A Shanghai hospital has unveiled China’s first artificial intelligence system designed to mimic the diagnostic reasoning of the world’s leading cardiologists in a bid to tackle the country’s overwhelming demand for cardiac care.
CardioMind, developed jointly by Fudan University-affiliated Zhongshan Hospital and the Shanghai Academy of Artificial Intelligence for Science, compares patient histories and test results with global research to generate diagnostic suggestions.
According to its creators, the system is not intended to be a replacement for physicians, but rather a “co-pilot” that can help overburdened doctors work faster and more accurately.
“We’re feeding it cardiovascular data and teaching it to think like a top expert cardiologist,” said Ge Junbo, a leading cardiologist and academician with the Chinese Academy of Sciences, who spearheaded the project.
“With the help [of AI], our doctors can serve more patients, reduce the overall workload, and improve the quality of diagnosis and treatment,” Ge told Yicai Global News.
While AI tools like CardioMind are proliferating globally, proponents argue that they could be especially transformative in China, where a shrinking workforce and ageing population are straining medical resources.
The urgency is clear. For example, the Zhongshan cardiology department’s 136 physicians last year handled 820,000 outpatient visits – a ratio underscoring the pressures facing China’s top public hospitals.
Unlike open-source AI models such as China’s DeepSeek, CardioMind has been trained specifically for cardiovascular diseases, drawing on decades of internal data including hundreds of thousands of anonymised patient records from the hospital’s archives.
The system – which has also been trained on the latest international treatment guidelines and research papers spanning coronary artery disease, heart failure, and other subfields – synthesises electrocardiograms, ultrasound images and blood tests to draft structured medical reports and recommend diagnoses.
CardioMind is in use only at Zhongshan Hospital, where doctors review all of its outputs and make the final decisions on patient treatments and care, according to the Yicai Global News report.
While CardioMind’s developers emphasise the system’s rigorous testing, there are still lingering ethical concerns. Patient privacy, algorithmic bias, and questions of liability for AI-introduced errors remain unresolved hurdles in China and elsewhere.
According to a Citic Securities report, China has launched more than 50 medical large language models since 2023. Most aim to streamline paperwork or assist diagnostics, though critics caution that real-world clinical validation remains limited.
Google’s Med-Gemini – which processes text, images and biosignals – is leading global performance benchmarks, outpacing OpenAI’s GPT-4 in medical accuracy by 44 per cent, according to recent studies.
However, Chinese institutions are closing the gap, with development driven by vast patient data sets and government backing for AI-based innovations. South China Morning Post