In what experts are saying is a potentially enormous development in global cancer technology, researchers in China have announced an artificial intelligence (AI) technology that can detect early-stage stomach cancer through standard CT scans. No other company has made such a claim, in China or the United States.
If this discovery holds up under more global scrutiny, the finding could transform cancer detection worldwide, save countless lives, and put China near the top of the cancer food chain.
Approximately 445,013 people die from stomach cancer each year in China, according to the National Institutes of Health (NIH). This figure represents a mortality rate of 32.06 per 100,000 people.
Early stomach cancer has been virtually invisible. But just days ago, Alibaba’s Damo Academy and the Zhejiang Cancer Hospital in Hangzhou, both in China, announced that they have created the first AI model capable of detecting gastric (stomach) cancer, including early-stage cases, by analyzing computed tomography (CT) scans.
The model, called Grape — short for “gastric cancer risk assessment procedure”— utilizes a deep-learning framework to process three-dimensional CT images, identifying and segmenting potential gastric tumors.
This breakthrough comes as Chinese scientists continue to push to take the global lead in cancer-based AI-powered healthcare. But there is an enormous amount of company. There are many companies, including a bunch in the United States, that are vying for the same kind of wins. Overall, the United States is still by far the lead in that race. But China is getting closer by design.
No matter how you slice it, what China has done here is very impressive.
“Identifying gastric cancer through AI algorithms is highly challenging due to the significant variations in the stomach’s distension and shape,” said Xia Yingda, an algorithm expert from DAMO Academy Medical AI Team, in a press release.
In one real-world case cited by Damo Academy, a patient with late-stage gastric cancer could have received a diagnosis six months earlier if Grape had been used, highlighting the system’s potential to reduce diagnostic delays.
The team hopes that by integrating this AI tool into standard clinical workflows, early detection rates can be significantly improved, ultimately leading to better patient outcomes.
Traditionally, a stomach cancer diagnosis relied on endoscopy, an invasive procedure involving a camera and biopsy tools inserted through the throat.
However, in China, fewer than 30% of patients comply with doctor’s recommendations for endoscopy due to its invasive nature, according to Cheng Xiangdong, a gastric surgeon at Zhejiang Cancer Hospital.
In a press release, the Chinese hospitals explain that the scientists are using GRAPE, an AI-based system, for early gastric cancer (GC) screening from non-contrast CT scans.
GRAPE has demonstrated the ability to predict GC from these scans months before clinical diagnosis, according to a study published in Nature Medicine.
The study shows AI can identify stomach tumors from non-contrast CT imaging. Up to now this method was not very effective.
According to the National Cancer Center of China, the country records 360,000 new cases and 260,000 deaths annually, with late-stage diagnoses contributing to a dismal five-year survival rate below 30 percent.
Patients diagnosed at stage one have survival odds exceeding 90 percent. However, traditional screening methods depend on invasive endoscopies, which are expensive, uncomfortable and often deter participation, the scientists say.
Trained on nearly 100,000 scans from 20 medical centers across China, the AI achieved 85.1 percent sensitivity in spotting early tumors, outperforming human radiologists by a substantial margin.
“We registered contrast-enhanced CT scans to non-contrast CT, which enabled the AI to gain intense precision information, accurately identifying gastric cancer locations,” explained Xia, the paper’s co-first author, in the press release.
The GRAPE addresses these challenges by turning routine CT scans, already widely used in hospitals and clinics, into powerful screening tools.
In real-world trials at two hospitals, the system boosted detection rates to as high as 24.5 percent. Notably, approximately 40 percent of the detected cases were stomach cancer patients without symptoms, according to the study.
“The AI model makes image-based gastric cancer screening possible for the first time,” said Cheng Xiangdong in an interview with China National News. Xiangdong is a lead researcher from Zhejiang Cancer Hospital.
One should have plenty of skepticism about this finding based on past trials, which have not always been as stringent as the US model. But multiple sources in China say that China is catching up and they are getting whatever they need to make it happen.
China has not been doing this nearly as long as the US. But they are making great progress. There are some who are skeptical of China’s race for the top of the mountain in cancer. But several sources I have spoken with say that they are coming up to speed.
I personally hope this “race” could be a positive one, not unlike the Race in Space versus the old Soviet Union, which in some ways brought us together.
Said one prominent oncologist and researcher that I know, “This is giving patients real hope, but historically there has been a different standard with some of the China trials compared to the trials in the US. But that is changing quickly. There is a big distance between China’s health sector and what we have in the United States.”
She continued, “They still have many miles to go. These kinds of successful trials have not been around as long as their US counterparts, or as thorough as what we see in the United States. But that is changing very, very quickly, and that is good for the entire world.”
GRAPE — the AI Model for Stomach Cancer Detection
Function: GRAPE is an AI system that analyzes 3D CT scans to identify early signs of stomach cancer.
Development: It was developed by Alibaba in collaboration with Zhejiang Cancer Hospital.
Performance: GRAPE demonstrates high sensitivity (85.1%) and specificity (96.8%) in detecting stomach cancer, potentially outperforming human radiologists.
Impact: The AI aims to improve early detection rates and reduce the need for invasive procedures like endoscopy, particularly in regions with high rates of gastric cancer.
Deployment: GRAPE is being deployed in screening programs across Zhejiang and Anhui provinces in China.
Application: Researchers are using AI to predict the characteristics of grape seedlings at an early stage, enabling faster and more efficient grape breeding.
Benefits: This approach can significantly reduce the breeding cycle and improve the accuracy of identifying desirable traits in grapes.
Example: A study from the Chinese Academy of Agricultural Sciences found that AI-powered grape breeding can increase efficiency by up to four times compared to traditional methods.
Source: Zhejiang Cancer Hospital
