- Email: [email protected]
Coronary CT angiography-derived quantitative markers for predicting in-stent restenosis
Journal of Cardiovascular Computed Tomography 10 (2016) 377e383
Contents lists available at ScienceDirect
Journal of Cardiovascular Computed Tomography journal homepage: www.JournalofCardiovascularCT.com
Research paper
Coronary CT angiography-derived quantitative markers for predicting in-stent restenosis Christian Tesche a, b, Carlo N. De Cecco a, c, Rozemarijn Vliegenthart a, d, Taylor M. Duguay a, Andrew C. Stubenrauch a, Russell D. Rosenberg a, e, Akos Varga-Szemes a, Richard R. Bayer 2nd a, e, Junjie Yang a, f, Ullrich Ebersberger a, b, Moritz Baquet g, David Jochheim g, Ellen Hoffmann b, Daniel H. Steinberg e, Salvatore A. Chiaramida e, U. Joseph Schoepf a, e, * a
Division of Cardiovascular Imaging, Department of Radiology and Radiological Science, Medical University of South Carolina, Charleston, SC, USA Department of Cardiology and Intensive Care Medicine, Heart Center Munich-Bogenhausen, Munich, Germany c Department of Radiological Sciences, Oncology and Pathology, University of Rome “Sapienza”, Rome, Italy d University of Groningen, University Medical Center Groningen, Department of Radiology, Groningen, The Netherlands e Division of Cardiology, Department of Medicine, Medical University of South Carolina, Charleston, SC, USA f Department of Cardiology, People’s Liberation Army General Hospital, Beijing, China g Department of Cardiology, Hospital of the Ludwig-Maximilians-University, Munich, Germany b
a r t i c l e i n f o
a b s t r a c t
Article history: Received 30 June 2016 Accepted 3 July 2016 Available online 6 July 2016
Objective: To evaluate quantitative markers derived from coronary CT angiography (coronary CTA) performed prior to percutaneous coronary intervention (PCI) with stent placement for predicting in-stent restenosis (ISR) as defined by quantitative coronary angiography (QCA). Materials and methods: We retrospectively analyzed the data of 74 patients (60 ± 12 years, 72% male) who had undergone dual-source coronary CTA within 3 months prior to a PCI procedure that included stent placement. Quantitative markers of the target vessel were derived from coronary CTA: Total plaque volume (TPV), calcified and non-calcified plaque volumes (CPV and NCPV), plaque burden (PB in %), remodeling index (RI), and lesion length (LL). Marker performance for predicting ISR, as defined by QCA at follow-up, was assessed. Results: Twenty-one of 74 stented lesions showed ISR on follow-up (mean 616 ± 447 days). When comparing stent length and LL in patients with ISR, a trend towards less complete stent coverage of the target lesion was observed in cases with ISR (17/21 vs. 4/53 cases, p ¼ 0.07). In multivariate analysis (corrected for dyslipidemia), the following markers showed predictive value for ISR (odds ratio [OR]): NCPV (OR 1.08, p ¼ 0.045), LL (OR 1.38, p ¼ 0.0024), and RI (OR 1.13, p ¼ 0.0019). Sensitivity and specificity for ISR were: NCPV 65% and 80%, LL 74% and 74%, and RI 71% and 78%. At receiver-operating characteristics analysis, NCPV (0.72, p ¼ 0.001), LL (0.77, p < 0.0001), and RI (0.79, p < 0.0001) showed discriminatory power for predicting ISR. A combination of these markers showed incremental predictive value (AUC 0.89, p < 0.0001) with sensitivity and specificity of 90% and 84%, respectively. Conclusion: Coronary CTA-derived NCPV, LL, and RI portend predictive value for ISR with incremental predictive value when combining these parameters. © 2016 Society of Cardiovascular Computed Tomography. Published by Elsevier Inc. All rights reserved.
