- Email: [email protected]
In vitro evaluation of 56 coronary artery stents by 256-slice multi-detector coronary CT
European Journal of Radiology 80 (2011) 143–150
Contents lists available at ScienceDirect
European Journal of Radiology journal homepage: www.elsevier.com/locate/ejrad
In vitro evaluation of 56 coronary artery stents by 256-slice multi-detector coronary CT Henning Steen a,∗,1 , Florian André a,1 , Grigorios Korosoglou a,1 , Dirk Mueller b , Waldemar Hosch c,1 , Hans-Ulrich Kauczor c,1 , Evangelos Giannitsis a,1 , Hugo A. Katus a,1 a
University of Heidelberg, Department of Cardiology, Im Neuenheimer Feld 410, Heidelberg 69120, Germany Philips GmbH Healthcare Division, Luebeckertordamm 5, Hamburg 20099, Germany c University of Heidelberg, Department of Diagnostic and Interventional Radiology, Im Neuenheimer Feld 410, Heidelberg 69120, Germany b
a r t i c l e
i n f o
Article history: Received 30 March 2010 Received in revised form 27 July 2010 Accepted 2 August 2010 Keywords: Multislice CT Coronary stents CT angiography Artifacts Coronary artery disease In-stent Restenosis
a b s t r a c t Objective: We sought to investigate stent lumen visibility of 56 coronary stents with the newest 256multi-slice-CT (256-MDCT) technology for different reconstruction algorithms in an in vitro model. Background: Early identification of in-stent restenosis (ISR) is important to avoid recurrent ischemia and prevent acute myocardial infarction (AMI). Since angiography has the disadvantage of high costs and its invasiveness, MDCT could be a convenient and safe non-invasive alternative for detection of ISR. Material and methods: Percentages of in-stent lumen diameter and in-stent signal attenuation (measured as contrast-to-noise ratio (CNR)) of 56 coronary stents (group A ≤2.5 mm; group B = 2.75–3.0 mm; group C = 3.5–4.0 mm) were evaluated in a coronary vessel in vitro phantom (iodine-filled plastic tubes) employing four different reconstruction algorithms (XCD, CC, CD, XCB) on a novel 256-MDCT (Philips-iCT, collimation = 128 mm × 0.625 mm; rotation time = 270 ms; tube current = 800 mA s with 120 kV). Analysis was conducted with the semi-automatical full-width-at-half-maximum (FWHM) method. P-values <0.05 were regarded statistically significant. Results: In-stent lumen diameter >60% for group C stents was significantly larger and CNR was significantly lower (both p < 0.05) for sharp kernels (CD; XCD) when compared to groups A/B. The FWHM-method showed significantly smaller in-stent lumen diameter (p < 0.05) when compared to the manual method. Conclusion: 256-MDCT could potentially be employed for clinical assessment of stent patency in stents >3.0 mm when analysed with cardio-dedicated sharp kernels, although clinical studies corroborating this claim should be performed. However, stents ≤3.0 mm reconstructed by soft kernels revealed insufficient in-stent lumen visualisation and should not be used in clinical practice. Further improvements in spatial and temporal image resolution as well as reductions of radiation exposure and image noise have to be accomplished for the ambitious goal of characterising both CT coronary artery anatomy and in-stent lumen. © 2010 Elsevier Ireland Ltd. All rights reserved.
1. Introduction
Abbreviations: AMI, acute myocardial infarction; BMS, bare metal stent; CAD, coronary artery disease; DES, drug eluting stent; ISR, in-stent restenosis; MDCT, multi-detector CT; PCI, percutaneous coronary intervention. ∗ Corresponding author. Tel.: +49 6221 56 38686; fax: +49 6221 56 5513. E-mail addresses: [email protected] (H. Steen), [email protected] (F. André), [email protected] (G. Korosoglou), [email protected] (D. Mueller), [email protected] (W. Hosch), [email protected] (H.-U. Kauczor), [email protected] (E. Giannitsis), [email protected] (H.A. Katus). 1 Te.: +49 6221 560. 0720-048X/$ – see front matter © 2010 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.ejrad.2010.08.007
Percutaneous coronary intervention (PCI) is the treatment of choice in (a) patients with acute coronary syndrome (ACS), (b) in symptomatic patients with chronic coronary artery disease (CAD) and symptoms of concomitant myocardial ischemia as well as in (c) acute myocardial infarction (AMI) [1,2]. Stent placement during PCI has become the reference procedure for clinical treatment of these patients rendering former balloon angioplasty a niche procedure. Despite the widespread use of drug eluting stents (DES), in-stent restenosis (ISR) has remained the major limitation of PCI [3]. Since ISR, aside from causing angina, is also associated with significant morbidity and mortality [3], precise diagnosis of potential ISR is of high clinical relevance. Approximately 600,000 stent procedures are annually performed in the US [1] leading
144
H. Steen et al. / European Journal of Radiology 80 (2011) 143–150
to ISR in about 20–30% of bare metal stents (BMS) and ∼10% of drug eluting stents (DES) [4]. Therefore, a large number of patients have to be examined by conventional coronary angiography for exclusion of ISR, contributing considerably to health care cost and exposing the patient to the risk of an invasive procedure [5]. Therefore, a non-invasive assessment of ISR would be of great clinical and considerable socio-economic importance [6]. Recently, novel 64-slice multi-detector single- and dual-source CT systems [7–10] (MDCT/DSCT) offered isotropic voxel resolution up to 0.4 mm, 83 ms temporal resolution and cardio-dedicated sharp kernel reconstruction methods [11] which led to considerably improved diagnostic accuracy for the detection of significant CAD [12]. In previously stented patients, there was encouraging evidence from 64-MDCT not only with respect to detecting stent occlusion but also to visualise low attenuation filling defects which are noted as indirect signs of ISR [8,13]. The influence of stent strut design, stent material and lumen diameter on CT image quality is well known from former 16- to 64MDCT in vitro studies [9,10,14,15]. From these studies there was clear evidence that with increasing CT-detector rows stent lumen visibility would gradually improve. Hence, this is the first in vitro study with 256-MDCT to investigate lumen visibility of various recent commercially available and experimental coronary artery stents of different types and sizes with state-of-the-art 256-MDCT reconstruction algorithms.
2. Materials and methods Fifty-six coronary artery stents of different manufacturers, materials and strut designs were studied and are summarized in Table 1. Stents were expanded at the nominal pressure of their respective stent delivery systems in a coronary vessel phantom which was made of a plastic tube with a wall thickness of 0.5 mm. Stents with a diameter up to 3 mm were implanted into tubes with an inner diameter of 3 mm. Larger stents were positioned in tubes with an inner diameter of 4 mm. As previously published [9], after stent deployment the tube lumen was filled with contrast media (Ultravist 370, Schering AG, Berlin, Germany) and diluted to a radio-density of approximately 250 HU (Hounsfield units). Tubes were sealed at both ends and positioned in a plastic container filled with vegetable oil with an adjusted radio-density of approximately −70 HU by adding iodine, simulating epicardial fat tissue. Stents were placed parallel to the z-axis of the scanner. Imaging was performed in a 256-slice CT-scanner (Brilliance iCT, Philips Healthcare, Cleveland, OH, USA) with the following parameters: helical scan mode, collimation = 128 mm × 0.625 mm, pitch = 0.18, tube rotation time = 270 ms, effective tube current = 800 mA s, tube voltage = 120 kV, field-of-view = 180 mm, matrix = 512 × 512, slice thickness = 0.67 mm, artificial ECG-gating (60/min) with retrospective reconstruction at 75%. Images were reconstructed with four different convolution kernels: (a) Xres detailed stent (XCD), (b) cardiac sharp (CC), (c) cardiac detailed stent (CD) and (d) Xres standard (XCB) on a dedicated CT workstation (Extended Brilliance Workspace V 3.5.3.1020, Philips Healthcare, Cleveland, OH, USA) with a window setting of 300/1200 HU as previously described [9]. While XCB is usually employed for assessment of coronary artery disease, XCD and CD are dedicated stent kernels. Stent lumen diameters were evaluated with two methods: (a) by subjectively drawing the estimated lumen diameter with an electronic caliper (manual method) and (b) semi-quantitatively by application of the full-width-at-half-maximum method (FWHM-
method) which has been validated previously [9]. In either method lumen diameters were assessed in axial orientation at proximal, middle and distal parts of the stent and mean values were employed for statistical analysis. To test the influence of different stent diameters on CT lumen visibility, all coronary stents were assorted into three groups of incremental diameter size: group A = 2.25–2.5 mm, group B = 2.75–3.0 mm and group C = 3.5–4.0 mm. 2.1. Lumen attenuation Signal densities inside a deployed stent lumen are most often increased or rarely decreased due to beam hardening and partial volume effects, both described as attenuation [9,10]. To quantify this effect, two types of attenuation were defined: (a) tube lumen attenuation (TLA) inside the tube and (b) stent lumen attenuation (SLA) inside the deployed stent, both filled with contrast media (∼250 HU). Values for TLA and SLA were measured in a coronary view by a region of interest (ROI) technique. For TLA, two ROIs were positioned outside the stent inside the tube lumen and a weighted mean value was calculated. For SLA, a third ROI was placed inside the stent lumen omitting strut artifacts. 2.2. Contrast- and signal-to-noise ratio (CNR and SNR) Contrast-to-noise ratio (CNR) was defined as the difference between the values of SLA minus TLA, divided by the image noise. Image noise was defined as the mean value of the standard deviations of three axial ROIs in the surrounding oily fluid outside the tube lumen. CNR = (SLA − TLA)/Stdnoise . Furthermore, coronary stents were subdivided into three types according to previously suggested thresholds [10] for the assessment of stent lumen visibility: type I = good visualisation with measured stent diameter > 60% of nominal diameter, type II = moderate visualisation between 50% and 60% and type III = insufficient visualisation with less than 50%. 2.3. Statistics Data were expressed as mean ± 1 standard deviation. Differences between any two groups were compared by Student’s t-test for continuous variables. Continuous variables among more than two groups were compared by one-way ANOVA with post hoc analysis with Bonferroni adjustment for multiple comparisons. For all analyses, p < 0.05 was regarded statistically significant. All statistical analyses were carried out using MedCalc 9.4.1.0 (MedCalc Statistical Software bvba, Belgium). 3. Results All 56 coronary stents were reconstructed using the aforementioned four convolution kernels. Fig. 1 shows longitudinal through-plane 0.67 mm thick reformations of all stents using the XCD kernel. 3.1. Percent lumen diameter of the FWHM and the manual measurements Mean lumen diameters for the FWHM and manual method for the four different kernels are given in Fig. 2. In general, the FWHM method showed significantly smaller percentages of lumen diameter for three of the four kernels (XCD, CC, XCB, p < 0.05), when
H. Steen et al. / European Journal of Radiology 80 (2011) 143–150
145
Table 1 List of different manufacturers, stent-types, dimensions, materials and drugs of 56 coronary artery stents. No.
