Imaging challenges of carotid artery in-stent restenosis
- 1Department of Neuroradiology, Massachusetts General Hospital, Harvard Medical School, Boston, Massachusetts, USA
- 2Department of Neurology and Psychiatry, Sapienza University of Rome, Rome, Italy
- 3Department of Interventional Neuroradiology, Massachusetts General Hospital, Harvard Medical School, Boston, Massachusetts, USA
- 4Neurovascular Laboratory, Massachusetts General Hospital, Harvard Medical School, Boston, Massachusetts, USA
- Correspondence to Dr J M Romero, Department of Neuroradiology, Massachusetts General Hospital, Harvard Medical School, Boston, MA 02114, USA;
- Received 6 December 2012
- Accepted 7 December 2012
- Published Online First 7 January 2013
Internal carotid artery stenosis is an established risk factor for stroke. Therefore, carotid artery revascularization has an important role in the prevention and treatment of stroke. For the treatment of carotid artery stenosis, carotid artery stenting (CAS) has currently gained acceptance as a safe alternative to carotid endarterectomy (CEA), particularly in patients at high surgical risk. Duplex ultrasonography (DUS) is a non-invasive technique with standardized criteria used for the diagnosis of carotid atheromatous disease as well as for the detection of restenosis after carotid revascularization. Restenosis rates vary widely in the literature. Different studies indicated that restenosis following CAS was higher than following CEA, although the Carotid Revascularization Endarterectomy Versus Stenting Trial (CREST) reported similar restenosis frequency after 2 years of follow-up. Given these results, DUS may have a significant role in the follow-up of CAS patients. Conventional carotid artery DUS velocity criteria are thought to be less accurate in patients who have undergone CAS and many authors proposed different criteria for grading in-stent restenosis (ISR). This review presents the advantages of CAS, the current practice of carotid revascularization, CAS complications and risks, and DUS criteria for carotid artery ISR. After analyzing multiple relevant studies that proposed sonographic criteria for grading at least 70% ISR, we can conclude that a peak systolic velocity value of 300–350 cm/s could be used as a relatively good and sensitive predictor of high grade ISR.
Internal carotid artery (ICA) stenosis accounts for approximately 10–15% of ischemic strokes. While most plaques remain asymptomatic, some undergo disruption causing thrombosis and secondary embolization.1 Duplex ultrasonography (DUS) is a non-invasive technique with standardized criteria used for the diagnosis and long term follow-up of patients at high risk of atheromatous disease of the carotid arteries. DUS is also useful for follow–up after carotid intervention for the detection of restenosis. Table 1 shows the restenosis rates of different studies, indicating that restenosis following carotid artery stenting (CAS) is higher than following carotid endarterectomy (CEA).2–4 The most recent trial, guided by the investigators of the Carotid Revascularization Endarterectomy Versus Stenting Trial (CREST),5 reported a similar restenosis frequency after CAS and CEA after 2 years of follow-up. Given these results, DUS may have a significant role in the follow-up of CAS patients. Conventional carotid artery DUS velocity criteria are thought to be less accurate in patients who have undergone CAS.6 Tables 2 and 3 demonstrate the grading criteria many authors have proposed for in-stent restenosis (ISR).
In this review, we discuss the advantages of CAS, the current practice of carotid revascularization, the complications and risks associated with the stenting procedure, and the DUS criteria for carotid artery ISR. We also briefly discuss other imaging techniques that can be used to diagnose ISR.
Epidemiology and costs of stenting in the USA
Approximately 780 000 people/year have a new or recurrent stroke with an estimated annual cost of US$62.7 billion in the USA.7 Based on the US Nationwide Inpatient Sample for 2005–2007, more than 130 000 carotid interventions/year were performed, 88.6% of which were CEAs and 11.4% were CAS; 92% of cases were male asymptomatic patients aged ≥70 aged. The median total hospital charges for each CAS procedure was $12 000–$13 500 more than each CEA (p<0.0001). Of interest, the postoperative stroke and mortality rates were statistically significantly higher among the CAS group compared with the CEA group in years 2005 and 2006, while in 2007 no statistically significant difference was observed between the two groups because of an improvement in the CAS results.8 An overall decrease in the rates of carotid revascularization from1998 to 2008 was also noted, with a reduction by 36% of CEA and an increase by 5% in CAS.9 A projected 10 year model based analysis10 of the costs and quality adjusted life expectancy demonstrated that CAS would result in a small mean incremental cost of US$524/patient and a minimal reduction in quality adjusted life expectancy of 0.008 years compared with CEA.