Keywords: Coronary artery disease Quantitative coronary angiography In-stent restenosis Coronary computed tomography angiography
1. Introduction
* Corresponding author. Heart & Vascular Center, Medical University of South Carolina, Ashley River Tower, 25 Courtenay Drive, Charleston, SC 29425-2260, USA. E-mail address: [email protected] (U.J. Schoepf).
Percutaneous coronary intervention (PCI) with stent placement is a standard therapy for myocardial revascularization of hemodynamically significant coronary artery disease (CAD).1 However, instent restenosis (ISR) after stent placement is possible and clinically relevant. Newer generation drug-eluting stents (DES) have
http://dx.doi.org/10.1016/j.jcct.2016.07.005 1934-5925/© 2016 Society of Cardiovascular Computed Tomography. Published by Elsevier Inc. All rights reserved.
378
C. Tesche et al. / Journal of Cardiovascular Computed Tomography 10 (2016) 377e383
Abbreviations BMI CAD CTA CI CPV DSCT ICA ISR IVUS LAD
Body mass index Coronary artery disease CT angiography Confidence interval Calcified plaque volume Dual-source CT Invasive catheter angiography In-stent restenosis Intravascular ultrasound Left anterior descending
decreased the incidence of ISR as compared to bare metal stents (BMS), with ISR rates of up to 5% and 30%, respectively.2e5 The development of ISR can be attributed to procedural factors, lesionrelated characteristics, and patient factors.6,7 Previous studies using pre-interventional angiographic and intravascular ultrasound (IVUS) measurements during invasive coronary angiography (ICA) have shown that lesion-related markers such as lesion length, plaque burden, plaque composition, and positive remodeling were independent predictors of ISR.8e10 Recent studies using coronary CT angiography (CTA) for plaque quantification have proposed that quantitative morphological and functional markers derived from coronary CTA for coronary artery stenosis characterization may enhance the performance of coronary CTA for predicting hemodynamically significant stenosis, acute coronary syndrome, and future cardiac events.11e15 However, the potential role of coronary CTA-derived lesion-specific markers to predict ISR has been insufficiently investigated to date. Thus, we sought to evaluate the performance of coronary CTAderived quantitative morphological and functional markers for predicting ISR as defined by quantitative catheter angiography (QCA).
LCX LL LM NCPV NPV OR PPV QCA RCA RI SD TPV
Left circumflex artery Lesion length Left main Non-calcified plaque volume Negative predictive value Odds ratio Positive predictive value Quantitative coronary angiography Right coronary artery Remodeling index Standard deviation Total plaque volume
voltage; tube current, 75 mA; 3-mm slice thickness with 1.5 mm increment). For the subsequent contrast-enhanced coronary CTA, image acquisition parameters were as follows: a retrospectively ECG-gated protocol for the 1st generation DSCT scanner and a prospectively ECG-triggered sequential scan protocol with padding window for the 2nd generation DSCT scanner; tube voltage of 100e120 kV, tube current of 320e412 mA, temporal resolution of 83 or 75 ms, and 2 32 0.6 mm or 2 64 0.6 mm collimation with z-flying focal spot. Contrast enhancement was achieved by injecting 50e80 mL iopromide (Ultravist 370mgI/mL, Bayer, Wayne, NJ) at 4e6 mL/s followed by a 30 mL saline bolus chaser. Beta-blockers and nitroglycerine were used at the discretion of the attending physician. Weighted filtered back projection image reconstruction was performed in the cardiac phase with the least motion unsing a section thickness of 0.75 mm, reconstruction increment of 0.5 mm and a smooth convolution kernel (B26f). Effective radiation dose was measured by multiplying the doselength-product by the chest-specific conversion coefficient k ¼ 0.014 mSv/Gy/cm.
