Detection of in-stent restenosis of coronary stents using 40-detector row computed tomography in vitro

OBJECTIVE: To evaluate the performance of 40-detector row computed tomography (CT) in the detection of in-stent stenosis of coronary stents. METHODS: Seven patent vascular models, 7 stenotic models, and 7 obstructed models were scanned with a 40-detector CT. We made the vascular models using 3 types of stent (Bx-Velocity, Express2, Driver) with an inner diameter of about 2.5, 3.0, or 3.5 mm. We measured the stent lumen diameter and evaluated the in-stent stenosis visually for the 21 vascular models. We evaluated attenuation values of the stent lumen of the 9 patent models of 2.5-mm diameter. RESULTS: The inner diameters of the vascular models were underestimated by CT with mean measurement errors of -1.19 to -1.49 mm. The absolute mean overall measurement error decreased as the inner diameter increased. The direct visualization of in-stent stenosis was possible for the 3.0- and 3.5-mm diameter models, but impossible for the 2.5-mm diameter models. For patent vascular models of 2.5-mm diameter, the CT attenuation inside the stent was significantly higher than that of the unstented portion (P < 0.0001). For obstructed vascular models of 2.5-mm diameter, the CT attenuation inside the stent was significantly lower than that of the unstented portion (P < 0.0001). Also for stenotic vascular models of 2.5-mm diameter, the CT attenuation inside the stent was lower than that of the unstented portion. CONCLUSIONS: Visualization of stent lumen on CT is affected by the stent diameter. Measurement of stent lumen is useful for detection of in-stent stenosis, when the direct visualization of in-stent stenosis is impossible.