URUMQI, March 11 (Xinhua) — A research team in northwestern China’s Xinjiang Uyghur Autonomous Region has developed a new diagnostic technique that combines spectral analysis and artificial intelligence to quickly and accurately distinguish between two deadly and easily confused cardiac emergencies: aortic dissection and myocardial infarction.
Their method required just five to 10 minutes of blood sample analysis and achieved a diagnostic accuracy of 94.06 percent in differentiating acute myocardial infarction from aortic dissection, according to the study published in the journal Engineering Applications of Artificial Intelligence.
The study was conducted by a team from Xinjiang Uyghur Autonomous Region People’s Hospital led by Professor Yang Yining in collaboration with a team from Xinjiang University led by Professor Liu Xiaoyi.
Both myocardial infarction and aortic dissection cause sudden, severe chest pain, but their treatments are fundamentally contradictory. Myocardial infarction is caused by occlusion of a coronary artery and requires immediate clot-busting therapy to restore blood flow. In contrast, aortic dissection involves laceration of the aorta and such drugs are strictly contraindicated as they can cause fatal hemorrhage. Therefore, misdiagnosis can be fatal.
Traditional diagnosis relied heavily on imaging techniques such as contrast-enhanced CT scans. These methods require expensive equipment and a lot of time, making them difficult to implement in ambulances and primary care facilities, said Yan Lei, a member of the research team. With mortality rates increasing for both diseases the longer it takes to receive effective treatment, it is clear that fast and portable diagnostic tools are an urgent priority.
The team’s breakthrough lies in capturing the distinct molecular fingerprints these diseases leave in the blood. The researchers employed two complementary techniques, Raman spectroscopy and infrared spectroscopy, to detect biochemical information from patient serum samples.
To further improve diagnostic efficiency, the team developed a deep learning model that integrates data from both spectroscopy methods and enables rapid classification of the two diseases.
A diagnostic prototype based on this technology is currently undergoing multicenter clinical validation. The researchers say portable devices based on this technology could one day become standard equipment in ambulances and community clinics, allowing early intervention and buying valuable time for patients facing life-threatening situations. ■