Current practice of CEA and stenting in the USA
Based on the results of the North American Symptomatic Carotid Endarterectomy Trial (NASCET),11 the European Carotid Surgery Trial (ECST),12 and the Asymptomatic Carotid Atherosclerosis Study (ACAS),13 CEA has been established as the gold standard treatment for patients with symptomatic severe stenosis (≥70%) (class I—level of evidence B) and for asymptomatic patients who have more than 70% stenosis with a low risk of perioperative stroke, myocardial infarction (MI), or death (class IIa—level of evidence A).14
A pooled analysis of NASCET and ECST15 showed significant benefits of CEA compared with medical treatment after 5 years of follow-up, both in the group with severe (70–99%) stenosis (absolute risk reduction of 16%, p<0.001) and moderate (50–69%) stenosis (absolute risk reduction of 4.6%, p=0.04). In patients with mild (≤50%) stenosis, surgery was not beneficial (absolute risk reduction of 3.2%, p=0.6). On the other hand, ACAS13 demonstrated CEA to be the treatment of choice for asymptomatic moderate to severe stenosis, reporting a 5 year risk of ipsilateral stroke and any perioperative stroke or death of 5.1% for surgical patients and 11.0% for patients treated medically. Moreover, the benefit of CEA versus medical therapy was confirmed at 5 and at 10 years by the Asymptomatic Carotid Surgery Trial-1 (ACST-1).16
The influence of patient age on surgical risk is unclear. NASCET showed that patients aged >75 years with 50–99% symptomatic carotid stenosis benefited more from CEA than younger patients17 but this result was not confirmed by ACST-116 or ECST.18 Some authors reported higher risks of complications among older patients undergoing CEA,19 while others showed similar risks of perioperative stroke and death between patients >75 years of age with few cardiovascular risk factors and younger patients.20
In the ESCT study, women undergoing CEA were associated with a higher operative risk than men.18 In ACAS-1, CEA for asymptomatic carotid stenosis had significant benefit both for men and women16 while in NASCET, women with >70% symptomatic ICA stenosis had similar long term benefit from CEA as men, although the perioperative risks were higher.21
The following randomized controlled trials have compared CAS and CEA. The Carotid and Vertebral Artery Transluminal Angioplasty Study (CAVATAS)22 involved 504 patients with symptomatic carotid artery stenosis who were randomly assigned to CAS or CEA. Although the combined stroke or death rate at 30 days was 10% in both groups, the CAS group experienced less ‘minor complications’, including cranial neuropathy, major hematoma, MI, and pulmonary embolism. Stroke or death at 3 years was similar in both groups (14.3% vs 14.2%).
The Stenting and Angioplasty with Protection in Patients at High Risk for Endarterectomy (SAPPHIRE) trial,23 which randomized 334 patients at high surgical risk to undergo either stenting or CEA, demonstrated a lower incidence of stroke, death, or MI within 30 days after CAS compared with after CEA (12.2% vs 20.1%; p=0.05), and the incidence of reintervention was higher after CEA than after CAS (4.3% vs 0.6%; p=0.04).
The Carotid Revascularization with Endarterectomy or Stenting Systems (CARESS) trial24 enrolled 397 patients (143 CAS and 254 CEA) and reported no significant differences in combined death/stroke rates between the CAS and CEA group at 30 days (2.1% vs 3.6%) or at 1 year (10.0% vs 13.6%). Similarly, there was no significant difference in the composite endpoint of death, stroke, or MI at 30 days (2.1% vs 4.4%) or at 1 year (10.9% vs 14.3%). A 4 year analysis25 showed no significant differences between the two groups in the occurrence of any stroke (8.6% vs 9.6%) or in the composite endpoints described above (21.7% vs 27.0%). In addition, the incidence of the composed endpoint complications was higher in the CEA group among patients <80 years of age, but there were no differences between the two groups in patients aged >80 years.
