technology could be licensed to hospitals and imaging centers, integrated into cloud service offerings for healthcare clients, or used to build a comprehensive diagnostic platform. The success of the multi-cancer screening pathway could enhance investor confidence in Alibaba's ability to enter and compete in the high-growth health tech sector, where rivals like Huawei and Tencent are also making inroads.
This article is for informational purposes only and does not constitute investment advice.
