New diagnostic technique in multi-slice computed tomography for in-stent restenosis: Pixel count method

Objectives A diagnostic technique to objectively determine coronary in-stent stenosis was developed with multi(16)-slice computed tomography (MSCT), and it was compared with coronary angiography (CAG) in clinical cases. Background MSCT is expected to replace coronary angiography as a new non-invasive examination. Evaluation of highly calcified or in-stent lesions with CT is generally thought to be difficult. Methods Twenty lesions among 16 consecutive patients that were implanted with coronary stents were examined with both MSCT and CAG at follow-up. The minor axis cross sections of the stents were reconstructed at intervals of 1.5 mm with multiplanar reformation (MPR). The pixel with a higher CT value than the lowest CT value in the standard cross section at the proximal site out of stent was counted to determine the presence/absence of a stenotic lesion. Results Among 20 lesions, one case was not able to be evaluated with MSCT. MSCT correctly detected 3 of 4 cases with in-stent stenosis (sensitivity 75%), and 14 of 16 cases with no in-stent stenosis (specificity 88%, negative predictive value 93%, positive predictive value 75%). If analysis was made per-artery, sensitivity and specificity were 100% (3 of 3) and 87% (13 of 15), respectively, for detection of in-stent stenosis. Conclusions This study was performed to examine a unique diagnostic technique: pixel count method, for coronary in-stent stenosis with MSCT. It showed that the coronary in-stent stenosis could be determined when stent struts were clearly imaged. Further examination is required with various stents, especially those with a diameter of 3.0 mm or smaller.