A meta-analysis26 of pooled data of 3433 patients with symptomatic carotid stenosis from the Endarterectomy versus Angioplasty in Patients with Symptomatic Severe Carotid Stenosis (EVA-3S) trial, Stent Supported Percutaneous Angioplasty of the Carotid Artery versus Endarterectomy (SPACE) trial, and the International Carotid Stenting Study (ICSS), suggested that CAS was associated with a higher periprocedural risk of stroke or death rate than CEA (8.9% vs 5.8%; p=0.27). This result was attributed to a significant difference in stroke risk in the two treatment groups. The occurrence of non-disabling strokes was higher in the CAS group while there were no significant differences in the occurrence of disabling stroke or death (4.8% vs 3.7%; p=0.94), or in all-cause death (1.9% vs 1.3%; p=0.07). However, the investigators found a significantly higher risk of stroke or death in the CAS group in patients older than 70 years.
The recently published CREST27 is the largest trial comparing the utility of CEA and CAS. This trial enrolled patients with either symptomatic or asymptomatic carotid artery stenosis. Similar rates of any periprocedural stroke, MI, or death, and subsequent ipsilateral stroke, between CAS and CEA (p=0.38) were noted. The stenting group presented a higher incidence of periprocedural stroke (4.1% vs 2.3; p=0.01) and lower incidence of periprocedural MI (1.1% vs 2.3%; p=0.03). There was an increased risk of an adverse event after CAS in patients aged ≥70 years. Also, periprocedural risk of events seemed to be higher in women who had CAS than those who had CEA whereas there was little difference in men.28
A pooled analysis of CREST, EVA-3S, SPACE, and ICSS29 showed substantially more periprocedural deaths with CAS than with CEA (3.72% vs 1.04%; p=0.04) and more periprocedural stroke and death (2.25 vs 1.40; p<0.0001).
Review of these trials leaves room for differing perspectives. What is close to universally agreed to is that CAS is a reasonable alternative to CEA to treat high grade carotid stenosis in patients with a high perioperative risk (class I—level of evidence B) and with neck anatomy unfavorable for arterial surgery (eg, high position of carotid bifurcation, previous cervical radiation therapy) (class IIa—level of evidence B).14
Further studies should be carried out with a sufficient numbers of women and older patients to define the efficacy of CAS and CEA in different demographic subsets of patients with ICA stenosis.
New stent technology and experienced operators will surely improve the outcomes of CAS. A trial with these improvements is required.
Major complications associated with CAS
CAS procedure may be associated with several complications, the main ones being brain embolism and ISR.
New diffusion weighted imaging brain lesions occur with a higher incidence in patients submitted to CAS than in patients treated with CEA.30 However, the frequency of ischemic events can be reduced by using distal or proximal embolic protection devices.31 Several authors have reported that plaques associated with large lipid pools or soft extracellular lipid, which appeared soft and uniformly or predominantly echolucent at DUS evaluation, and plaques with surface irregularity were more predisposed to embolization.32–34
Restenosis after CAS likely results from vessel trauma with subsequent neointimal hyperplasia which occurs within 24 months of the postoperative period or de novo neoatherosclerosis thereafter.35 Neointimal hyperplasia is a stable condition, usually peaking 3 months after surgery, and followed by a regression. It develops from the continuous interaction between the stent and the vessel which causes physical irritation, endothelial dysfunction, and chronic inflammation.36 This restenosis is secondary to hypersensitivity reactions, chronic inflammation, late malapposition related to positive remodeling, and delayed arterial healing with incomplete endothelialization.37 Although few studies reported data about the exact time of ISR occurrence, it seems to occur mostly within the 1 year after CAS.3 ,38 Clinical ISR complication rates vary widely in the literature; stroke rates were between 0% and 25% and death rates from 0% to 11.1%.39
CAVATAS found an increased rate of restenosis ≥70% at 1 year and 5 years in the CAS group compared with the CEA group (see table 1). The incidence of ipsilateral stroke or transient ischemic attack was significantly higher in patients with restenosis ≥70% compared with those without ISR.2 A higher restenosis incidence in the CAS group was also seen in the EVA-3S4 and SPACE3 trials. Also, the CARESS trial showed that CAS had a twofold higher restenosis rate (defined as ≥75% narrowing documented by DUS or symptomatic narrowing >50% that required secondary treatment) compared with CEA at 4 years (14.7 vs 5.9; p=0.014).25 An analysis from CREST5 showed that the cumulative 2 year rates of severe (≥70%) restenosis or occlusion were low and did not differ between the CEA and CAS groups (table 1). Although 70% stenosis is known to increase the risk of stroke in native atheromatous disease, the risk of stroke for ISR ≥70% is still unknown. Further trials are recommended to asses this risk.
