Introduction
Coronary stents are used to reduce restenosis after balloon angioplasty [1]. However, in-stent restenosis can occur in 20—35% of cases after implantation of bare-metal stents and 5—10% for drug-eluting stents, as demonstrated by intravascular ultrasound [2].
The increasing number of patients who had coronary stent implantation yearly necessitate a reliable method of diagnosis. Invasive coronary angiography is the gold standard for detection of in-stent restenosis. However, coronary angiography has limitations due to its invasiveness and association with potential risks of morbidity and mortality.
Multislice computed tomography (MSCT) has emerged as a non-invasive tool in diagnosis of coronary artery disease with high accuracy [3].
However, imaging of coronary stents by MSCT is more difficult than native coronary artery due to various artefacts complicating interpretation [4]. Visibility of stent lumen is influenced by many factors including site of the stent, angulation, thickness of struts, composition and dimensions of stent [5, 6].
Although many studies have shown that MSCT may be used to evaluate stent patency, blooming artefacts markedly affect this evaluation [7]. Thus, MSCT is not considered for routine clinical evaluation of in-stent restenosis with exception of assessment of stents in the left main coronary artery [8].
We assume an excellent visibility of proximal LAD stent [9, 10] because it has nearly the same angulation as the left main coronary artery [11, 12] and relatively protected from motion artefact [5, 6]. Additionally, it is a continuation of the left main artery and usually stented with stent >3 mm.
This study was conducted to assess the accuracy of multislice computed tomography angiography in diagnosis of in-stent restenosis in proximal left anterior descending artery compared to conventional coronary angiography.
Material and methods
All procedures were following the ethical standards of our institutional and/or national research committee, Helsinki declaration 1964 and its later amendments or comparable ethical standards. Informed consent was obtained from all participants included in the study.
This prospective observational study was conducted at the Cardiology Department of the Ain Shams University Hospitals for the period from January 1, 2017 to November 30, 2017. There were 50 patients scheduled for coronary angiography and CT angiography for typical angina pectoris after previous stenting of proximal LAD (>3 months ago) and suspected in-stent restenosis.
Exclusion criteria: acute coronary syndrome, hemodynamic instability, chronic kidney disease (serum creatinine >1.4 mg/dl or glomerular filtration rate <50 ml/min/1.73 m2), allergy to contrast agent, atrial fibrillation or other rhythm irregularity, heavy calcification at the site of the stent, multiple motion artefacts, inability to perform breath holds or another patient status. All patients signed an informed consent, and the study protocol was approved by local ethics committee.
Beta-blocker (oral atenolol 100—200 mg or i.v. propranolol 10—20 mg) was used if heart rate was ≥65 beats/min.
All patients were examined under a 64-slice scanner (Aquilion; Toshiba Medical Systems, Otawara, Japan) using the following parameters: collimation width 64´0.6 mm, rotation time 350 ms, tube voltage 120 kV, effective tube current 430 mA and pitch 0.2. Tube current modulation was deactivated to allow maximal flexibility in choosing image reconstruction intervals.
All scans were contrast-enhanced within one single breath hold using a 50 ml bolus of non-ionic iso-osmolar iodine-containing contrast (Ultravist 370 injection; Bayer Schering Pharma AG, Leverkusen, Germany). Contrast agent was injected into antecubital vein using the dual-head injector at a rate of 5 ml/s followed by 50 ml of saline chaser.
The entire volume of the heart was acquired during one breath-hold and retrospective ECG gating.
Axial images were reconstructed in mid-diastole at different points of cardiac cycle with minimal motion artefacts. A slice thickness of 0.75 mm with a 0.4-mm increment, smooth and sharp reconstruction kernels were used. Images were transferred to a workstation (Vitrea Fx, Vital Images, USA).
Only proximal segment of left anterior descending artery was included in the study. An expert performed blind analysis of images. Significant ISR was diagnosed in luminal diameter reduction ≥50% or total occlusion of stent in axial slices, cross-sectional multi-planar reconstruction and curved-planar reconstruction images (fig. 1, a).