Manufacturer
Name
Diameter (mm)
Length (mm)
Strut thickness (mm)
Material
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56
Abbott Vascular Abbott Vascular Abbott Vascular Abbott Vascular Abbott Vascular B. Braun B. Braun Biotronik Biotronik Biotronik Biotronik Biotronik BostonScientific BostonScientific BostonScientific BostonScientific Cordis Cordis Cordis Cordis Cordis Medtronic Medtronic Medtronic Medtronic Medtronic Medtronic Medtronic Medtronic Medtronic Medtronic Medtronic Medtronic Medtronic Medtronic Medtronic Medtronic Medtronic Medtronic Medtronic Medtronic Medtronic Medtronic Medtronic Medtronic Micro Science Medical Micro Science Medical Micro Science Medical Terumo Terumo Terumo Terumo Terumo Translumina Translumina Translumina
Multi-Link Mini Vision Multi-Link Mini Vision Multi-Link Vision Multi-Link Vision Multi-Link Vision Coroflex Blue Coroflex Please PRO-Kinetic PRO-Kinetic PRO-Kinetic PRO-Kinetic PRO-Kinetic Liberté Liberté Taxus Liberté Taxus Liberté Cyper Select Plus Cyper Select Plus Cyper Select Plus Presillion Presillion Micro-Driver Micro-Driver Micro-Driver Micro-Driver Driver Driver Driver Driver Driver Driver Endeavor Endeavor Resolute Endeavor Resolute Endeavor Resolute Endeavor Resolute Endeavor Resolute Endeavor Resolute Endeavor Resolute Endeavor Resolute Endeavor Resolute Endeavor Sprint Endeavor Sprint Endeavor Sprint Endeavor Sprint Experimental 1a Experimental 2a Experimental 3a Tsunami Gold Tsunami Gold Tsunami Gold Tsunami Gold Tsunami Gold SV Yukona Yukon Choicea Yukon Choice CCa
2.50 2.50 2.75 3.50 3.50 3.00 3.00 2.50 3.00 3.50 4.00 4.00 3.00 3.00 3.00 3.00 2.50 3.50 3.50 3.00 3.00 2.25 2.50 2.75 2.75 3.00 3.50 4.00 4.00 4.00 4.00 3.50 2.25 2.25 2.50 2.50 2.75 2.75 3.00 3.50 3.50 2.75 2.75 3.50 3.50 3.00 3.00 3.00 3.00 3.00 3.50 3.50 2.50 3.00 3.00 3.00
18 28 8 12 28 19 19 22 30 20 13 20 16 20 16 20 18 13 18 12 17 24 24 18 24 12 9 9 12 24 30 9 8 12 12 24 8 30 12 9 24 14 24 18 24 16 16 16 18 30 10 18 20 18 18 18
0.081 0.081 0.081 0.081 0.081 0.065 0.120 0.060 0.060 0.080 0.080 0.080 0.097 0.097 0.097 0.097 0.140 0.140 0.140 0.073 0.073 0.091 0.091 0.091 0.091 0.091 0.091 0.091 0.091 0.091 0.091 0.091 0.091 0.091 0.091 0.091 0.091 0.091 0.091 0.091 0.091 0.091 0.091 0.091 0.091 0.080 0.080 0.080 0.080 0.080 0.080 0.080 0.080 0.115 0.097 0.100
Cocr Cocr CoCr CoCr CoCr CoCr stainless steel 316L CoCr + Si:C coating CoCr + Si:C coating CoCr + Si:C coating CoCr + Si:C coating CoCr + Si:C coating stainless steel 316L stainless steel 316L stainless steel 316L stainless steel 316L stainless steel 316L stainless steel 316L stainless steel 316L CoCr CoCr Cobalt alloy Cobalt alloy Cobalt alloy Cobalt alloy Cobalt alloy Cobalt alloy Cobalt alloy Cobalt alloy Cobalt alloy Cobalt alloy CoCr CoCr CoCr CoCr CoCr CoCr CoCr CoCr CoCr CoCr CoCr CoCr CoCr CoCr stainless steel 316L + Tantal coating stainless steel 316L + Tantal coating stainless steel 316L + Tantal coating stainless steel 316L stainless steel 316L stainless steel 316L stainless steel 316L stainless steel 316L stainless steel 316L stainless steel 316L CoCr
a
Drug
Paclitaxel
Paclitaxel Paclitaxel Sirolimus Sirolimus Sirolimus
Zotarolimus Zotarolimus Zotarolimus Zotarolimus Zotarolimus Zotarolimus Zotarolimus Zotarolimus Zotarolimus Zotarolimus Zotarolimus Zotarolimus Zotarolimus Zotarolimus
Stent without a deliviering system which were expanded by a ballon catheter.
compared to the manual lumen approach. Only the CD kernel showed similar stent lumen diameter results. Although the manual approach already revealed a good coefficient of variation (5.3%), the even more reproducible index was given by the FWHM-method (1.8%) with an intra-observer concordance of 0.95. Examples in coronal (inserted frames) and longitudinal views of one stent (Endeavour Resolute) with a diameter of 3 mm reconstructed with four different kernels are given in Fig. 3. Due to the better reproducibility of the FWHM-method, we focused on this method in the following.
3.2. Interrelation between stent diameter and lumen visibility Comparison of stent groups A and B employing the four reconstruction kernels showed no significant difference in the percentages of stent lumen diameters (Fig. 4a, p = n.s.). However, in all four kernels there was a significant difference in the percentages of stent lumen diameters when groups A or B were compared to group C (p < 0.01). The highest percentage of lumen diameter for group A and B stents were measured for the CD kernel, for group C stents with the CD and XCD kernels (all p < 0.05).
146
H. Steen et al. / European Journal of Radiology 80 (2011) 143–150
Fig. 1. Comparison of 56 different coronary artery stents of different manufacturers, materials and strut designs. Using a sharp XCD kernel, longitudinal through-plane reformations of 0.67-mm slice thickness reconstruction are shown.
Fig. 2. Comparison of mean lumen diameters assessed with both the FWHM method and the manual approach for the four different kernels. In general, the FWHM method showed significantly smaller percentages of lumen diameter for three of the four kernels (XCD, CC, XCB, p < 0.05) when compared to the manual lumen approach. Only the CD kernel showed similar stent lumen diameters, which would enable the comparison of stent data from other studies. The largest diameters were measured for the sharp kernels XCD and CD.
3.3. Artificial lumen narrowing The significantly lower percentage of lumen diameter of the FWHM-method held true for the XCD, CC, XCB kernel (p < 0.05), whereas the CD kernel showed comparable percentages of lumen diameter (Fig. 2, p = n.s.). With the CD and XCD kernels, the largest mean percentage of lumen diameter could be measured. An overview of the mean
diameters of the different groups measured by the FWHM-method is given in Table 2. Furthermore, a list of the artificial lumen narrowing values for all stents measured by the FWHM-method as well as by the manual approach can be found in Appendix A. With the XCD kernel reconstruction, for group A the visible percentage of lumen diameter ranged from 40% (PRO-Kinetic) to 54% (Endeavor Resolute), for group B from 44% (Tsunami Gold) to 54%
H. Steen et al. / European Journal of Radiology 80 (2011) 143–150
147
Table 2 Mean lumen diameter of different stent groups and kernels. XCD Group A Group B Group C
46 (40/54) ±4 48 (44/54) ±3 62 (58/66) ±3
CC 42 (36/46) ±3 41 (38/45) ±2 53 (48/56) ±2
CD 50 (42/57) ±4 52 (47/57) ±3 63 (59/68) ±3
XCB 46 (37/53) ±4 46 (42/51) ±3 58 (52/62) ±3
The diameters are given as mean (minimum/maximum) ± 1 standard deviation in % of the nominal lumen as measured by the FWHM-method.
Fig. 3. Exemplary images taken from the Endeavour Resolute stent (3 mm diameter) for all four reconstruction kernels (XCD, CC, CD, XCB) in the axial (parallel to z-axis) and in small windows cross-sectional (orthogonal to the z-axis) views. Even visually, one can already see the superior in-stent lumen visualisation for the XCD and CD reconstructions.
