Background In-stent restenosis is a common complication after stent implantation. However, the assessment of stent lumen in computed tomography （CT） coronary angiography is limited by multiple factors. Our study aimed to evaluate the accuracy and the suspected affecting factors in diagnosing coronary in-stent restenosis by dual-source CT （DSCT） compared with coronary angiography. Methods One hundred and fifteen stents in 50 patients were evaluated with DSCT before coronary angiography for the detection of coronary in-stent restenosis （〉 50% luminal narrowing）. Patency of each stent was analyzed by two independent expert radiologists blinded to the results of coronary angiography. The relationship between diagnostic accuracy and the suspected factors including age, body mass index （BMI）, heart rate, variation of heart rate, radiation dose, image quality, location and stent characteristics （type, material, diameter, length and strut thickness） was assessed with both univariate and multivariate analysis. The fitting of a Logistic regression model was evaluated using a receiver operating characteristic （ROC） curve. Results Mean stent diameter was （2.9±0.4） mm. Sensitivity, specificity, positive and negative predictive values and accuracy of DSCT in detection of in-stent restenosis were 69.2%, 91.2%, 50.0%, 95.9%, and 88.7%, respectively. In a subgroup of stents with a diameter 〉3.0 mm, sensitivity, specificity, positive and negative predictive values and accuracy were 100.0%, 96.5%, 75.0%, 100.0%, and 96.8%, respectively. Stent diameter 〈3.0 mm and poor image quality were associated with poor diagnostic accuracy （P 〈0.05）. The area under curve of ROC was 0.79. Conclusion DSCT can provide high accuracy for the assessment of in-stent restenosis in stents with a diameter 〉3.0 mm, and can play an important role in ruling out in-stent restenosis.