2.3. Analysis of coronary CTA data and quantitative markers 2. Material and methods 2.1. Study population This study was approved by the local Institutional Review Board with a waiver of informed consent. We retrospectively analyzed the data of a patient cohort with suspected or known CAD who had undergone dual-source coronary CTA within 3 months prior to a PCI procedure that included stent placement between March 2007 and May 2012. Another inclusion criterion was the availability of ICA follow-up for the evaluation of ISR. Patients were excluded if they had a prior percutaneous coronary stent implantation or coronary artery bypass grafting (CABG) in the vessel of interest at the time of coronary CTA. In addition, patients with stent placement in bifurcation stenoses, ostial lesions, bypass grafts, and/or with nondiagnostic CT image quality were excluded. Covariates, including cardiac risk factors and patient baseline characteristics were obtained from medical records. 2.2. Coronary CTA data acquisition All CT examinations were performed by using either a 1st or 2nd generation dual-source CT (DSCT) system (Somatom Definition or Somatom Definition Flash, Siemens Healthineers, Forchheim, Germany). All patients initially underwent a non-contrast enhanced calcium scoring scan (collimation, 32 1.2 mm; 120 kV tube
CT data were transferred to a post-processing workstation (syngo.via VA30, Siemens, Forchheim, Germany) for further analysis. For this retrospective study, two observers who were blinded to the patients’ history independently analyzed the lesion characteristics. Discordant cases were resolved by a consensus reading. Transverse sections and automatically generated curved multiplanar reformations were available for assessment. In accordance with current guidelines,16,17 the target lesion on ICA where the stent was eventually deployed was evaluated on coronary CTA using the 18-segment coronary artery model. As the reference for diameter and area stenosis determination, average dimensions of nonaffected vessel segments immediately proximal and distal to the lesion of interest were measured at sections free of atherosclerotic plaque. Coronary artery stenoses were graded as mild (25e49% stenosis), moderate (50e69% stenosis), severe (70e99% stenosis), and totally occluded. For the analysis of morphological characteristics a dedicated semi-automatic software prototype was used (Coronary Plaque Analysis 2.0.3 syngo.via FRONTIER, Siemens). This analysis software uses automated segmentation based on attenuation values prevailing in the target anatomy within user-defined proximal and distal boundaries to compute a comprehensive array of quantitative atherosclerosis lesion descriptors. The start and end of a target lesion was defined as the proximal and distal non-diseased section with absence of atherosclerotic changes, adjacent to the atheroma shoulders. Lesion length (LL), total plaque volume (TPV), calcified
C. Tesche et al. / Journal of Cardiovascular Computed Tomography 10 (2016) 377e383
plaque volume (CPV), and non-calcified plaque volume (NCPV) were automatically measured by the software with the following cut-off values (HU) for the different plaque characteristics: lipidrich (17-70HU), fibrotic (71-124HU), vessel lumen (125-511HU), and calcified (>511HU).18 Plaque burden (PB in %) was measured as follows: PB ¼ [plaque area/vessel area] x 100. On vessel crosssections, remodeling index (RI) was measured as the ratio of the vessel area of the lesion over the proximal luminal reference area.19 Coverage of the atheromatous plaque by the implanted stent was measured as the ratio of LL derived from coronary CTA and stent length. 2.4. Invasive procedures PCI, stent placement, and QCA were performed using standard techniques in our catheterization laboratory.1 At least two views were obtained of each major epicardial vessel in the same projections both during PCI and ICA/QCA follow-up. Additional IVUS to guide stent placement was at the discretion of the interventional cardiologist. ISR on QCA follow-up was defined as luminal stenosis >50% within the previously implanted stent. Balloon angioplasty was performed for focal ISR, while repeated stent placement was used for diffuse ISR or total stent occlusion. 