(Micro-Driver), for group C from 58% (Driver) to 66% (Endeavor Resolute). Mean percentages of lumen diameter for groups A, B and C were 46 ± 4%, 48 ± 3% and 62 ± 3%. Using the CC kernel, for group A the visible percentage of lumen diameter ranged from 36% (PRO-Kinetic) to 46% (Endeavor Resolute), for group B from 38% (Tsunami Gold) to 45% (Micro-Driver), for group C from 48% (Driver) to 56% (Endeavor Resolute). Mean percentages of lumen diameter for groups A, B and C were 42 ± 3%, 41 ± 2% and 53 ± 2%. For the CD kernel, for group A the visible percentage of lumen diameter ranged from 42% (PRO-Kinetic) to 57% (Endeavor Reso-
lute), for group B from 47% (PRO-Kinetic) to 57% (Micro-Driver), for group C from 59% (Driver) to 68% (Endeavor Resolute). Mean percentages of lumen diameter for groups A, B and C were 50 ± 4%, 52 ± 3% and 63 ± 3%. For the XCB kernel, for group A the visible percentage of lumen diameter ranged from 37% (PRO-Kinetic) to 53% (Endeavor Resolute), for group B from 42% (Tsunami Gold) to 51% (Endeavor Sprint), for group C from 52% (Tsunami Gold) to 62% (Endeavor Resolute). Mean percentages of lumen diameter for groups A, B and C were 46 ± 4%, 46 ± 3% and 58 ± 3%. In our study, the Endeavor Resolute, Endeavor Sprint and the Micro-Driver stents showed best results both for small as well as larger coronary stent diameters (up to 68%). The least visible stents were the PRO-Kinetic, Tsunami Gold and the Driver stents (as low as 36%). Stents with type I in-stent lumen visibility were different for the three groups and the different reconstruction algorithms (Fig. 4b). For the XCD kernel, there was only one type II stent (Endeavor Resolute) in group A, whereas 10 stents showed insufficient type III artificial lumen narrowing. In group B, out of 23 stents only 6 showed type II lumen narrowing whereupon 17 were insufficient (type III). From 22 group C stents 16 showed type I and 6 type II lumen narrowing. For the CC kernel, all stents from groups A and B showed insufficient type III artificial lumen narrowing. From 22 group C stents 16 stents showed only type II and 6 insufficient type III lumen narrowing. For the CD kernel, at least 6 type II stents in group A could be assessed, whereas 5 stents showed type III artificial lumen narrowing. In group B, out of 23 only 4 stents showed type III lumen narrowing whereupon 17 were moderate (type II). From 22 group C stents 16 stents showed type I and 6 type II lumen narrowing. For the XCB kernel, there was again only one type II stent (Endeavor Resolute) in group A, whereas 10 stents showed insufficient type III artificial lumen narrowing. In group B, out of 23 only 2 stents showed type II lumen narrowing whereupon 21 were insufficient (type III). From 22 group C stents 6 stents showed type I and 16 type II lumen narrowing. 3.4. CNR of the depicted stent lumen
Fig. 4. (a) The interrelation between stent diameter and percentages of lumen visibility for the three stent groups and the four different kernels. Obviously, stent groups A and B for small and intermediate coronary artery branches exhibit significantly smaller (p < 0.01) in-stent lumen diameter when compared to the group C stents for larger coronary arteries. Again the preferable use of the XCD and CD reconstruction is reconfirmed due to significantly larger (p = 0.03) in-stent lumen measurements. (b) Groups 1 and 2 stents are sub-summarized and show inferior quality for the visualisation of the in-stent diameters. Group 3 stents exhibit promising results, especially for the sharp reconstruction kernels.
Since a similar attenuation between SLA and TLA would be most desired, the optimal CNR values in our study should be small or in the vicinity of zero. Mean CNR values for the three groups of coronary stents using the four reconstruction protocols are given in Fig. 5a. For all kernels, there was a clear and significant (p < 0.05) relation between stent lumen diameter and CNR. For all kernels, the decreases in CNR between the three stent groups were all significant (p < 0.05). An overview of the results is given in Table 3. When the XCD kernel was employed, CNR values for group A stents ranged from 1.4 (Endeavour Resolute) to 6.9 (Micro-Driver), for group B stents from 0.8 (Endeavour Sprint) to 6.4 (Endeavour Resolute) and group C stents from 0.8 (Tsunami Gold) to 5.2 (Multilink Vision).
148
H. Steen et al. / European Journal of Radiology 80 (2011) 143–150
both FMWH and manual measurements for one exemplary stent type with different sizes (i.e. Micro-Driver/Driver 2.25–4 mm). In general, for smaller stent diameters (2.25–2.5 mm) the FWHM method showed significantly larger diameters due to resolution and partial volume effects. Between 2.5 and 3.0 mm, both methods were almost equally good. For stents ≥3 mm CNR was negligible. For stents smaller than 2.75 mm attenuation effects become more prevalent and objective assessment of the in-stent lumen cannot be reliably conducted. 3.6. Image noise The mean noise values for all four kernels were significantly different (p = 0.049). As expected, the lowest noise values were measured for the smoother XCB (15.7 ± 1.8) and CC kernels (19.8 ± 2.2) when compared to the previously described sharper kernels XCD (21.0 ± 2.9) and CD (34.8 ± 4.7). 4. Discussion
Fig. 5. (a)Mean CNR values for the three groups of coronary stents using the four reconstruction protocols. For all kernels, larger stent diameters lead to the preferred lower CNR. Furthermore, there was a significant CNR decrease between all three stent groups (p < 0.05). Taking into account that CNR values near zero generate the least in-stent artifacts, the use of CD- and XCD-kernels are most desirable (p = n.s.) for all three stent groups. Compared to the CC and XCB reconstruction kernels, the CNR values were highly significant for all three groups (p < 0.001). (b) Exemplary (MicroDriver/Driver 2.25–4 mm) inverse linear dependence between stent diameter and CNR for both measurement methods. For stents ≥3 mm, CNR between stent lumen and tube lumen is very low to negligible and therefore a potential clinical evaluation of proximal ISR may hold great promise. For stents smaller than 2.75 mm, attenuation becomes more prevalent due to partial volume and beam hardening effects so that objective assessment of the in-stent lumen cannot be reliably conducted. Table 3 CNR value ranges for different stent groups and kernels.
Group A Group B Group C
XCD
CC
CD
XCB
1.4–7.9 0.8–6.4 0.8–5.2
3.5–20.8 −0.1–5.5 −8.3–3.7
1.2–5.9 0.3–4.8 −1.5–1.3
10.4–26.2 5.7–20.3 0.3–11.5
Using the CC kernel, CNR values for group A stents ranged from 3.5 (Endeavour Resolute) to 20.8 (PRO-Kinetic), for group B stents from −0.1 (PROKinetik) to 5.5 (Tsunami Gold) and group C stents from −8.3 (ML Vision) to 3.7 (MSM 13) The CD kernel showed CNR values for group A stents ranging from 1.2 (Endeavour Resolution) to 5.9 (Micro-Driver), for group B stents from 0.3 (Taxus Liberté) to 4.8 (Endeavour Resolution) and group C stents from −1.5 (Endeavour) to 1.3 (MSM 8). Lastly, the XCB kernel CNR values for group A stents ranged from 10.4 (Cypher Select) to 26.2 (PRO-Kinetic), for group B stents from 5.7 (Yukon Choice) to 20.3 (Endeavour Resolution) and group C stents from 0.3 (Driver) to 11.5 (Endeavour Resolution). Taking into account that CNR values near zero represent best instent lumen visualisation, the CD- and XCD-kernels are of similar quality and most recommendable (p = n.s.) in all three stent groups. Compared to the CC and XCB reconstruction kernels, the differences of CNR values were highly significant for all three groups (p < 0.001). 3.5. Dependence between CNR and stent diameter Fig. 5b shows an almost inverse linear dependence between stent diameter and CNR, when lumen diameters were assessed with
Recently, a novel 256-MDCT has been introduced to clinical practice with 8 cm coverage and further improved spatial and temporal resolution when compared to previous CT generations. With this CT technology we have recently shown that quantification of lumen narrowing and assessment of atherosclerotic plaque is feasible, using coronary angiography as a gold-standard technique [16]. However, with 256-MDCT little data is available so far for the visualisation of coronary stents. In our study, 56 recent coronary artery stents with various diameters, strut designs and materials were assessed in vitro regarding their lumen diameter visualisation and in-stent lumen signal to answer the question whether state-of-the-art 256-MDCT could precisely distinguish coronary stent components from the contrastfilled lumen. We sought to investigate (a) the performance of coronary artery stents on 256-MDCT to create a catalogue of stent performances for the most clinically used stents as well as some experimental ones (MicroScienceMedical) but also (b) the interrelation between nominal and measured stent diameters for the usefulness of CT angiography in the visualisation of already implanted stents. For luminal stent assessment we introduced a previously validated FWHM method [17] of high reproducibility for CT stent lumen measurements in contrast to recently published data [9,10] where a manual approach was utilized. Similarly to Maintz et al., in our study we confirmed the high variability of stent appearances and diameters that were reported for 16-to-64-MDCT scanners [18,19]. We also validated the superiority of sharp reconstruction kernels (XCD, CD) for in-stent lumen visualisation. But although we utilised sharp CT kernels, 256-MDCT was unable to visualise more than 68% of the true lumen with the FWHM method. On the other hand promising data could be provided for stents >3.0 mm where most stent lumen measurements showed at least 60% of the stent lumen. One major finding in our study was the quantitative relationship between true stent lumen diameter and its visualisation on 256-MDCT. When applying the threshold of ≥50% stent lumen visibility to sufficiently exclude significant (>50% lumen diameter) stenosis as suggested by Maintz et al. [9,10], none of group A stents and only a minority of group B stents fulfilled the criteria. Therefore, the conclusion has to be drawn that luminal assessment of stents ≤3.0 mm by CT is presently not adequate. This is in line with recently published findings of the CORE-64 multicenter trial assessment of ISR, where sensitivity to identify ISR was between 20% and 25% for stents between 2.5 and 3.0 mm and 67% for stents >3.0 mm [20]. Our study showed an improvement in reproducibility when
H. Steen et al. / European Journal of Radiology 80 (2011) 143–150
a quantitative method was utilized (FWHM) as opposed to a manual method, while the CORE-64 results showed no improvement from the manual assessment when they used their quantitative method. However, the threshold of 50% lumen visibility is derived from the definition of restenosis in clinical trials and is not related to clinical outcome. Moreover, novel post-processing data algorithms, improved reconstruction methods, faster gantry rotations and innovative image noise reduction kernels may lead to improved in-stent visualisation in the near future. So far, there was unanimous consensus about the ability of highperformance MDCT to detect stent occlusion [18] but scarce and controversial data exists on the ability of detection or even exclusion of ISR [19]. In our study there is encouraging evidence for larger stents (group C) that in-stent lumen visibility might be sufficient at least for sharp kernels (XCD, CD) since all investigated stents showed at least ≥50% in-stent lumen visibility, the majority even ≥60% lumen visibility. When taking into account that group C stents ≥3 mm are usually placed in proximal coronary arteries or bypass grafts assuring blood supply to large myocardial territories, 256MDCT could contribute to non-invasive ruling-out of relevant ISR. Although there are alternative methods to evaluate stent patency e.g. the measurement of the amount of contrast media distal to the stent, stent lumen visualisation should be aimed at. All other approaches might lead to false-negative results because of uncontrolled influences as e.g. retrograde vessel filling. In our study we described the difference of signal intensities between the stent lumen (SLA) and the tube lumen (TLA) as CNR, a phenomenon previously described as attenuation [9,10]. Lowest CNR values could be measured for sharp kernels, especially for group C stents. Again, stents ≤3.0 mm might suffer from attenuation artifacts inside the lumen due to beam hardening and partial volume effects which could conceal in-stent atherosclerotic processes. Future in vitro and subsequent in-vivo studies with defined ISR need to be conducted to verify the potential of high-performance MDCT to correctly identify and classify in-stent lesions. Also, the diagnostic impact of negative CNR values for CC and CD kernels in group C stents need to be further validated since their diagnostic value is yet unclear. All four reconstruction settings represent protocols for routine clinical coronary examinations. From our results, we recommend sharp reconstruction kernels in patients with coronary stents. While the relatively high content of high frequencies in the modulation-transfer function may explain the greater values for stent lumen diameter, the second effect of lower CNR values may be due to decreased exaggeration in the lower frequency values. In our study the Endeavor Resolute and Sprint (drug eluting) stent as well as the Micro-Driver (cobalt alloy) showed best results both for small as well as larger coronary stent diameters (up to 68%). The least visible stents were the Pro-Kinetic (CoCr + SiC coating), Tsunami Gold (stainless steel) and the Driver stents (cobalt alloy) (as low as 36%). We did not investigate other innovative materials like magnesium (i.e. Magic stent) which previously exhibited the least artifacts and consequently best lumen visibility compared with conventional stents [10]. At the moment clinical application of this novel stent is yet limited, cost-intensive and clinical benefit needs to be further investigated. However, there is promising evidence that magnesium-containing stents dramatically improve in-stent visualisation and therefore could be attractive for non-invasive stent lumen assessment.
