Background <p>Predicting anastomotic occlusion remains a major challenge in the management of patients after coronary artery bypass grafting (CABG). This study aimed to evaluate the prognostic value of the transluminal attenuation gradient (TAG), a parameter derived from coronary computed tomography angiography (CCTA), for this complication.</p> Methods <p>In this retrospective cohort study, data from 160 patients with 327 anastomoses were analyzed. Postoperative CCTA was performed to assess graft characteristics and TAG. The primary endpoint was anastomosis occlusion, as confirmed by follow-up CCTA. Predictors of occlusion were identified through univariate and multivariate analyses.</p> Results <p>During follow-up, 57 (17.4%) anastomoses occluded. Compared with patent anastomoses, occluded grafts exhibited significantly more negative TAG values, shorter lengths, and smaller anastomotic diameters—including both proximal and distal dimensions. Multivariable analysis identified higher TAG (OR 0.892), longer graft length (OR 0.989), and larger distal anastomotic diameter (OR 0.157) as independent protective factors, whereas elevated total cholesterol (OR 9.404) and HbA1c (OR 1.672) emerged as independent risk factors. The predictive performance improved sequentially across models: a clinical model (AUC = 0.746), a model enhanced with anatomical variables (AUC = 0.864), and the final model incorporating TAG, which achieved the highest discriminatory power (AUC = 0.886). Nomograms were constructed to facilitate individualized risk prediction.</p> Conclusion <p>TAG is a potent, independent predictor of anastomosis patency after CABG. Its integration into models combining clinical and anatomical variables significantly enhances risk stratification, offering a potential strategy for the early identification of high-risk patients who may benefit from targeted interventions to prolong graft patency.</p>