2.5. Statistical analysis MedCalc (MedCalc Software, version 15, Ostend, Belgium) was used for all statistical analyses. Continuous variables were presented as mean ± standard deviation or median with interquartile range when not normally distributed. Student t-test and MannWhitney U test were used for parametric or non-parametric data. Receiver-operating characteristics (ROC) analysis was used to characterize predictors of ISR. For the evaluation of discriminatory power, the area under the ROC curve (AUC) was measured according to the method of DeLong.20 Multivariate logistic regression analysis was performed with angiographic ISR as a dichotomous outcome, with coronary CTA-derived lesion-related markers as independent variables to identify possible independent predictors. Correction for clinical risk factors was performed for significant (p < 0.05) univariable variables. Using the Youden index for the optimal threshold, sensitivity, specificity, positive predictive value, and negative predictive value from ROC curve analysis were measured and presented with a 95% confidence interval. Statistical
379
significance was assumed with a p-value 0.05. 3. Results 3.1. Patient characteristics In this single-center retrospective study, a total of 74 stents in 74 patients (60 ± 12 years, 72% male) who had undergone dual-source coronary CTA within 3 months prior to PCI with stent placement were included. Clinical indications for initial ICA/QCA and follow up catheterization included abnormal exercise or nuclear stress test results, recurring chest pain, and/or routine follow-up. A flow diagram of the study is shown in Fig. 1. Further patient demographics and baseline characteristics are presented in Table 1. 3.2. QCA with stent placement Mean time to the initial ICA/PCI after coronary CTA was 14.7 ± 28.8 days (ISR mean 15.3 ± 31.7 days, no ISR mean 13.4 ± 26.7 days, p ¼ 0.247). A total of 74 lesions in 74 patients were treated with stent placement: 47 in the LAD (64%), 9 in the LCX (12%), and 18 in the RCA (24%). Forty-six (62%) DES and 28 (38%) BMS were deployed. Mean stent length was 18.3 ± 5.8 mm without significant differences between patients with ISR (18.7 ± 5.9 mm) and without ISR (17.9 ± 5.8 mm, p ¼ 0.55). Mean stent diameter was 3.0 ± 0.5 mm with no significant differences between patients with ISR (3.1 ± 0.5 mm) and without (2.9 ± 0.4 mm, p ¼ 0.21). On ICA/ QCA follow-up, 21 stents showed significant ISR (14 DES, 7 BMS). Of these, 11 were in the LAD, 3 in the LCX, and 7 in the RCA. The mean time to ICA/QCA follow-up was 617 ± 447 days with no statistical differences between the groups (ISR mean 484 ± 511 days, no ISR mean 692 ± 395 days, p ¼ 0.09). Plaque coverage was sufficient in only 26 patients (35%) and insufficient in 48 patients (65%), with a considerable trend towards a mismatch (17/21 vs. 4/53, p ¼ 0.07) between stent length and LL for patients with ISR versus patients without ISR. Location of stents with ISR throughout the vessel tree (proximal, medial, distal) showed no statistically significant (p ¼ 0.06) differences between groups. Further procedural results are shown in Tables 2 and 3. 3.3. Association of CT-derived quantitative markers and ISR Mean TPV was 126.9 ± 81.6 mm3 with significant differences
Fig. 1. Flow diagram of the study.
380
C. Tesche et al. / Journal of Cardiovascular Computed Tomography 10 (2016) 377e383
Table 1 Patient demographics and baseline characteristics of the study population (n ¼ 74). Parameter
All patients (n ¼ 74)
Patients with ISR (n ¼ 21)
Patients without ISR (n ¼ 53)
p value
Age (years) Male sex n (%) Body-mass-index (kg/m2) Cardiovascular risk factors Hypertension n (%) Diabetes n (%) Dyslipidemia n (%) Current smoker n (%) CAD family history n (%)
60 ± 12 53 (72%) 29.8 ± 6.4
59 ± 13 16 (30%) 28.2 ± 5.0
60 ± 11 37 (70%) 28.8 ± 6.7
0.65 0.48 0.74
53 24 49 36 42
17 10 19 15 13
36 14 30 21 29
0.23 0.09 0.006 0.07 0.57
(72%) (32%) (66%) (49%) (57%)
(32%) (42%) (39%) (42%) (31%)
(68%) (58%) (61%) (58%) (69%)
CAD ¼ coronary artery disease. Data presented as mean ± standard deviation or numbers with percentages (%).
