Technology and data acquisition

High spatial and temporal resolution is needed to evaluate coronary arteries, particularly in case of stented vessels. Therefore, ICA is the gold standard technique for ruling out ISR. In the late 1990s, MDCT with retrospective ECG gating was introduced. This technology, characterized by thinner and multiple sections and lower pitch factors (table feed/gantry rotation) consists of a continuous spiral scan of the heart with simultaneous ECG recording. This translates into higher spatial and temporal resolution that allows the evaluation of stented segments.

The main technical factors that influence diagnostic performance are the number of slices, slice thickness, gantry rotation time, scanning parameters, heart rate during the scan, the type of contrast injection protocol, and postprocessing analysis ( Tables 1 and 2 ). A progressive increase in the number of slices from 4 up to 40 allowed a contemporaneous increase in spatial resolution from 1.25 to 0.5 mm, providing promising results in ISR detection.

The temporal resolution of old-generation scanners was further improved in the mid-2000s with the introduction of 64-slice multidetector computed tomographic scanners and DSCT. Moreover, in addition to detector improvements, these scanners were equipped with much faster gantries, resulting in effective temporal resolution up to 83 ms.

However, matrices with such thin layers required very high tube current and voltage, translating to radiation exposure increases. The studies using 64-slices scanners included in this review reported effective x-ray doses ranging from 10 and 20.5 mSv. Several strategies have been proposed to reduce radiation dose, such as modulation dose protocols, dual-source scanning, and prospective ECG gating. With the latter technique, the table remains stationary while the x-ray tube rotates around the patient and is advanced for the subsequent scan only when data acquisition is completed. This allows a reduction of up to 5.7 mSv without impairing diagnostic accuracy for ISR detection.

Several studies demonstrated that heart rate has an important impact on the diagnostic performance of MDCT. Although the heart rate limit depends on the temporal resolution of the scanner used, a target heart rate ≤65 beats/min is always desirable. This was achieved with β blockers administered orally 60 to 90 minutes before the scan, intravenously immediately before, or both. Image quality has also been shown to improve with nitroglycerin, and therefore, some investigators administered sublingual nitrates immediately before scans to achieve maximum coronary vasodilatation.

With regard to the contrast injection protocol, no difference was found between the “bolus-tracking technique,” faster and more reliable, and the “test bolus technique.” With regard to postprocessing analysis, image data sets were analyzed in all studies using reconstruction protocols and vessel analysis ( Figure 1 ). To improve stent visibility and to decrease artifacts, dedicated kernels can be used. Reconstruction kernels are mathematical calculations applied to raw data that control low contrast detectability and spatial resolution in the reconstructed images. Although a smooth kernel is suitable for the evaluation of the vessel lumen, vessel wall, and surrounding tissue, the visualization of high-density structures such as stents requires a sharp kernel ( Figure 2 ). Indeed, this was used for ISR evaluation in almost all studies reported in this review.

(A) MDCT (volume rendering) showing stents (arrows) implanted in the first diagonal branch (D1) and left circumflex coronary artery (LCX). (B) Multiplanar reconstruction of the left main coronary artery (LM) and LCX showing stent patency, confirmed by short-axis view of the proximal (C) and distal (D) stent segments. LAD = left anterior descending coronary artery.

(A) MDCT (volume rendering) showing multiple stents (arrows) implanted in the proximal segment and first marginal branch of the left circumflex coronary artery (LCX). (B) The blooming effect caused by the metallic struts hinders adequate stent luminal visualization with multiplanar reconstruction. (C) Use of a dedicated sharp kernel reduces the blooming effect, with a significant improvement in stent luminal visualization.

Finally, 3 different methods were used to determinate the degree of ISR: qualitative, semiquantitative, and quantitative. With the qualitative technique, significant ISR (diameter reduction ≥50%) was visually detected. With the semiquantitative method, a 4-point scale ranging from 1 (patent stent lumen) to 4 (stent occlusion) was used. In the quantitative measurement, the percentage of stenosis was calculated as the ratio between the diameters, measured in the short axis, of the narrower stent lumen and of the proximal and distal reference segments. However, no significant difference was reported in the accuracy of the qualitative and quantitative method.

