PURPOSE To evaluate the diagnostic accuracy of multislice computed tomography coronary angiography (MSCT-CA) in the detection of in-stent restenosis. MATERIALS AND METHODS Forty-two patients (33 male, 9 female, mean age 58+/-8 years) previously subjected to percutaneous implantation of coronary stent with suspected in-stent restenosis, underwent a 16-row MSCT (Sensation 16, Siemens) examination. The average time between stent implantation and MSCT-CA was 7.4+/-5.3 months. The following scan parameters were used: collimation 16x0.75 mm, rotation time 0.42 s, feed 3.0 mm/rot., kV 120, mAs 500. After administration of iodinated contrast material (Iomeprol 400 mgI/ml, 100 ml at 4 ml/s) and bolus chaser (40 ml of saline at 4 ml/s) the scan was completed in <20 s. All segments with a stent were assessed by two observers in consensus and were graded according to the following scheme: patent stent, in-stent intimal hyperplasia (IIH) (lumen reduction <50%), in-stent restenosis (ISR) (=/>50%), in-stent occlusion (ISO) (100%). Consensus reading was compared with coronary angiography. RESULTS Forty-seven stents were assessed (16 in the right coronary artery; 4 in the left main; 22 in the left anterior descending; 5 in the circumflex). In 7 (17%) stents there was ISR (3) or ISO (4), and in 4 (10%) stents there was IIH. The sensitivity and negative predictive values for the detection of ISO were 80% and 98%, respectively, while for the detection of ISR+ISO they were 50% and 89%, respectively. CONCLUSIONS Although the results are encouraging, the follow-up of stent patency with MSCT-CA does not show a diagnostic accuracy suitable for clinical implementation.