Development and External Validation of a Nomogram Prediction Model based on Quantitative Coronary Angiography for Predicting Ischemic Lesions: A Multi-centre Study

Shuai Yang^1,2Shuang Leng^2,3Zhouchi Wang²Jiang Ming Fam^2,3Adrian Fatt Hoe Low^4,5Ru-San Tan^2,3Ping Chai^4,5Lynette Teo^5,6Chee Yang Chin^2,3John Carson Allen³Mark Yan-Yee Chan^4,5Khung Keong Yeo^2,3Aaron Sung Lung Wong^2,3Soo Teik Lim^2,3Qinghua Wu⁷Liang Zhong^2,3,8*

• ¹Department of Cardiology, Henan Provincial Chest Hospital, Zhengzhou, Henan Province, China
• ²National Heart Centre Singapore, Singapore, Singapore
• ³Duke-NUS Medical School, Singapore, Singapore
• ⁴Department of Cardiology, National University Heart Centre, Singapore, Singapore
• ⁵Yong Loo Lin School of Medicine, National University of Singapore, Singapore, Singapore
• ⁶Department of Diagnostic Imaging, National University Hospital, Singapore, Singapore
• ⁷Department of Cardiovascular Medicine, the Second Affiliated Hospital of Nanchang University, Nanchang, Jiangxi Province, China
• ⁸Department of Biomedical Engineering, College of Design and Engineering, National University of Singapore, Singapore, Singapore

Objectives: Quantitative coronary angiography (QCA) has significantly contributed to the diagnosis of coronary artery disease. This study aimed to construct and validate a QCA-based prediction model, represented as a nomogram, for predicting ischemic lesions defined by invasive fractional flow reserve (FFR)≤0.80. Methods: In this multi-centre study, we enrolled 220 patients with 303 interrogated vessels who underwent FFR measurements during clinically indicated invasive coronary angiography. QCA predictors for ischemic lesions were extracted to construct a nomogram model using Least Absolute Shrinkage and Selection Operator (LASSO) regression analysis of the development set (n=113 patients). An external validation (n=107 patients) was performed to assess the nomogram model’s discrimination and consistency. Results: Lesion length, minimal lumen diameter, stenosis flow reserve, percent diameter stenosis by visual estimation, and weight were included as predictors in the nomogram. The nomogram yielded an area under the curve (AUC) of 0.922 and 0.912 at per-vessel and per-patient levels, respectively, in the development set. In the validation set, it achieved an AUC of 0.915 and 0.912 at per-vessel and per-patient levels, respectively. Per-vessel accuracy, sensitivity, and specificity derived from the nomogram were 86.5%, 88.2%, 86.2% in the development cohort and 84.2%, 85.5%, and 83.1% in the validation cohort. For per-patient analysis, the corresponding values were 85.8%, 85.7%, 86.0% in the development cohort and 82.2%, 83.3%, 81.1% in the validation cohort. Conclusion: The nomogram may be useful for predicting ischemic lesions using QCA measurements and the LASSO regression algorithm, with external validation indicating potential predictive value in cardiology care settings.