There were some studies5 ,39–41 that identified several clinical risk factors in the development of ISR: older age, female sex, dyslipidemia, diabetes, tobacco, and peripheral artery occlusion disease. Technical and postprocedural risk factors included multiple stents implantation, postprocedural residual stenosis, and previous CEA. Close DUS follow-up is therefore appropriate in patients with these ISR risk factors.
DUS criteria for stented carotid stenosis
Doppler ultrasonography is a reliable tool for evaluation of ISR and correlates well with angiography.42 ,43 Doppler blood flow velocity criteria and the presence of plaque on gray scale and/or color DUS images are the parameters that should be used when diagnosing and grading ICA stenosis.44 DUS velocity threshold criteria for the diagnosis of ISR have not been well established. Differentiating between moderate and severe ISR plays an important role in deciding the need for further interventions.
Different reports6 ,43 ,45 indicate that CAS may cause changes in the stented artery compliance and blood flow, reducing compliance and causing an increase in the standard DUS velocity criteria (table 4), in the absence of angiographically proven restenosis. Furthermore, DUS velocities appear higher when a closed cell stent is placed than when using an open cell stent.46
Several studies were carried out to establish sonographic cut-off criteria for the diagnosis of different degrees of carotid ISR, using peak systolic velocity (PSV), end-diastolic velocity (EDV), and internal to common carotid artery PSV ratio (I/C). PSV is the most accurate criterion of ISR. In fact, it increases according to the increment in the stenosis.47 EDV and I/C are additional criteria and may increment the accuracy of the DUS evaluation.44
Systemic literature review
We searched Medline and PubMed for articles published from January 1990 to September 2012, with the keywords ‘carotid artery’, ‘in-stent’, ‘restenosis’, ‘Doppler’, and ‘criteria’. To be included in our analysis, studies had to fulfill the following criteria: (1) articles had to be written in English; (2) more than 25 stented carotid arteries had to be investigated; (3) patients with critical stenosis were studied with angiography or CT angiography (CTA); and (4) angiographic or CTA degree of carotid ISR had to be graded by the NASCET criteria. The abstract of each article was carefully studied and, if it matched the criteria, the full text was retrieved. Furthermore, all reference sections of those articles containing data on restenosis rates were checked. Carotid occlusions were counted as restenosis and PSV, EDV, and I/C PSV ratio values were calculated in relation to the number of arteries studied.
A total of 14 studies matched the above mentioned criteria (tables 2, 3).42 ,48–60 We analyzed 12 of these studies that proposed sonographic criteria for grading at least 70% ISR. This threshold was chosen because it is mostly used by vascular centers for intervention. Among these studies, high grade stenosis was defined differently as a diameter reduction of 70%, 75%, or 80% but its PSV threshold was relatively consistent at 300–350 cm/s (figure 1), EDV threshold varied slightly more at 90–140 cm/s, and the I/C ratio varied from 3.8 to 4.75. However, these parameters suffer from several exceptions. As summarized in table 2, Peterson et al50 adopted a PSV ≥170 cm/s and an EDV cut-off value of 120 cm/s in detecting ISR ≥70%, in association with an increase in PSV and EDV, respectively, of 50% compared with immediate postoperative values, reporting 100% sensitivity and 100% specificity. Chi et al53 used a PSV of 450 cm/s to identify ≥70% ISR, with a sensitivity of 67% and a specificity of 100%. Similar findings held true for an I/C ratio of 4.3. They found also that EDV had insignificant statistical relation to angiographic stenosis. Cumbie et al59 determined the threshold velocity criteria to detect ISR ≥80% using receiver operating characteristic (ROC) curve analyses. A PSV ≥205 cm/s yielded sensitivity and specificity of 100% and 96%, respectively, whereas an I/C ratio ≥2.6 yielded sensitivity and specificity of 100% and 94%, respectively. They also found that EDV was not a good predictor of ISR ≥80%, while the combination of PSV and I/C yielded 100% sensitivity and 97% specificity. Heck60 concluded that ISR ≥70% can reliably be detected using cut-off values of PSV ≥200 cm/s in association with EDV ≥125 cm/s or I/C ratio ≥4, without reporting any statistic analysis.