between patients with ISR (143.9 ± 105.5 mm3) and patients without (101.9 ± 51.1 mm3, p ¼ 0.045). Plaque composition revealed a mean CPV of 5.8 ± 3.1 mm3 with lower CPV in patients with ISR (4.9 ± 3.4 mm3) compared to patients without (6.9 ± 4.1 mm3, p < 0.001). Conversely, NCPV (mean 109.1 ± 62.0 mm3) was significantly (p ¼ 0.013) higher in ISR patients (136.0 ± 80.7 mm3) compared to patients with no ISR (92.2 ± 50.3 mm3) (Table 4). Mean LL was 21.2 ± 6.9 mm with a significant differences in patients with ISR (23.9 ± 6.0 mm) and patients without ISR (19.8 ± 6.5 mm, p < 0.001). Mean RI was 0.96 ± 0.17 with a significantly higher RI in patients with ISR (1.02 ± 0.10) compared to patients without (0.91 ± 0.19, p ¼ 0.0002) (Table 4). A representative example of the semi-automatic software prototype is shown in Fig. 2. At multivariate analysis (corrected for dyslipidemia), NCPV (OR 1.08 per mm3, p ¼ 0.045), LL (OR 1.38 per mm, p ¼ 0.0024), and RI (OR 1.13, p ¼ 0.0019) were significant predictors of ISR, whereas TPV and CPV were not (OR 1.01 per mm3, p ¼ 0.12 and OR 0.91 per mm3, p ¼ 0.36) (Table 5). ROC curve analysis showed discriminatory
Table 2 Procedural results. Quantitative coronary angiography No. of lesions with stent placement n Left anterior descending artery n (%) Left circumflex artery n (%) Right coronary artery n (%) Stent placement after CTA (days) Fluoroscopy time (min) Contrast agent (cc) ICA/QCA mean follow-up time (days) No. of lesions with in-stent restenosis n Left anterior descending artery n (%) Left circumflex artery n (%) Right coronary artery n (%) Coronary CT Angiography Heart rate (bpm) Agatston coronary artery calcium score Dose-length-product (mGy*cm) Estimated effective radiation dose (mSv)
74 47 (64%) 9 (12%) 18 (24%) 14.7 ± 28.8 18.9 ± 12.5 154.3 ± 65.7 616.9 ± 447.4 21 11 (52%) 3 (14%) 7 (33%) 69.8 ± 12.6 348.2 ± 484.6 481 ± 54.8 6.7 ± 0.4
Data presented as mean ± standard deviation or numbers with percentages (%).
Table 3 Analysis of stent characteristics. Parameter
All lesions (n ¼ 74)
Lesions with ISR (n ¼ 21)
Lesions without ISR (n ¼ 53)
p value
DES BMS Stent length Stent diameter
46 (62%) 28 (38%) 18.3 ± 5.8 3.0 ± 0.5
12 (26%) 9 (32%) 18.7 ± 5.9 3.1 ± 0.5
34 (74%) 19 (68%) 17.9 ± 5.8 2.9 ± 0.4
0.56 0.57 0.55 0.21
Data are presented as mean ± standard deviation or numbers (%); DES ¼ drug eluting stent, BMS ¼ bare metal stent.