5. Limitations For the used in vitro phantom some limitations have to be considered. Scans were only performed perpendicular to the stent axis
149
which is not the representative orientation for all 15 coronary segments but only for parts of the LAD. However, for 16-MDCT [21], stents at orientations of 0◦ , 45◦ and 90◦ toward the Z-axis were analysed and there was evidence that the parallel orientation (0◦ to the Z-axis) was advantageous over the other angulations. Therefore, stent-related artifacts in oblique or orthogonal Z-axis orientation can be assumed and have to be further investigated. Window settings were not chosen automatically but in regard to previously published studies [9,10] (window width = 1200; window center = 300), which was a compromise between acceptable contrast and low metal artifacts. The FWHM method is not influenced by these settings as the measured values for the radiodensity are independent from the used gray scale. Heart rate was static and evaluated only at 60 bpm. The 256MDCT scanner with its wider coverage and faster rotation time of 270 ms may be able to image accurately at higher heart rates than single-source 64-slice systems. This was not evaluated in this study. A recent phantom study performed with 64-slice MDCT at varying heart rates (60, 75 and 90 bpm) [22] found that both spatial and temporal resolution, and therefore the level of heart rate, had a negative effect on the diagnostic accuracy. 64-MDCT significantly overestimated the degree of stenosis, underestimated the stent lumen and increased in-stent attenuation. Although the scanner parameters are according to the ones used in clinical routine, the radiation dose was not measured. Further studies may evaluate the potential of advanced dose saving algorithms for stent visualisation. As the experimental set-up and the method to measure the lumen diameter differ from other phantom studies dealing with the visualisation of stents in coronary CT, the results are not fully comparable. 6. Conclusion With the new generation of 256-MDCT, assessment of stent patency in stents larger than 3 mm could potentially be possible for clinical routine practice when analysed with cardio-dedicated sharp kernels. However, clinical studies corroborating this claim should be performed. In stents ≤3.0 mm, we experienced insufficient in-stent lumen visualisation and therefore assessment of in-stent patho-morphology is impractical. Further improvements in spatial and temporal image resolution as well as reductions of radiation exposure and image noise have to be accomplished for the ambitious goal of characterising both CT coronary artery anatomy and in-stent lumen. Acknowledgements The authors would like to thank all stent companies for instructive help and Philips Medical Systems for technical support. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.ejrad.2010.08.007. References [1] Patel MR, Dehmer GJ, Hirshfeld JW, Smith PK, Spertus JA. American College of Cardiology Foundation Appropriateness Criteria Task Force. J Am Coll Cardiol 2009 Feb (10);53(6):530–53. [2] Lloyd-Jones D, Adams R, Carnethon M, et al. American Heart Association Statistics Committee and Stroke Statistics Subcommittee. Heart disease and stroke statistics–2009 update: a report from the American Heart Association Statistics Committee and Stroke Statistics Subcommittee. Circulation 2009;119(January (3)). [3] Lemesle G, Maluenda G, Collins SD, Waksman R. Drug-eluting stents: issues of late stent thrombosis. Cardiol Clin 2010;28(February (1)):97–105.
150
H. Steen et al. / European Journal of Radiology 80 (2011) 143–150
[4] Pocock SJ, Lansky AJ, Mehran R, et al. Angiographic surrogate end points in drug-eluting stent trials: a systematic evaluation based on individual patient data from 11 randomized, controlled trials. J Am Coll Cardiol 2008;51(January (1)):23–32. [5] Scanlon PJ, Faxon DP, Audet AM, et al. ACC/AHA guidelines for coronary angiography. A report of the American College of Cardiology/American Heart Association Task Force on practice guidelines (Committee on Coronary Angiography). Developed in collaboration with the Society for Cardiac Angiography and Interventions. J Am Coll Cardiol 1999;33(May (6)): 1756–824. [6] Mowatt G, Cummins E, Waugh N, et al. Systematic review of the clinical effectiveness and cost-effectiveness of 64-slice or higher computed tomography angiography as an alternative to invasive coronary angiography in the investigation of coronary artery disease. Health Technol Assess 2008;12(May (17)). [7] Rixe J, Achenbach S, Ropers D, et al. Assessment of coronary artery stent restenosis by 64-slice multidetector computed tomography. Eur Heart J 2006;27(November (21)):2567–72. [8] Van Mieghem CA, Cademartiri F, Mollet NR, et al. Multislice spiral computed tomography for the evaluation of stent patency after left main coronary artery stenting: a comparison with conventional coronary angiography and intravascular ultrasound. Circulation 2006;114(August (7)):645–53. [9] Maintz D, Seifarth H, Raupach R, et al. 64-slice multidetector coronary CT angiography: in vitro evaluation of 68 different stents. Eur Radiol 2006;16(April (4)):818–26. [10] Maintz D, Burg MC, Seifarth H, et al. Update on multidetector coronary CT angiography of coronary stents: in vitro evaluation of 29 different stent types with dual-source CT. Eur Radiol 2009;19(January (1)):42–9. [11] Johnson TR, Nikolaou K, Wintersperger BJ, et al. Dual-source CT cardiac imaging: initial experience. Eur Radiol 2006;16(July (7)):1409–15. [12] Gaspar T, Halon DA, Lewis BS, et al. Diagnosis of coronary in-Stent restenosis with multidetector row spiral computed tomography. J Am Coll Cardiol 2005;46(October (8)):1573–9.
[13] Hecht HS, Zaric M, Jelnin V, Lubarsky L, Prakash M, Roubin G. Usefulness of 64-detector computed tomographic angiography for diagnosing in-stent restenosis in native coronary arteries. Am J Cardiol 2008;101(March (6)):820–4. [14] Suzuki S, Furui S, Kuwahara S, et al. Assessment of coronary stents using 40detector row CT: comparison with 16-detector row CT in vitro. J Comput Assist Tomogr 2007;31(November–December (6)):910–6. [15] Suzuki S, Furui S, Kuwahara S, et al. Detection of in-stent restenosis of coronary stents using 40-detector row computed tomography in vitro. J Comput Assist Tomogr 2008;32(March April (2)):252–8. [16] Korosoglou G, Mueller D, Lehrke S, et al. Quantitative assessment of stenosis severity and atherosclerotic plaque composition using 256-slice computed tomography. Eur Radiol 2010;(March (20)) [Epub ahead of print]. [17] Flohr T, Stierstorfer K, Raupach R, Ulzheimer S, Bruder H. Performance evaluation of a 64-slice CT system with z-flying focal spot. Rofo 2004;176(December (12)):1803–10. [18] Sun Z, Davidson R, Lin CH. Multi-detector row CT angiography in the assessment of coronary in-stent restenosis: a systematic review. Eur J Radiol 2009;69(March (3)):489–95. [19] Pugliese F, Weustink AC, Van Mieghem C, et al. Dual source coronary computed tomography angiography for detecting in-stent restenosis. Heart 2008;94(July (7)):848–54. [20] Wykrzykowska JJ, Arbab-Zadeh A, Godoy G, et al. Assessment of in-stent restenosis using 64-MDCT: analysis of the CORE-64 Multicenter International Trial. AJR 2010;194:85–92. [21] Maintz D, Seifarth H, Flohr T, et al. Improved coronary artery stent visualization and in-stent stenosis detection using 16-slice computed-tomography and dedicated image reconstruction technique. Invest Radiol 2003;38(December (12)):790–5. [22] Kerl JM, Schoepf UJ, Vogl TJ, et al. In vitro evaluation of metallic coronary artery stents with 64-MDCT using an ECG-gated cardiac phantom: relationship between in-stent visualization, stent type, and heart rate. AJR Am J Roentgenol 2010;194(March (3)):W256–62.