Coronary angiography
At least 4 different views were obtained. Assessment of proximal LAD in-stent restenosis was done using quantitative coronary angiography in the Right Anterior Oblique Cranial view. Stenosis of coronary stent ≥50% was defined as significant ISR (fig. 1, b).
Statistical analysis
All data were analyzed using SPSS 18 software package (SPSS Inc., Chicago, Illinois, USA). Quantitative variables were expressed as mean ± standard deviation and range. Incidence is presented as percentages. Quantitative variables were compared by Student’s t-test, while qualitative variables were compared by chi-square test or Fisher’s exact test as appropriate. P-value <0.05 was considered significant.
Results
This study prospectively assessed 50 patients who had previous PCI in proximal segment of the left anterior descending artery. All patients underwent CT and coronary angiography to assess in-stent restenosis.
Mean age of patients was 55.7±10.4 years (range 45—76). Most patients were males (86%).
The most common risk factor was smoking (76% of patients). Dyslipidemia was the second common risk factor (72%). Diabetes mellitus (DM) and hypertension were observed in 50% and 52% of patients, respectively (table 1).
Variable | Number | Percent | |
Hypertension: | |||
negative | 24 | 48 | |
positive | 26 | 52 | |
Diabetes mellitus: | |||
negative | 25 | 50 | |
positive | 25 | 50 | |
Dyslipidemia: | |||
negative | 14 | 28 | |
positive | 36 | 72 | |
Smoking: | |||
negative | 12 | 24 | |
positive | 38 | 76 |
Stent diameter <3 mm was found in 30% of patients, >3 mm — 70% of cases.
We found that 42% of patients had significant ISR according to coronary angiography data. Non-significant ISR (<50%) was observed in 58% of patients (table 2).
Variable | Number | Percent |
Diameter of stent: | ||
<3 mm | 15 | 30 |
>3 mm | 35 | 70 |
ISR at QCA: | ||
non-significant | 30 | 60 |
significant | 20 | 40 |
ISR at CT angiography: | ||
non-significant | 26 | 52 |
significant | 24 | 48 |
According to CT angiography data, 48% of patients had significant ISR and 52% had non-significant ISR. An expert who was blinded from the results of coronary angiography and the patient clinical data (table 2) estimated CT data.
There was a correlation between stent diameter and significant ISR. We found 21 non-significant and 14 significant ISRs among patients with stent diameter >3 mm according to coronary angiography and CT data, respectively.
In contrast, we found different data in patients with stent diameter <3 mm (9 non-significant and 6 significant ISR on coronary angiography vs. 5 non-significant and 10 significant ISR on CT angiography).
We compared the results of coronary angiography and CT angiography. Coronary angiography revealed significant ISR in 20 stents, CT angiography — in 24 stents. Sensitivity of CT angiography was 100%, specificity — 86%, positive predictive value — 83%, negative predictive value — 100%, test accuracy — 92%.
When these results were correlated with stent diameter, sensitivity, specificity, positive predictive value and negative predictive value were 100, 55, 60 and 100%, respectively in patients with stent diameter <3 mm. Overall test accuracy was 73%.
Regarding stents >3 mm, CT has 100% value for sensitivity, specificity, positive predictive value and negative predictive value. Overall accuracy in detection of significant ISR was 100% (table 3, fig. 2).
Variable | Sensitivity, % | Specificity, % | Positive predictive Value, % | Negative predictive Value, % | Accuracy, % |
CT angiography | 100 | 86 | 83 | 100 | 92 |
Less than 3 mm | 100 | 55 | 60 | 100 | 73 |
Over 3 mm | 100 | 100 | 100 | 100 | 100 |
Discussion
Widespread use of drug-eluting stents for complex lesions and multiple-vessel coronary artery disease has led to a great number of patients with ISR. This aspect demands an accurate diagnostic technique to assess stent patency.