power for RI (AUC 0.79, p < 0.0001), LL (AUC 0.77, p < 0.0001), and NCPV (AUC 0.72, p ¼ 0.001). The combination of these markers (RI þ LL þ NCPV) yielded an AUC of 0.89 (p < 0.0001) with a corresponding sensitivity and specificity of 90% and 84%, respectively. ROC curve analyses and the calculated sensitivity, specificity, PPV, and NPV for the markers are shown in Fig. 3 and Table 6. 4. Discussion This study evaluated coronary lesion characteristics derived from coronary CTA regarding their ability to predict restenosis after treatment by stent placement. Our results show that non-calcified plaque volume (NCPV) portends predictive value for ISR, resulting in an AUC value of 0.72 (p ¼ 0.001) and a corresponding sensitivity and specificity of 65% and 80%, respectively. RI and LL as markers also showed predictive power with AUC values of 0.79 (p < 0.0001) and 0.77 (p < 0.0001). Sensitivity and specificity were 71% and 78% (RI) and 74% and 74% (LL), respectively. These results, obtained by non-invasive coronary CTA, are in accordance with previous studies based on invasive angiography and IVUS, which showed a significant association of anatomic and morphological markers with ISR.6,8,9 We demonstrated a discriminatory power of NCPV (AUC 0.72, p ¼ 0.001) for predicting ISR. Farb et al. showed that stent strut penetration into the plaque lipid core of non-calcified plaques increases inflammatory endothelial reaction and is associated with higher neointimal thickening, in turn leading to more frequent ISR.8 Our findings appear to confirm these results and further support the prognostic value of NCPV as a morphological marker. Contrary to previous reports based on invasive testing,9,21 TPV and CPV failed to show discriminatory power to predict ISR in our investigation (p ¼ 0.12 and p ¼ 0.36). Furthermore, total plaque burden yielded no significant difference between lesions with or without ISR (p ¼ 0.16), which is different from previous invasive studies demonstrating a significant impact of total plaque burden on ISR.9,22 Table 4 Analysis of coronary CTA-derived lesion-related markers. Parameter
All lesions (n ¼ 74)
Lesions with ISR (n ¼ 21)
Lesions without ISR (n ¼ 53)
p value
TPV (mm3) CPV (mm3) NCPV (mm3) LL (mm) PB (%) RI
126.9 ± 81.6 5.8 ± 3.1 109.1 ± 62.0 21.2 ± 6.9 63.8 ± 12.3 0.96 ± 0.17
143.9 ± 105.5 4.9 ± 3.4 136.0 ± 80.7 23.9 ± 6.0 66.8 ± 10.2 1.02 ± 0.10
101.9 ± 51.1 6.9 ± 4.1 92.2 ± 50.3 19.8 ± 6.5 62.1 ± 13.3 0.91 ± 0.19
0.045 <0.001 0.013 <0.001 0.16 0.0002
Data presented as mean ± standard deviation; LL ¼ lesion length, TPV ¼ total plaque volume, CPV ¼ calcified plaque volume, NCPV ¼ non-calcified plaque volume, PB ¼ plaque burden, RI ¼ remodeling index.
C. Tesche et al. / Journal of Cardiovascular Computed Tomography 10 (2016) 377e383
381
Fig. 2. 55-year old man presenting with chest pain. (A) Coronary CTA demonstrates severe stenosis of the distal LAD (arrow). (B) Color-coded automated plaque evaluation of the causative lesion by the analysis software quantitates (C) the predominantly non-calcified composition of the underlying atheroma. Vessel RI is measured at 1.2. Subsequent ICA confirms severe stenosis (C, arrow) which was successfully treated by stent placement (D, arrow). The patient returned 14 months after PCI with recurrent angina. (F) Repeat ICA shows severe ISR (arrow), which was subsequently treated with balloon angioplasty and repeated stent placement.