Contents lists available at ScienceDirect
European Journal of Radiology journal homepage: www.elsevier.com/locate/ejrad
In vitro evaluation of 56 coronary artery stents by 256-slice multi-detector coronary CT Henning Steen a,∗,1 , Florian André a,1 , Grigorios Korosoglou a,1 , Dirk Mueller b , Waldemar Hosch c,1 , Hans-Ulrich Kauczor c,1 , Evangelos Giannitsis a,1 , Hugo A. Katus a,1 a
University of Heidelberg, Department of Cardiology, Im Neuenheimer Feld 410, Heidelberg 69120, Germany Philips GmbH Healthcare Division, Luebeckertordamm 5, Hamburg 20099, Germany c University of Heidelberg, Department of Diagnostic and Interventional Radiology, Im Neuenheimer Feld 410, Heidelberg 69120, Germany b
a r t i c l e
i n f o
Article history: Received 30 March 2010 Received in revised form 27 July 2010 Accepted 2 August 2010 Keywords: Multislice CT Coronary stents CT angiography Artifacts Coronary artery disease In-stent Restenosis
a b s t r a c t Objective: We sought to investigate stent lumen visibility of 56 coronary stents with the newest 256multi-slice-CT (256-MDCT) technology for different reconstruction algorithms in an in vitro model. Background: Early identification of in-stent restenosis (ISR) is important to avoid recurrent ischemia and prevent acute myocardial infarction (AMI). Since angiography has the disadvantage of high costs and its invasiveness, MDCT could be a convenient and safe non-invasive alternative for detection of ISR. Material and methods: Percentages of in-stent lumen diameter and in-stent signal attenuation (measured as contrast-to-noise ratio (CNR)) of 56 coronary stents (group A ≤2.5 mm; group B = 2.75–3.0 mm; group C = 3.5–4.0 mm) were evaluated in a coronary vessel in vitro phantom (iodine-filled plastic tubes) employing four different reconstruction algorithms (XCD, CC, CD, XCB) on a novel 256-MDCT (Philips-iCT, collimation = 128 mm × 0.625 mm; rotation time = 270 ms; tube current = 800 mA s with 120 kV). Analysis was conducted with the semi-automatical full-width-at-half-maximum (FWHM) method. P-values <0.05 were regarded statistically significant. Results: In-stent lumen diameter >60% for group C stents was significantly larger and CNR was significantly lower (both p < 0.05) for sharp kernels (CD; XCD) when compared to groups A/B. The FWHM-method showed significantly smaller in-stent lumen diameter (p < 0.05) when compared to the manual method. Conclusion: 256-MDCT could potentially be employed for clinical assessment of stent patency in stents >3.0 mm when analysed with cardio-dedicated sharp kernels, although clinical studies corroborating this claim should be performed. However, stents ≤3.0 mm reconstructed by soft kernels revealed insufficient in-stent lumen visualisation and should not be used in clinical practice. Further improvements in spatial and temporal image resolution as well as reductions of radiation exposure and image noise have to be accomplished for the ambitious goal of characterising both CT coronary artery anatomy and in-stent lumen. © 2010 Elsevier Ireland Ltd. All rights reserved.
1. Introduction
Abbreviations: AMI, acute myocardial infarction; BMS, bare metal stent; CAD, coronary artery disease; DES, drug eluting stent; ISR, in-stent restenosis; MDCT, multi-detector CT; PCI, percutaneous coronary intervention. ∗ Corresponding author. Tel.: +49 6221 56 38686; fax: +49 6221 56 5513. E-mail addresses: [email protected] (H. Steen), [email protected] (F. André), [email protected] (G. Korosoglou), [email protected] (D. Mueller), [email protected] (W. Hosch), [email protected] (H.-U. Kauczor), [email protected] (E. Giannitsis), [email protected] (H.A. Katus). 1 Te.: +49 6221 560. 0720-048X/$ – see front matter © 2010 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.ejrad.2010.08.007
Percutaneous coronary intervention (PCI) is the treatment of choice in (a) patients with acute coronary syndrome (ACS), (b) in symptomatic patients with chronic coronary artery disease (CAD) and symptoms of concomitant myocardial ischemia as well as in (c) acute myocardial infarction (AMI) [1,2]. Stent placement during PCI has become the reference procedure for clinical treatment of these patients rendering former balloon angioplasty a niche procedure. Despite the widespread use of drug eluting stents (DES), in-stent restenosis (ISR) has remained the major limitation of PCI [3]. Since ISR, aside from causing angina, is also associated with significant morbidity and mortality [3], precise diagnosis of potential ISR is of high clinical relevance. Approximately 600,000 stent procedures are annually performed in the US [1] leading
144
H. Steen et al. / European Journal of Radiology 80 (2011) 143–150
to ISR in about 20–30% of bare metal stents (BMS) and ∼10% of drug eluting stents (DES) [4]. Therefore, a large number of patients have to be examined by conventional coronary angiography for exclusion of ISR, contributing considerably to health care cost and exposing the patient to the risk of an invasive procedure [5]. Therefore, a non-invasive assessment of ISR would be of great clinical and considerable socio-economic importance [6]. Recently, novel 64-slice multi-detector single- and dual-source CT systems [7–10] (MDCT/DSCT) offered isotropic voxel resolution up to 0.4 mm, 83 ms temporal resolution and cardio-dedicated sharp kernel reconstruction methods [11] which led to considerably improved diagnostic accuracy for the detection of significant CAD [12]. In previously stented patients, there was encouraging evidence from 64-MDCT not only with respect to detecting stent occlusion but also to visualise low attenuation filling defects which are noted as indirect signs of ISR [8,13]. The influence of stent strut design, stent material and lumen diameter on CT image quality is well known from former 16- to 64MDCT in vitro studies [9,10,14,15]. From these studies there was clear evidence that with increasing CT-detector rows stent lumen visibility would gradually improve. Hence, this is the first in vitro study with 256-MDCT to investigate lumen visibility of various recent commercially available and experimental coronary artery stents of different types and sizes with state-of-the-art 256-MDCT reconstruction algorithms.
2. Materials and methods Fifty-six coronary artery stents of different manufacturers, materials and strut designs were studied and are summarized in Table 1. Stents were expanded at the nominal pressure of their respective stent delivery systems in a coronary vessel phantom which was made of a plastic tube with a wall thickness of 0.5 mm. Stents with a diameter up to 3 mm were implanted into tubes with an inner diameter of 3 mm. Larger stents were positioned in tubes with an inner diameter of 4 mm. As previously published [9], after stent deployment the tube lumen was filled with contrast media (Ultravist 370, Schering AG, Berlin, Germany) and diluted to a radio-density of approximately 250 HU (Hounsfield units). Tubes were sealed at both ends and positioned in a plastic container filled with vegetable oil with an adjusted radio-density of approximately −70 HU by adding iodine, simulating epicardial fat tissue. Stents were placed parallel to the z-axis of the scanner. Imaging was performed in a 256-slice CT-scanner (Brilliance iCT, Philips Healthcare, Cleveland, OH, USA) with the following parameters: helical scan mode, collimation = 128 mm × 0.625 mm, pitch = 0.18, tube rotation time = 270 ms, effective tube current = 800 mA s, tube voltage = 120 kV, field-of-view = 180 mm, matrix = 512 × 512, slice thickness = 0.67 mm, artificial ECG-gating (60/min) with retrospective reconstruction at 75%. Images were reconstructed with four different convolution kernels: (a) Xres detailed stent (XCD), (b) cardiac sharp (CC), (c) cardiac detailed stent (CD) and (d) Xres standard (XCB) on a dedicated CT workstation (Extended Brilliance Workspace V 3.5.3.1020, Philips Healthcare, Cleveland, OH, USA) with a window setting of 300/1200 HU as previously described [9]. While XCB is usually employed for assessment of coronary artery disease, XCD and CD are dedicated stent kernels. Stent lumen diameters were evaluated with two methods: (a) by subjectively drawing the estimated lumen diameter with an electronic caliper (manual method) and (b) semi-quantitatively by application of the full-width-at-half-maximum method (FWHM-
method) which has been validated previously [9]. In either method lumen diameters were assessed in axial orientation at proximal, middle and distal parts of the stent and mean values were employed for statistical analysis. To test the influence of different stent diameters on CT lumen visibility, all coronary stents were assorted into three groups of incremental diameter size: group A = 2.25–2.5 mm, group B = 2.75–3.0 mm and group C = 3.5–4.0 mm. 2.1. Lumen attenuation Signal densities inside a deployed stent lumen are most often increased or rarely decreased due to beam hardening and partial volume effects, both described as attenuation [9,10]. To quantify this effect, two types of attenuation were defined: (a) tube lumen attenuation (TLA) inside the tube and (b) stent lumen attenuation (SLA) inside the deployed stent, both filled with contrast media (∼250 HU). Values for TLA and SLA were measured in a coronary view by a region of interest (ROI) technique. For TLA, two ROIs were positioned outside the stent inside the tube lumen and a weighted mean value was calculated. For SLA, a third ROI was placed inside the stent lumen omitting strut artifacts. 2.2. Contrast- and signal-to-noise ratio (CNR and SNR) Contrast-to-noise ratio (CNR) was defined as the difference between the values of SLA minus TLA, divided by the image noise. Image noise was defined as the mean value of the standard deviations of three axial ROIs in the surrounding oily fluid outside the tube lumen. CNR = (SLA − TLA)/Stdnoise . Furthermore, coronary stents were subdivided into three types according to previously suggested thresholds [10] for the assessment of stent lumen visibility: type I = good visualisation with measured stent diameter > 60% of nominal diameter, type II = moderate visualisation between 50% and 60% and type III = insufficient visualisation with less than 50%. 2.3. Statistics Data were expressed as mean ± 1 standard deviation. Differences between any two groups were compared by Student’s t-test for continuous variables. Continuous variables among more than two groups were compared by one-way ANOVA with post hoc analysis with Bonferroni adjustment for multiple comparisons. For all analyses, p < 0.05 was regarded statistically significant. All statistical analyses were carried out using MedCalc 9.4.1.0 (MedCalc Statistical Software bvba, Belgium). 3. Results All 56 coronary stents were reconstructed using the aforementioned four convolution kernels. Fig. 1 shows longitudinal through-plane 0.67 mm thick reformations of all stents using the XCD kernel. 3.1. Percent lumen diameter of the FWHM and the manual measurements Mean lumen diameters for the FWHM and manual method for the four different kernels are given in Fig. 2. In general, the FWHM method showed significantly smaller percentages of lumen diameter for three of the four kernels (XCD, CC, XCB, p < 0.05), when
H. Steen et al. / European Journal of Radiology 80 (2011) 143–150
145
Table 1 List of different manufacturers, stent-types, dimensions, materials and drugs of 56 coronary artery stents. No.