Among all of the studies analyzed, the prospective study of Setacci et al56 is notable for being the largest one to date, involving 814 patients over 6 years. Additionally, they had extraordinary follow-up, with DUS examination being performed at months 1, 3, 6, 9, and 12, and then yearly in the same vascular laboratory, with the same ultrasound machines. Only patients with a PSV ≥200 cm/s or a PSV value greater than three times the baseline value underwent angiography, and blinded operators compared the results of all angiographies to the ultrasound velocity data within 48 h, with a total of 6427 DUS examinations and 1123 angiographies performed. They proposed a PSV value of ≥300 cm/s in association with EDV ≥140 cm/s and an I/C ratio ≥3.8 to detect ISR ≥70%. Using these parameters, the authors identified 95 patients with ISR (73 with ISR ≥50%, 22 with ISR ≥70%). ROC analysis was performed to evaluate the sensitivity and specificity for the identified threshold values of PSV, EDV, and ICA/common carotid artery (CCA) in the case of stenosis ≥70%. They reported PSV, EDV, and ICA/CCA significant values from the areas under the ROC curve. In detail, area under the curve for PSV was 0.99, for EDV 0.98 and for ICA/CCA 0.99. Using the combination of the three parameters, a specificity of 99% and a sensitivity of 98% for ISR ≥70% were achieved. The Pearson coefficient of correlation (R) between any velocity criteria (PSV, EDV, I/C) and the percentage of angiographic ISR was always statistically significant (p=0.0001) (PSV vs % stenosis: R=0.55; EDV vs % stenosis: R=0.49; I/C vs % stenosis: R=0.71). In addition, they studied the PSV trend in relation to the percentage of angiographic stenosis, identifying a subpopulation with ISR ≥50% (11.7% of patients) with R coefficient of 0.87 (p<0.0001). Despite the large number of patients examined, the weak point of the study was the small number of patients with high grade stenosis.
A PSV threshold of ≥300 cm/s was also used by both EVA-3S4 and CREST5 to identify the frequency of ISR ≥70%. CAVATAS chose a PSV threshold of ≥210 cm/s2 while the SPACE trial did not specify the criteria used to quantify ISR.
Based on the current available literature, we can conclude that a PSV value of 300–350 cm/s could be used as a relatively good and sensitive predictor of high grade ISR.
DUS, CTA, and MRA imaging in detecting ISR: pearls and pitfalls
DUS is a reproducible, affordable, non-invasive method that can be used as both the initial diagnostic test to detect carotid artery atherosclerotic disease and to assess restenosis during follow-up after the CAS procedure.42 ,43 For evaluating restenosis, DUS may be performed at 1 and 6 months, and annually thereafter.14 Given the similar rates of restenosis of CEA and CAS seen in the CREST trial,5 it is our practice to recommend a follow-up post-CAS similar to the one used for post-CEA. DUS can be hampered by the technique's limitations, such as anatomic difficulties in reaching the distal end of the stent.61 However, because of the good sensitivity, non-invasiveness, and cost benefit, carotid DUS remains the first imaging tool for follow-up after CAS42 (figures 2, 3). As previously mentioned, the high velocity values of the stented carotid artery can lead to a wrong diagnosis. Immediate post-stenting DUS examination provides a baseline value for future follow-up comparisons, in order to reduce the number of false restenoses. The examination is operator and machine dependent; therefore, the velocity recorded in one laboratory may differ significantly from other laboratories. Thus each laboratory needs to develop internally validated Doppler criteria. Correlative angiograms for the validation are not available or realistic in all institutions. For this reason, recommended Doppler thresholds for diagnosis of ICA stenosis should be assessed.44 In this review, we suggest a PSV value of 300–350 cm/s as the threshold value to detect high grade restenosis, and it can be applied at laboratories that cannot validate their own Doppler criteria. If the diagnostic information obtained by performing DUS is not reliable, CTA may be helpful for surveillance after CAS. It should also be required to confirm the presence of high grade ISR before reintervention, although digital subtraction angiography (DSA) is the gold standard for the evaluation of occlusive carotid disease. However, DSA is associated with access site complications, risk of thromboembolism, and non-ionic iodinated contrast adverse reactions. It has a relatively high risk of morbidity and mortality, ranging from 1% to 4% in patients with atherosclerosis,62 and is less commonly done in the workup of carotid atherosclerotic disease.