LL and RI as functional markers showed significant discriminatory power and emerged as predictors for ISR. LL demonstrated an AUC of 0.77 (p < 0.0001) and thus provided incremental diagnostic value. The validity of LL as an independent predictor of ISR has been demonstrated in several studies based on invasive testing.23,24 Goldberg et al. demonstrated that a longer baseline lesion length on ICA is a strong predictor for diffuse ISR.25 Van Velzen et al. showed that mean LL on coronary CTA was significantly longer compared to QCA measurements and also longer than the mean length of the stents eventually deployed, resulting in insufficient plaque coverage.26 However, the authors of this previous investigation did not systematically analyze the prognostic implications of this observation for ISR. Analogously, in our study LL analysis of atheromatous plaque extent on coronary CTA yielded a strong trend (17/21 vs. 4/53, p ¼ 0.07) linking a mismatch between stent length
and LL, resulting in insufficient coverage of the atheromatous area, with the eventual development of ISR. However, it is worth noting that other ICA studies reported no significant role of LL for predicting ISR.27 Positive vessel remodeling on QCA evaluation has previously shown an association with ISR.8,9 Sahara et al. demonstrated that on IVUS evaluation, RI was the most powerful predictor for ISR, showing that positively remodeled lesions tend to develop into ISR after stenting.10 In our study, lesions with ISR also showed a higher RI compared to lesions without ISR with RI demonstrating significant prognostic power (AUC 0.79, p < 0.0001). The combination of anatomic and morphological markers (RI þ LL þ NCPV) yielded incremental predictive power (AUC of 0.89, p < 0.0001) and further corroborate the prognostic value of these coronary CTA-derived parameters for predicting ISR.
382
C. Tesche et al. / Journal of Cardiovascular Computed Tomography 10 (2016) 377e383
Table 5 Multivariate logistic regression analysis of coronary CTA-derived lesion-related markers (corrected for dyslipidemia). Parameter
Odds ratio
95% CI of Odds ratio
p value
TPV (mm3) CPV (mm3) NCPV (mm3) LL (mm) RI
1.01 0.91 1.08 1.38 1.13
0.73e1.31 0.59e1.40 0.99e1.21 1.12e1.71 1.03e1.22
0.12 0.36 0.045 0.0024 0.0019
Data presented as mean ± standard deviation; CI ¼ confidence interval, LL ¼ lesion length, TPV ¼ total plaque volume, CPV ¼ calcified plaque volume, NCPV ¼ noncalcified plaque volume, RI ¼ remodeling index.
Several limitations of this study need to be taken into consideration. We performed a retrospective investigation which is therefore subject to limitations inherent in this study design. A relatively small number of patients were included; therefore, larger, possibly prospective studies will be necessary to validate our findings. Our results on multivariate analysis may be underpowered by the limited number of observations per variable included.28 Therefore, this data generated in this study should only be hypothesis generating. We only evaluated one lesion in each patient, where stent placement was performed. Furthermore, we excluded bypass grafts, bifurcation lesions, left main revascularization, and ostial lesions, which are prone to higher restenosis rates.9 Additionally, we did not evaluate the influence of procedural results
Table 6 Diagnostic performance of coronary CTA-derived lesion-related markers for the prediction of ISR according to QCA. Parameter
LL > 21.1 (95%CI)
RI > 1.03 (95%CI)
NCPV > 131.7 (95%CI)
Sensitivity
74% (55e88%) 74% (59e86%) 66% (48e81%) 81% (66e91%)
71% (52e86%) 78% (64e89%) 69% (50e84%) 80% (65%e90%
65% (45e81%) 80% (65e90%) 69% (49e85%) 77% (62e88%)
Specificity PPV NPV
CI ¼ confidence interval, LL ¼ lesion length, NCPV ¼ non-calcified plaque volume, RI ¼ remodeling index. Data presented as percentages (%).
(e.g.: stent deployment, stent malapposition, stent fracture) and clinical parameters which are known to affect ISR.29 We did not correlate our findings on coronary CTA with intracoronary imaging techniques, as IVUS-guided stent placement had been performed in only 6 of the 21 cases with ISR and optical coherence tomography is not routinely performed at our institution. In conclusion this study demonstrates that coronary CTAderived NCPV, LL, and RI as quantitative markers of target lesion anatomy and CAD disease activity portend predictive value for ISR with incremental value of combination of these parameters. Obtaining these markers prior to PCI may guide the selection of an
Fig. 3. Diagnostic performance of Coronary CTA-derived lesion-related markers for the prediction of ISR.