Manufacturer
Name
Diameter (mm)
Length (mm)
Strut thickness (mm)
Material
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56
Abbott Vascular Abbott Vascular Abbott Vascular Abbott Vascular Abbott Vascular B. Braun B. Braun Biotronik Biotronik Biotronik Biotronik Biotronik BostonScientific BostonScientific BostonScientific BostonScientific Cordis Cordis Cordis Cordis Cordis Medtronic Medtronic Medtronic Medtronic Medtronic Medtronic Medtronic Medtronic Medtronic Medtronic Medtronic Medtronic Medtronic Medtronic Medtronic Medtronic Medtronic Medtronic Medtronic Medtronic Medtronic Medtronic Medtronic Medtronic Micro Science Medical Micro Science Medical Micro Science Medical Terumo Terumo Terumo Terumo Terumo Translumina Translumina Translumina
Multi-Link Mini Vision Multi-Link Mini Vision Multi-Link Vision Multi-Link Vision Multi-Link Vision Coroflex Blue Coroflex Please PRO-Kinetic PRO-Kinetic PRO-Kinetic PRO-Kinetic PRO-Kinetic Liberté Liberté Taxus Liberté Taxus Liberté Cyper Select Plus Cyper Select Plus Cyper Select Plus Presillion Presillion Micro-Driver Micro-Driver Micro-Driver Micro-Driver Driver Driver Driver Driver Driver Driver Endeavor Endeavor Resolute Endeavor Resolute Endeavor Resolute Endeavor Resolute Endeavor Resolute Endeavor Resolute Endeavor Resolute Endeavor Resolute Endeavor Resolute Endeavor Sprint Endeavor Sprint Endeavor Sprint Endeavor Sprint Experimental 1a Experimental 2a Experimental 3a Tsunami Gold Tsunami Gold Tsunami Gold Tsunami Gold Tsunami Gold SV Yukona Yukon Choicea Yukon Choice CCa
2.50 2.50 2.75 3.50 3.50 3.00 3.00 2.50 3.00 3.50 4.00 4.00 3.00 3.00 3.00 3.00 2.50 3.50 3.50 3.00 3.00 2.25 2.50 2.75 2.75 3.00 3.50 4.00 4.00 4.00 4.00 3.50 2.25 2.25 2.50 2.50 2.75 2.75 3.00 3.50 3.50 2.75 2.75 3.50 3.50 3.00 3.00 3.00 3.00 3.00 3.50 3.50 2.50 3.00 3.00 3.00
18 28 8 12 28 19 19 22 30 20 13 20 16 20 16 20 18 13 18 12 17 24 24 18 24 12 9 9 12 24 30 9 8 12 12 24 8 30 12 9 24 14 24 18 24 16 16 16 18 30 10 18 20 18 18 18
0.081 0.081 0.081 0.081 0.081 0.065 0.120 0.060 0.060 0.080 0.080 0.080 0.097 0.097 0.097 0.097 0.140 0.140 0.140 0.073 0.073 0.091 0.091 0.091 0.091 0.091 0.091 0.091 0.091 0.091 0.091 0.091 0.091 0.091 0.091 0.091 0.091 0.091 0.091 0.091 0.091 0.091 0.091 0.091 0.091 0.080 0.080 0.080 0.080 0.080 0.080 0.080 0.080 0.115 0.097 0.100
Cocr Cocr CoCr CoCr CoCr CoCr stainless steel 316L CoCr + Si:C coating CoCr + Si:C coating CoCr + Si:C coating CoCr + Si:C coating CoCr + Si:C coating stainless steel 316L stainless steel 316L stainless steel 316L stainless steel 316L stainless steel 316L stainless steel 316L stainless steel 316L CoCr CoCr Cobalt alloy Cobalt alloy Cobalt alloy Cobalt alloy Cobalt alloy Cobalt alloy Cobalt alloy Cobalt alloy Cobalt alloy Cobalt alloy CoCr CoCr CoCr CoCr CoCr CoCr CoCr CoCr CoCr CoCr CoCr CoCr CoCr CoCr stainless steel 316L + Tantal coating stainless steel 316L + Tantal coating stainless steel 316L + Tantal coating stainless steel 316L stainless steel 316L stainless steel 316L stainless steel 316L stainless steel 316L stainless steel 316L stainless steel 316L CoCr
a
Drug
Paclitaxel
Paclitaxel Paclitaxel Sirolimus Sirolimus Sirolimus
Zotarolimus Zotarolimus Zotarolimus Zotarolimus Zotarolimus Zotarolimus Zotarolimus Zotarolimus Zotarolimus Zotarolimus Zotarolimus Zotarolimus Zotarolimus Zotarolimus
Stent without a deliviering system which were expanded by a ballon catheter.
compared to the manual lumen approach. Only the CD kernel showed similar stent lumen diameter results. Although the manual approach already revealed a good coefficient of variation (5.3%), the even more reproducible index was given by the FWHM-method (1.8%) with an intra-observer concordance of 0.95. Examples in coronal (inserted frames) and longitudinal views of one stent (Endeavour Resolute) with a diameter of 3 mm reconstructed with four different kernels are given in Fig. 3. Due to the better reproducibility of the FWHM-method, we focused on this method in the following.
3.2. Interrelation between stent diameter and lumen visibility Comparison of stent groups A and B employing the four reconstruction kernels showed no significant difference in the percentages of stent lumen diameters (Fig. 4a, p = n.s.). However, in all four kernels there was a significant difference in the percentages of stent lumen diameters when groups A or B were compared to group C (p < 0.01). The highest percentage of lumen diameter for group A and B stents were measured for the CD kernel, for group C stents with the CD and XCD kernels (all p < 0.05).
146
H. Steen et al. / European Journal of Radiology 80 (2011) 143–150
Fig. 1. Comparison of 56 different coronary artery stents of different manufacturers, materials and strut designs. Using a sharp XCD kernel, longitudinal through-plane reformations of 0.67-mm slice thickness reconstruction are shown.
Fig. 2. Comparison of mean lumen diameters assessed with both the FWHM method and the manual approach for the four different kernels. In general, the FWHM method showed significantly smaller percentages of lumen diameter for three of the four kernels (XCD, CC, XCB, p < 0.05) when compared to the manual lumen approach. Only the CD kernel showed similar stent lumen diameters, which would enable the comparison of stent data from other studies. The largest diameters were measured for the sharp kernels XCD and CD.
3.3. Artificial lumen narrowing The significantly lower percentage of lumen diameter of the FWHM-method held true for the XCD, CC, XCB kernel (p < 0.05), whereas the CD kernel showed comparable percentages of lumen diameter (Fig. 2, p = n.s.). With the CD and XCD kernels, the largest mean percentage of lumen diameter could be measured. An overview of the mean
diameters of the different groups measured by the FWHM-method is given in Table 2. Furthermore, a list of the artificial lumen narrowing values for all stents measured by the FWHM-method as well as by the manual approach can be found in Appendix A. With the XCD kernel reconstruction, for group A the visible percentage of lumen diameter ranged from 40% (PRO-Kinetic) to 54% (Endeavor Resolute), for group B from 44% (Tsunami Gold) to 54%
H. Steen et al. / European Journal of Radiology 80 (2011) 143–150
147
Table 2 Mean lumen diameter of different stent groups and kernels. XCD Group A Group B Group C
46 (40/54) ±4 48 (44/54) ±3 62 (58/66) ±3
CC 42 (36/46) ±3 41 (38/45) ±2 53 (48/56) ±2
CD 50 (42/57) ±4 52 (47/57) ±3 63 (59/68) ±3
XCB 46 (37/53) ±4 46 (42/51) ±3 58 (52/62) ±3
The diameters are given as mean (minimum/maximum) ± 1 standard deviation in % of the nominal lumen as measured by the FWHM-method.
Fig. 3. Exemplary images taken from the Endeavour Resolute stent (3 mm diameter) for all four reconstruction kernels (XCD, CC, CD, XCB) in the axial (parallel to z-axis) and in small windows cross-sectional (orthogonal to the z-axis) views. Even visually, one can already see the superior in-stent lumen visualisation for the XCD and CD reconstructions.
