A combined model of the triglyceride-glucose index and carotid ultrasound radiomics features can non-invasively predict coronary artery lesion severity in CHD patients, achieving an AUC of 0.823 in the training set and 0.730 in the test set.
The TyG index was identified as an independent predictor of coronary artery lesion severity in CHD patients.
Sex and hypertension status were also identified as independent predictors of coronary artery lesion severity.
A radiomics score (Rad score) based on 17 essential carotid plaque ultrasound features achieved moderate discriminative performance for coronary lesion severity.
The combined model incorporating sex, TyG index, hypertension status, and Rad score outperformed the Rad score alone in predicting coronary artery lesion severity.
The study used invasive coronary angiography (ICA) as the gold standard reference for classifying coronary artery lesion severity.
This research suggests that a combination of a simple blood-based metabolic marker (the triglyceride-glucose, or TyG, index) and detailed image analysis of carotid artery (neck artery) ultrasound scans can help predict how severe blockages in the heart's coronary arteries are, without the need for invasive heart catheterization. The study analyzed 381 patients with coronary heart disease who had all undergone the standard invasive diagnostic procedure (coronary angiography). Using a mathematical technique called radiomics, researchers extracted 17 key features from ultrasound images of plaques in patients' carotid arteries and combined these with clinical factors—including the TyG index, sex, and whether the patient had high blood pressure—to build a prediction model. The combined model performed substantially better than the image-based score alone, correctly distinguishing between patients with more versus less severe coronary disease about 73–82% of the time (as measured by the area under the curve, or AUC). The TyG index stood out as the strongest individual predictor, meaning that patients with higher levels of this marker—which reflects both blood sugar and fat metabolism—were nearly four times more likely to have more severe coronary artery disease. This research suggests that non-invasive tools like routine ultrasound imaging combined with standard blood test results could serve as a practical screening method to identify patients at higher risk for severe coronary artery disease, potentially helping clinicians decide who may need further invasive testing or more aggressive treatment. Because the study was conducted at a single center with a relatively modest sample size, further validation in larger and more diverse populations would be needed before widespread clinical adoption.
Hu H, Zhu R, Shen D, Zhang A, Hu W, Li X, et al.. (2026). Predicting Coronary Artery Stenosis Severity in Coronary Heart Disease: A Combined Model of Triglyceride-Glucose Index and Carotid Ultrasound Radiomics.. British journal of hospital medicine (London, England : 2005). https://doi.org/10.31083/BJHM49951