Diagnostic performance of 4-, 16-, and 40-slice MDCT

Two of 6 studies included in the present review showed that multidetector computed tomographic scanners of this generation were unable to correctly identify all stent segments. Indeed, up to 18% of stents were not recognized in these studies, as well as in other studies not included in this review because of the exclusion criteria. The reasons were mainly motion artifacts and severe calcifications. Therefore, analysis was performed on the correctly identified stents only.

Table 3 lists the percentage of nonevaluable coronary stented segments and the diagnostic accuracy parameters reported in the 6 published studies that investigated the diagnostic performance of 4-, 16-, and 40-slice MDCT. The rate of coronary stents classified as not evaluable because of poor image quality ranged from 2.6% and 23.5%. The overall feasibility of MDCT was 90.4%. Image quality was impaired mainly by motion artifacts related to patients’ inability to hold their breath or refrain from chest movement. These were the reasons for uninterpretable stented segments in 100% of the cases in a study by Cademartiri et al and in 40% of the cases in a study by Gaspar et al and represented the first cause of nonevaluability in a study by Kefer et al. This is due to the lower temporal resolution of scanners of this generation, causing longer scan times requiring that patients hold their breath longer.

Another cause of impaired quality was the high-density artifacts generated by metallic stent struts. Three studies found that stent diameter may influence the evaluation of stent lumen on MDCT. In a study by Tedeschi et al, only 1 of 9 stents (11%) with diameters <3 mm was judged evaluable, compared with 60 of 81 stents (74%) with diameters ≥3 mm. In a study by Soon et al, all nonevaluable stents had diameters <3 mm. Similarly, in the study by Kefer et al, nonevaluable stents were of significantly smaller diameters. In the latter study, nonassessable stents had thinner struts (87 vs 110 μm). As listed in Table 3 , the overall diagnostic accuracy of the 6 studies included in this review was 90%. As shown in the study by Kefer et al, stent diameter plays a significant role in the correct evaluation of stent patency.

In conclusion, most of the studies using 4-, 16-, and 40-slice MDCT reported high overall feasibility for the evaluation of stent patency. However, some studies not included in this review because of the exclusion criteria showed more relevant rates of nonevaluable stents: 46% (106 of 232) in a study by Gilard et al and 77% (50 of 65) in a study by Schuijf et al.

Feasibility and imaging quality of 64-slice MDCT and DSCT

Table 4 lists the percentages of nonevaluable coronary stented segments reported in the 18 published studies that investigated the diagnostic performance of 64-slice MDCT and DSCT. With the exception of a very high rate (42%) of nonevaluable stents reported in a study published in 2006 by Rixe et al, the rate of coronary stents classified in the other studies as not evaluable because of poor image quality ranged from 0% to 19.5%. The overall feasibility of MDCT was 90.4%.

The main reasons for impaired image quality and stent assessment were high-density artifacts generated by metallic stent struts, followed by motion artifacts related to patients’ inability to hold their breath or refrain from chest movement, misalignment of slices due to heart rate variation or premature beats, and blooming effects due to vessel calcifications.

Some investigators evaluated stent and patient characteristics that may have a negative impact on image quality with MDCT. A number of stent characteristics, such as diameter, type, strut thickness, material, cell shape, and the complexity of the interventional procedure (bifurcation or overlapping stenting) were evaluated. Almost all studies showed that stent diameter plays a major role in the evaluation of the stent lumen. In a study performed at our institution, we found a significant increase in evaluability in stents with diameters ≥3 mm compared to those with diameters <3 mm. Accordingly, the mean diameter of evaluable stents was significantly larger than that of nonevaluable stents. Rixe et al confirmed these findings, showing that the mean diameter of evaluable stents was significantly larger than the diameter of nonevaluable stents (3.28 vs 3.03 mm). Two studies performed by Oncel et al, the first with 64-slice MDCT and the second with DSCT, and a study by Carbone et al showed a significant difference in the percentage of evaluable stented segments between ≥3-mm-diameter stents and <3-mm-diameter stents. In a study by Pflederer et al, the mean diameter of stents classified as assessable (3.30 mm) was significantly higher than that of stents classified as nonassessable (3.0 mm); in the same study, multivariate analysis showed that diameter was significantly associated with stent assessability. Finally, Pugliese et al, using DSCT, demonstrated that all stents classified as not evaluable had diameters ≤2.75 mm, confirming the key role of stent size.