(Micro-Driver), for group C from 58% (Driver) to 66% (Endeavor Resolute). Mean percentages of lumen diameter for groups A, B and C were 46 ± 4%, 48 ± 3% and 62 ± 3%. Using the CC kernel, for group A the visible percentage of lumen diameter ranged from 36% (PRO-Kinetic) to 46% (Endeavor Resolute), for group B from 38% (Tsunami Gold) to 45% (Micro-Driver), for group C from 48% (Driver) to 56% (Endeavor Resolute). Mean percentages of lumen diameter for groups A, B and C were 42 ± 3%, 41 ± 2% and 53 ± 2%. For the CD kernel, for group A the visible percentage of lumen diameter ranged from 42% (PRO-Kinetic) to 57% (Endeavor Reso-
lute), for group B from 47% (PRO-Kinetic) to 57% (Micro-Driver), for group C from 59% (Driver) to 68% (Endeavor Resolute). Mean percentages of lumen diameter for groups A, B and C were 50 ± 4%, 52 ± 3% and 63 ± 3%. For the XCB kernel, for group A the visible percentage of lumen diameter ranged from 37% (PRO-Kinetic) to 53% (Endeavor Resolute), for group B from 42% (Tsunami Gold) to 51% (Endeavor Sprint), for group C from 52% (Tsunami Gold) to 62% (Endeavor Resolute). Mean percentages of lumen diameter for groups A, B and C were 46 ± 4%, 46 ± 3% and 58 ± 3%. In our study, the Endeavor Resolute, Endeavor Sprint and the Micro-Driver stents showed best results both for small as well as larger coronary stent diameters (up to 68%). The least visible stents were the PRO-Kinetic, Tsunami Gold and the Driver stents (as low as 36%). Stents with type I in-stent lumen visibility were different for the three groups and the different reconstruction algorithms (Fig. 4b). For the XCD kernel, there was only one type II stent (Endeavor Resolute) in group A, whereas 10 stents showed insufficient type III artificial lumen narrowing. In group B, out of 23 stents only 6 showed type II lumen narrowing whereupon 17 were insufficient (type III). From 22 group C stents 16 showed type I and 6 type II lumen narrowing. For the CC kernel, all stents from groups A and B showed insufficient type III artificial lumen narrowing. From 22 group C stents 16 stents showed only type II and 6 insufficient type III lumen narrowing. For the CD kernel, at least 6 type II stents in group A could be assessed, whereas 5 stents showed type III artificial lumen narrowing. In group B, out of 23 only 4 stents showed type III lumen narrowing whereupon 17 were moderate (type II). From 22 group C stents 16 stents showed type I and 6 type II lumen narrowing. For the XCB kernel, there was again only one type II stent (Endeavor Resolute) in group A, whereas 10 stents showed insufficient type III artificial lumen narrowing. In group B, out of 23 only 2 stents showed type II lumen narrowing whereupon 21 were insufficient (type III). From 22 group C stents 6 stents showed type I and 16 type II lumen narrowing. 3.4. CNR of the depicted stent lumen
Fig. 4. (a) The interrelation between stent diameter and percentages of lumen visibility for the three stent groups and the four different kernels. Obviously, stent groups A and B for small and intermediate coronary artery branches exhibit significantly smaller (p < 0.01) in-stent lumen diameter when compared to the group C stents for larger coronary arteries. Again the preferable use of the XCD and CD reconstruction is reconfirmed due to significantly larger (p = 0.03) in-stent lumen measurements. (b) Groups 1 and 2 stents are sub-summarized and show inferior quality for the visualisation of the in-stent diameters. Group 3 stents exhibit promising results, especially for the sharp reconstruction kernels.
Since a similar attenuation between SLA and TLA would be most desired, the optimal CNR values in our study should be small or in the vicinity of zero. Mean CNR values for the three groups of coronary stents using the four reconstruction protocols are given in Fig. 5a. For all kernels, there was a clear and significant (p < 0.05) relation between stent lumen diameter and CNR. For all kernels, the decreases in CNR between the three stent groups were all significant (p < 0.05). An overview of the results is given in Table 3. When the XCD kernel was employed, CNR values for group A stents ranged from 1.4 (Endeavour Resolute) to 6.9 (Micro-Driver), for group B stents from 0.8 (Endeavour Sprint) to 6.4 (Endeavour Resolute) and group C stents from 0.8 (Tsunami Gold) to 5.2 (Multilink Vision).
148
H. Steen et al. / European Journal of Radiology 80 (2011) 143–150
both FMWH and manual measurements for one exemplary stent type with different sizes (i.e. Micro-Driver/Driver 2.25–4 mm). In general, for smaller stent diameters (2.25–2.5 mm) the FWHM method showed significantly larger diameters due to resolution and partial volume effects. Between 2.5 and 3.0 mm, both methods were almost equally good. For stents ≥3 mm CNR was negligible. For stents smaller than 2.75 mm attenuation effects become more prevalent and objective assessment of the in-stent lumen cannot be reliably conducted. 3.6. Image noise The mean noise values for all four kernels were significantly different (p = 0.049). As expected, the lowest noise values were measured for the smoother XCB (15.7 ± 1.8) and CC kernels (19.8 ± 2.2) when compared to the previously described sharper kernels XCD (21.0 ± 2.9) and CD (34.8 ± 4.7). 4. Discussion
Fig. 5. (a)Mean CNR values for the three groups of coronary stents using the four reconstruction protocols. For all kernels, larger stent diameters lead to the preferred lower CNR. Furthermore, there was a significant CNR decrease between all three stent groups (p < 0.05). Taking into account that CNR values near zero generate the least in-stent artifacts, the use of CD- and XCD-kernels are most desirable (p = n.s.) for all three stent groups. Compared to the CC and XCB reconstruction kernels, the CNR values were highly significant for all three groups (p < 0.001). (b) Exemplary (MicroDriver/Driver 2.25–4 mm) inverse linear dependence between stent diameter and CNR for both measurement methods. For stents ≥3 mm, CNR between stent lumen and tube lumen is very low to negligible and therefore a potential clinical evaluation of proximal ISR may hold great promise. For stents smaller than 2.75 mm, attenuation becomes more prevalent due to partial volume and beam hardening effects so that objective assessment of the in-stent lumen cannot be reliably conducted. Table 3 CNR value ranges for different stent groups and kernels.
Group A Group B Group C
XCD
CC
CD
XCB
1.4–7.9 0.8–6.4 0.8–5.2
3.5–20.8 −0.1–5.5 −8.3–3.7
1.2–5.9 0.3–4.8 −1.5–1.3
10.4–26.2 5.7–20.3 0.3–11.5
Using the CC kernel, CNR values for group A stents ranged from 3.5 (Endeavour Resolute) to 20.8 (PRO-Kinetic), for group B stents from −0.1 (PROKinetik) to 5.5 (Tsunami Gold) and group C stents from −8.3 (ML Vision) to 3.7 (MSM 13) The CD kernel showed CNR values for group A stents ranging from 1.2 (Endeavour Resolution) to 5.9 (Micro-Driver), for group B stents from 0.3 (Taxus Liberté) to 4.8 (Endeavour Resolution) and group C stents from −1.5 (Endeavour) to 1.3 (MSM 8). Lastly, the XCB kernel CNR values for group A stents ranged from 10.4 (Cypher Select) to 26.2 (PRO-Kinetic), for group B stents from 5.7 (Yukon Choice) to 20.3 (Endeavour Resolution) and group C stents from 0.3 (Driver) to 11.5 (Endeavour Resolution). Taking into account that CNR values near zero represent best instent lumen visualisation, the CD- and XCD-kernels are of similar quality and most recommendable (p = n.s.) in all three stent groups. Compared to the CC and XCB reconstruction kernels, the differences of CNR values were highly significant for all three groups (p < 0.001). 3.5. Dependence between CNR and stent diameter Fig. 5b shows an almost inverse linear dependence between stent diameter and CNR, when lumen diameters were assessed with
Recently, a novel 256-MDCT has been introduced to clinical practice with 8 cm coverage and further improved spatial and temporal resolution when compared to previous CT generations. With this CT technology we have recently shown that quantification of lumen narrowing and assessment of atherosclerotic plaque is feasible, using coronary angiography as a gold-standard technique [16]. However, with 256-MDCT little data is available so far for the visualisation of coronary stents. In our study, 56 recent coronary artery stents with various diameters, strut designs and materials were assessed in vitro regarding their lumen diameter visualisation and in-stent lumen signal to answer the question whether state-of-the-art 256-MDCT could precisely distinguish coronary stent components from the contrastfilled lumen. We sought to investigate (a) the performance of coronary artery stents on 256-MDCT to create a catalogue of stent performances for the most clinically used stents as well as some experimental ones (MicroScienceMedical) but also (b) the interrelation between nominal and measured stent diameters for the usefulness of CT angiography in the visualisation of already implanted stents. For luminal stent assessment we introduced a previously validated FWHM method [17] of high reproducibility for CT stent lumen measurements in contrast to recently published data [9,10] where a manual approach was utilized. Similarly to Maintz et al., in our study we confirmed the high variability of stent appearances and diameters that were reported for 16-to-64-MDCT scanners [18,19]. We also validated the superiority of sharp reconstruction kernels (XCD, CD) for in-stent lumen visualisation. But although we utilised sharp CT kernels, 256-MDCT was unable to visualise more than 68% of the true lumen with the FWHM method. On the other hand promising data could be provided for stents >3.0 mm where most stent lumen measurements showed at least 60% of the stent lumen. One major finding in our study was the quantitative relationship between true stent lumen diameter and its visualisation on 256-MDCT. When applying the threshold of ≥50% stent lumen visibility to sufficiently exclude significant (>50% lumen diameter) stenosis as suggested by Maintz et al. [9,10], none of group A stents and only a minority of group B stents fulfilled the criteria. Therefore, the conclusion has to be drawn that luminal assessment of stents ≤3.0 mm by CT is presently not adequate. This is in line with recently published findings of the CORE-64 multicenter trial assessment of ISR, where sensitivity to identify ISR was between 20% and 25% for stents between 2.5 and 3.0 mm and 67% for stents >3.0 mm [20]. Our study showed an improvement in reproducibility when
H. Steen et al. / European Journal of Radiology 80 (2011) 143–150
a quantitative method was utilized (FWHM) as opposed to a manual method, while the CORE-64 results showed no improvement from the manual assessment when they used their quantitative method. However, the threshold of 50% lumen visibility is derived from the definition of restenosis in clinical trials and is not related to clinical outcome. Moreover, novel post-processing data algorithms, improved reconstruction methods, faster gantry rotations and innovative image noise reduction kernels may lead to improved in-stent visualisation in the near future. So far, there was unanimous consensus about the ability of highperformance MDCT to detect stent occlusion [18] but scarce and controversial data exists on the ability of detection or even exclusion of ISR [19]. In our study there is encouraging evidence for larger stents (group C) that in-stent lumen visibility might be sufficient at least for sharp kernels (XCD, CD) since all investigated stents showed at least ≥50% in-stent lumen visibility, the majority even ≥60% lumen visibility. When taking into account that group C stents ≥3 mm are usually placed in proximal coronary arteries or bypass grafts assuring blood supply to large myocardial territories, 256MDCT could contribute to non-invasive ruling-out of relevant ISR. Although there are alternative methods to evaluate stent patency e.g. the measurement of the amount of contrast media distal to the stent, stent lumen visualisation should be aimed at. All other approaches might lead to false-negative results because of uncontrolled influences as e.g. retrograde vessel filling. In our study we described the difference of signal intensities between the stent lumen (SLA) and the tube lumen (TLA) as CNR, a phenomenon previously described as attenuation [9,10]. Lowest CNR values could be measured for sharp kernels, especially for group C stents. Again, stents ≤3.0 mm might suffer from attenuation artifacts inside the lumen due to beam hardening and partial volume effects which could conceal in-stent atherosclerotic processes. Future in vitro and subsequent in-vivo studies with defined ISR need to be conducted to verify the potential of high-performance MDCT to correctly identify and classify in-stent lesions. Also, the diagnostic impact of negative CNR values for CC and CD kernels in group C stents need to be further validated since their diagnostic value is yet unclear. All four reconstruction settings represent protocols for routine clinical coronary examinations. From our results, we recommend sharp reconstruction kernels in patients with coronary stents. While the relatively high content of high frequencies in the modulation-transfer function may explain the greater values for stent lumen diameter, the second effect of lower CNR values may be due to decreased exaggeration in the lower frequency values. In our study the Endeavor Resolute and Sprint (drug eluting) stent as well as the Micro-Driver (cobalt alloy) showed best results both for small as well as larger coronary stent diameters (up to 68%). The least visible stents were the Pro-Kinetic (CoCr + SiC coating), Tsunami Gold (stainless steel) and the Driver stents (cobalt alloy) (as low as 36%). We did not investigate other innovative materials like magnesium (i.e. Magic stent) which previously exhibited the least artifacts and consequently best lumen visibility compared with conventional stents [10]. At the moment clinical application of this novel stent is yet limited, cost-intensive and clinical benefit needs to be further investigated. However, there is promising evidence that magnesium-containing stents dramatically improve in-stent visualisation and therefore could be attractive for non-invasive stent lumen assessment.