CTA is a non-invasive technique with high resolution and quick acquisition times. It demonstrates excellent sensitivity and specificity in the assessment of large vessel occlusion. It is the imaging technique of choice in the case of tortuous carotid, severe calcification, short neck, and high bifurcation63 (figures 4, 5). For severe stenosis (≥70%), CTA was found to be reliable, with sensitivity of 100% and specificity of 63%; the negative predictive value of CTA demonstrating <70% carotid artery stenosis was 100%.64 In a comparison study between CTA, DUS, and DSA in non-stented carotid arteries, an excellent correlation was demonstrated when comparing the results of CTA and DSA grading of stenosis and then between DUS and DSA.65 However, CTA has some limitations related to the need for external beam radiation and injection of intravenous iodinated contrast. Technical limitations include beam hardening from the metallic stent, which may make evaluation of the residual lumen difficult. New techniques such as spectral or dual source CT angiograms and postprocessing software may resolve this limitation.
MR angiography (MRA) is a safe, non-invasive, and high resolution imaging technique for carotid artery stenosis. In a comparison study of MRA, DUS, and DSA in non-stented carotid arteries, Nederkoorn et al66 demonstrated that MRA had a pooled sensitivity and specificity of 95% and 90%, and DUS of 86% and 87%, respectively, in detecting stenosis ≥70% versus <70%. For occlusion, the sensitivity and specificity for MRA were 98% and 100% and for DUS 96% and 100%, respectively. They concluded that MRA has a better discriminatory power compared with DUS in diagnosing stenosis ≥70%, and is a sensitive and specific test compared with DSA in the evaluation of carotid artery stenosis. For detecting occlusion, both DUS and MRA are very accurate. Pitfalls in MRA evaluation of extracranial carotid disease include overestimation of stenosis and inability to discriminate between subtotal and complete arterial occlusion. Contrast enhanced MRA appears to overcome the limitations seen with unenhanced MRA; however, for the screening and identification of patients with ICA stenosis >70%, contrast enhanced MRA does not offer significant advantages over two-dimensional time of flight MRA67 (figures 6, 7). Metallic related artifacts can hamper the use of MRA for evaluation of ISR. Thus CTA is to be preferred to MRA for surveillance of carotid in-stent restenosis.68
To summarize this review of the literature, DUS is the first test for surveillance of stented carotid artery. If the results of DUS are inconclusive, CTA is required. Different results with either of these techniques should be confirmed by DSA.68
The use of CAS for the treatment of carotid artery stenosis has gained acceptance currently as a safe alternative to CEA, particularly in patients at high surgical risk. New trials should be performed for defining the efficacy of revascularization in different demographic subsets. Carotid artery ISR is one of the major challenges associated with the CAS procedure. Risk factors or predictors of ISR should be identified. According to the existing literature and guidelines, DUS before discharge and at 1 and 6 months and then annually can be recommended. Although overestimation of post-CAS restenosis is a potential risk in a stented carotid artery, DUS remains a sensitive tool in this population. A PSV value of 300–350 cm/s may be used as the threshold criterion of high grade restenosis. Restenosis is generally a benign condition that does not require revascularization except in selected cases, such as when restenosis progresses to preocclusive grade or determinates neurological symptoms. Under these circumstances, it may be justifiable to repeat revascularization, either by CEA in the hands of an experienced surgeon or by CAS in the case of hostile neck anatomy.14 Further studies on the correct management of ISR continue to be warranted.
Contributors The study was designed by JMR and JAH. Analysis and interpretation of the data were performed by RP. RP drafted the article. JMR, JAH, and RP revised it critically for important intellectual content. All authors read and approved the final version to be published.
Competing interests None.
Provenance and peer review Not commissioned; not externally peer reviewed.