In this study, we have enrolled 50 patients who underwent CT and CA to assess proximal in-stent restenosis of the left anterior descending artery. Diameter ≥3 mm was noted for 70% of stents, <3 mm — 30% of stents.
CT angiography showed 100% accuracy for stents >3 mm. Regarding stents <3 mm, positive predictive valve declined from 100 to 60% and caused tends to over-estimation of ISR. In our study, sensitivity of CT for ISR detection was 100%, specificity 55%, positive predictive value 60% and negative predictive value 100%.
Our results support previous report by Das et al. [9]. These authors showed similar values of sensitivity (96.9%), specificity (88.0%), positive predictive value (77.5%), negative predictive value (98.5%) and accuracy (91%) for detection of in-stent restenosis by CT angiography [9].
In contrast, lower sensitivity and positive predictive value (75% and 71%, respectively) were reported by Carbone et al. [13]. This could be explained by a blooming artefact resulting non-interpretable segments. They exclude 9 out of 12 stented segments with diameter of 2.5 mm and 10 out of 23 stented segments with diameter of 2.75 mm due to blooming artefacts. Additionally, our study only included stents in proximal part of the left anterior descending artery characterized by favorable visibility compared to other segments. This was proven by Das et al. [9] who emphasized worse in-stent lumen visibility for the right coronary artery and circumflex artery. This is because proximal LAD is relatively protected from motion artefact and corresponds to scanning direction. This segment is usually passing in an axial plane whereas proximal circumflex artery runs in a slightly less favorable craniocaudal longitudinal plane [6]. Additionally, Dawoud et al. [10] reported that stent location affects visibility. Stents in proximal segments of the right coronary artery, LAD and circumflex artery were better visualized than those deployed within the distal segments of the same arteries.
ISR detection is affected by image quality as shown by Haraldsdottir et al. [14]. In their study, 4 stents with significant stenosis 70—100% were incorrectly classified as patent stents after CT angiography. None of these four stents had good or excellent image quality. These data resulted low sensitivity (27%) and high specificity (95%) of this method. Positive and negative predictive values were 67 and 78%, respectively.
Yoshimura et al. [15] used both visual assessment and quantitative stent restenosis index. The last one considered CT density at proximal and distal artery lumen from the stented region that is then corrected according to stent diameter. Sensitivity, specificity, positive predictive value, negative predictive value and accuracy of stent restenosis index were 82, 93, 64, 97 and 91%, respectively. These values were higher compared to those obtained by visual assessment (78%, 75%, 35%, 95% and 76%, respectively). However, there was no statistical significance that could be attributed to a small number of the stents recruited.
Oncel et al. [16] detected ISR using 64-slice CT scanner. Sensitivity, specificity, positive and negative predictive values were 89%, 95%, 94% and 90%, respectively. These values were applied for all coronary segments, rather proximal LAD only as in our study.
The influence of stent type on ISR assessment could not be detected in our study, as drug-eluting stents were implanted only. Moreover, this study was not powered to answer such sub-analysis and differences.
Our study is limited by assessment of stents in proximal part of the left anterior descending artery only. Other segments of coronary arteries could be more difficult to access due to blooming effect and motion artefacts.
Another limitation was small sample size, in addition to exclusion of patients with cardiac arrhythmias, rapid heart rate, renal failure and contraindication for injection of iodinated contrast agent. In practice, these constitute a significant proportion of patients with coronary stents, and their exclusion is likely to overestimate stent assess ability.
Conclusion
Accordingly, our study considered CT is a reliable and perfect tool to rule out the presence of ISR in proximal LAD with high negative predictive value. Thus, patients can avoid the risk of invasive coronary angiography. However, CT accuracy depends on stent diameter (100% for stent diameter >3 mm and less accuracy for stent diameter <3 mm).
No conflict of interests to declare.
Funding
This research did not receive any specific grant from funding agencies in the public, commercial or not-for-profit sectors.