5. Limitations For the used in vitro phantom some limitations have to be considered. Scans were only performed perpendicular to the stent axis
149
which is not the representative orientation for all 15 coronary segments but only for parts of the LAD. However, for 16-MDCT [21], stents at orientations of 0◦ , 45◦ and 90◦ toward the Z-axis were analysed and there was evidence that the parallel orientation (0◦ to the Z-axis) was advantageous over the other angulations. Therefore, stent-related artifacts in oblique or orthogonal Z-axis orientation can be assumed and have to be further investigated. Window settings were not chosen automatically but in regard to previously published studies [9,10] (window width = 1200; window center = 300), which was a compromise between acceptable contrast and low metal artifacts. The FWHM method is not influenced by these settings as the measured values for the radiodensity are independent from the used gray scale. Heart rate was static and evaluated only at 60 bpm. The 256MDCT scanner with its wider coverage and faster rotation time of 270 ms may be able to image accurately at higher heart rates than single-source 64-slice systems. This was not evaluated in this study. A recent phantom study performed with 64-slice MDCT at varying heart rates (60, 75 and 90 bpm) [22] found that both spatial and temporal resolution, and therefore the level of heart rate, had a negative effect on the diagnostic accuracy. 64-MDCT significantly overestimated the degree of stenosis, underestimated the stent lumen and increased in-stent attenuation. Although the scanner parameters are according to the ones used in clinical routine, the radiation dose was not measured. Further studies may evaluate the potential of advanced dose saving algorithms for stent visualisation. As the experimental set-up and the method to measure the lumen diameter differ from other phantom studies dealing with the visualisation of stents in coronary CT, the results are not fully comparable. 6. Conclusion With the new generation of 256-MDCT, assessment of stent patency in stents larger than 3 mm could potentially be possible for clinical routine practice when analysed with cardio-dedicated sharp kernels. However, clinical studies corroborating this claim should be performed. In stents ≤3.0 mm, we experienced insufficient in-stent lumen visualisation and therefore assessment of in-stent patho-morphology is impractical. Further improvements in spatial and temporal image resolution as well as reductions of radiation exposure and image noise have to be accomplished for the ambitious goal of characterising both CT coronary artery anatomy and in-stent lumen. Acknowledgements The authors would like to thank all stent companies for instructive help and Philips Medical Systems for technical support. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.ejrad.2010.08.007. References [1] Patel MR, Dehmer GJ, Hirshfeld JW, Smith PK, Spertus JA. American College of Cardiology Foundation Appropriateness Criteria Task Force. J Am Coll Cardiol 2009 Feb (10);53(6):530–53. [2] Lloyd-Jones D, Adams R, Carnethon M, et al. American Heart Association Statistics Committee and Stroke Statistics Subcommittee. Heart disease and stroke statistics–2009 update: a report from the American Heart Association Statistics Committee and Stroke Statistics Subcommittee. Circulation 2009;119(January (3)). [3] Lemesle G, Maluenda G, Collins SD, Waksman R. Drug-eluting stents: issues of late stent thrombosis. Cardiol Clin 2010;28(February (1)):97–105.
150
H. Steen et al. / European Journal of Radiology 80 (2011) 143–150
[4] Pocock SJ, Lansky AJ, Mehran R, et al. Angiographic surrogate end points in drug-eluting stent trials: a systematic evaluation based on individual patient data from 11 randomized, controlled trials. J Am Coll Cardiol 2008;51(January (1)):23–32. [5] Scanlon PJ, Faxon DP, Audet AM, et al. ACC/AHA guidelines for coronary angiography. A report of the American College of Cardiology/American Heart Association Task Force on practice guidelines (Committee on Coronary Angiography). Developed in collaboration with the Society for Cardiac Angiography and Interventions. J Am Coll Cardiol 1999;33(May (6)): 1756–824. [6] Mowatt G, Cummins E, Waugh N, et al. Systematic review of the clinical effectiveness and cost-effectiveness of 64-slice or higher computed tomography angiography as an alternative to invasive coronary angiography in the investigation of coronary artery disease. Health Technol Assess 2008;12(May (17)). [7] Rixe J, Achenbach S, Ropers D, et al. Assessment of coronary artery stent restenosis by 64-slice multidetector computed tomography. Eur Heart J 2006;27(November (21)):2567–72. [8] Van Mieghem CA, Cademartiri F, Mollet NR, et al. Multislice spiral computed tomography for the evaluation of stent patency after left main coronary artery stenting: a comparison with conventional coronary angiography and intravascular ultrasound. Circulation 2006;114(August (7)):645–53. [9] Maintz D, Seifarth H, Raupach R, et al. 64-slice multidetector coronary CT angiography: in vitro evaluation of 68 different stents. Eur Radiol 2006;16(April (4)):818–26. [10] Maintz D, Burg MC, Seifarth H, et al. Update on multidetector coronary CT angiography of coronary stents: in vitro evaluation of 29 different stent types with dual-source CT. Eur Radiol 2009;19(January (1)):42–9. [11] Johnson TR, Nikolaou K, Wintersperger BJ, et al. Dual-source CT cardiac imaging: initial experience. Eur Radiol 2006;16(July (7)):1409–15. [12] Gaspar T, Halon DA, Lewis BS, et al. Diagnosis of coronary in-Stent restenosis with multidetector row spiral computed tomography. J Am Coll Cardiol 2005;46(October (8)):1573–9.
[13] Hecht HS, Zaric M, Jelnin V, Lubarsky L, Prakash M, Roubin G. Usefulness of 64-detector computed tomographic angiography for diagnosing in-stent restenosis in native coronary arteries. Am J Cardiol 2008;101(March (6)):820–4. [14] Suzuki S, Furui S, Kuwahara S, et al. Assessment of coronary stents using 40detector row CT: comparison with 16-detector row CT in vitro. J Comput Assist Tomogr 2007;31(November–December (6)):910–6. [15] Suzuki S, Furui S, Kuwahara S, et al. Detection of in-stent restenosis of coronary stents using 40-detector row computed tomography in vitro. J Comput Assist Tomogr 2008;32(March April (2)):252–8. [16] Korosoglou G, Mueller D, Lehrke S, et al. Quantitative assessment of stenosis severity and atherosclerotic plaque composition using 256-slice computed tomography. Eur Radiol 2010;(March (20)) [Epub ahead of print]. [17] Flohr T, Stierstorfer K, Raupach R, Ulzheimer S, Bruder H. Performance evaluation of a 64-slice CT system with z-flying focal spot. Rofo 2004;176(December (12)):1803–10. [18] Sun Z, Davidson R, Lin CH. Multi-detector row CT angiography in the assessment of coronary in-stent restenosis: a systematic review. Eur J Radiol 2009;69(March (3)):489–95. [19] Pugliese F, Weustink AC, Van Mieghem C, et al. Dual source coronary computed tomography angiography for detecting in-stent restenosis. Heart 2008;94(July (7)):848–54. [20] Wykrzykowska JJ, Arbab-Zadeh A, Godoy G, et al. Assessment of in-stent restenosis using 64-MDCT: analysis of the CORE-64 Multicenter International Trial. AJR 2010;194:85–92. [21] Maintz D, Seifarth H, Flohr T, et al. Improved coronary artery stent visualization and in-stent stenosis detection using 16-slice computed-tomography and dedicated image reconstruction technique. Invest Radiol 2003;38(December (12)):790–5. [22] Kerl JM, Schoepf UJ, Vogl TJ, et al. In vitro evaluation of metallic coronary artery stents with 64-MDCT using an ECG-gated cardiac phantom: relationship between in-stent visualization, stent type, and heart rate. AJR Am J Roentgenol 2010;194(March (3)):W256–62.