Two studies evaluated the effect of strut thickness on stent evaluability, demonstrating a significant reduction in artifacts with thinner stents (<100 μm). The influence of stent material on image quality was assessed by Oncel et al, who found that the cobalt-chromium alloy produced fewer metal artifacts than did other materials.

The finding that stent luminal visibility greatly varies depending on the stent type was also reported in an in vitro study by Maintz et al. Using a 64-slice MDCT, they showed that all types of stents made of cobalt-chromium alloy had luminal visibility >66%. In another in vitro study performed with DSCT and using all 4 different type of kernel reconstruction protocols, Maintz et al confirmed the influence of stent material on image quality, showing that the mean luminal diameter was highest with stents made using magnesium (WE43) and lowest with those made using tantalum. A study by Schuijf et al showed that image quality was significantly lower in overlapping stents than in stents without overlap. Finally, no significant difference was found between stents with open or closed cell designs in the 2 studies that evaluated this feature.

Two studies performed with 64-slice MDCT showed higher feasibility when the heart rate during the scan was <60 versus ≥60 beats/min. In our study, heart rate was also a predictor of better feasibility on multivariate analysis. Also, the study by Schuijf et al confirmed this finding, showing that the mean heart rate in the interpretable stents was significantly lower than that in the noninterpretable stents (55 vs 72 beats/min). In contrast, the 2 studies performed with DSCT showed no significant difference between <70 and ≥70 beats/min during the scan. This was likely due to the increased temporal resolution of this type of scanner in comparison to single-source scanners.

In conclusion, concerning stent characteristics that may influence stent image quality, the role of stent diameter is critical, as shown by almost all studies included in this review. The role of strut thickness is still controversial, with studies in which this parameter affected stent assessment and other studies that found no impact. However, it is important to note that in the latter studies, the cutoff for distinguishing thin from thick strut stents was higher (110 vs 120 or <140 vs >140 μm). The role of stent material has been demonstrated by in vitro studies and confirmed by the study of Oncel et al. Finally, high-density artifacts generated by metallic stent struts are more evident in overlapping stents. Concerning patient characteristics, heart rate during the scan has a fundamental role on stent image quality, as in the case of native coronary arteries. Figure 3 shows 2 cases of nonevaluable stents due to high-density artifacts generated by metallic stent struts and to the misalignment of slices related to heart rate variability.

(A,B) Multidetector computed tomographic multiplanar reconstructions of the left anterior descending coronary artery showing high-density artifacts generated by metallic stent struts (arrows) that prevent the correct evaluation of the proximal stent lumen. (C,D) Multidetector computed tomographic multiplanar reconstructions of 2 overlapping stents implanted in the proximal right coronary artery showing artifacts due to slice misalignment caused by heart rate variations during the scan.

Only gold members can continue reading. Log In or Register to continue

Share this:

• Click to share on Twitter (Opens in new window)
• Click to share on Facebook (Opens in new window)

Related

Related posts:

1. Comparative Analysis of Cardiac Magnetic Resonance Viability Indexes to Predict Functional Recovery After Successful Percutaneous Coronary Intervention in Acute Myocardial Infarction
2. N-Terminal Pro–Brain Natriuretic Peptide and Outcomes in Patients Undergoing Surgical Ventricular Restoration
3. The RecordAF Study: Design, Baseline Data, and Profile of Patients According to Chosen Treatment Strategy for Atrial Fibrillation
4. Erratum for Shah et al. “Usefulness of Various Plasma Biomarkers for Diagnosis of Heart Failure in Children With Single Ventricle Physiology” Am J Cardiol 2009;104:1280–1284

Stay updated, free articles. Join our Telegram channel

Join