C. Tesche et al. / Journal of Cardiovascular Computed Tomography 10 (2016) 377e383
appropriate revascularization strategy, e.g., selection of stents that cover the entire atheroma or DES to treat lesions with high disease activity by coronary CTA. Disclosures Dr. Schoepf is a consultant for and/or receives research support from Astellas, Bayer, Bracco, GE, Guerbet, Medrad, and Siemens. Drs. De Cecco and Varga-Szemes are consultants for Guerbet. All other authors have no conflicts of interest to disclose. Christian Tesche is an exchange visiting scholar supported by a grant from the Fulbright Visiting Scholar Program of the U.S. Department of State, Bureau of Educational and Cultural Affairs (ECA). References 1. Smith Jr SC, Dove JT, Jacobs AK, et al. ACC/AHA guidelines for percutaneous coronary intervention (revision of the 1993 PTCA guidelines)-executive summary: a report of the American College of Cardiology/American Heart Association task force on practice guidelines (Committee to revise the 1993 guidelines for percutaneous transluminal coronary angioplasty) endorsed by the Society for Cardiac Angiography and Interventions. Circulation. 2001;103: 3019e3041. 2. Palmerini T, Biondi-Zoccai G, Della Riva D, et al. Clinical outcomes with bioabsorbable polymer- versus durable polymer-based drug-eluting and baremetal stents: evidence from a comprehensive network meta-analysis. J Am Coll Cardiol. 2014;63:299e307. 3. Vink MA, Van Nooijen FC, Laarman GJ, et al. Patency of paclitaxel-eluting versus bare metal stents long term after implantation in acute ST-segment elevation myocardial infarction. Am J Cardiol. 2011;108:1214e1219. 4. Cohen DJ, Van Hout B, Serruys PW, et al. Quality of life after PCI with drugeluting stents or coronary-artery bypass surgery. N Engl J Med. 2011;364: 1016e1026. 5. Stenestrand U, James SK, Lindback J, et al. Safety and efficacy of drug-eluting vs. bare metal stents in patients with diabetes mellitus: long-term follow-up in the Swedish Coronary Angiography and Angioplasty Registry (SCAAR). Eur Heart J. 2010;31:177e186. 6. Holmes Jr DR, Kereiakes DJ, Garg S, et al. Stent thrombosis. J Am Coll Cardiol. 2010;56:1357e1365. 7. Mehran R, Dangas G, Abizaid AS, et al. Angiographic patterns of in-stent restenosis: classification and implications for long-term outcome. Circulation. 1999;100:1872e1878. 8. Farb A, Weber DK, Kolodgie FD, Burke AP, Virmani R. Morphological predictors of restenosis after coronary stenting in humans. Circulation. 2002;105: 2974e2980. 9. Hoffmann R, Mintz GS, Mehran R, et al. Intravascular ultrasound predictors of angiographic restenosis in lesions treated with Palmaz-Schatz stents. J Am Coll Cardiol. 1998;31:43e49. 10. Sahara M, Kirigaya H, Oikawa Y, et al. Arterial remodeling patterns before intervention predict diffuse in-stent restenosis: an intravascular ultrasound study. J Am Coll Cardiol. 2003;42:1731e1738. 11. Bittencourt MS, Hulten E, Ghoshhajra B, et al. Prognostic value of nonobstructive and obstructive coronary artery disease detected by coronary computed tomography angiography to identify cardiovascular events. Circ
383
Cardiovasc Imaging. 2014;7:282e291. 12. Bauer RW, Thilo C, Chiaramida SA, Vogl TJ, Costello P, Schoepf UJ. Noncalcified atherosclerotic plaque burden at coronary CT angiography: a better predictor of ischemia at stress myocardial perfusion imaging than calcium score and stenosis severity. AJR Am J Roentgenol. 2009;193:410e418. 13. Wang R, Renker M, Schoepf UJ, et al. Diagnostic value of quantitative stenosis predictors with coronary CT angiography compared to invasive fractional flow reserve. Eur J Radiol. 2015;84:1509e1515. 14. Wang R, Baumann S, Schoepf UJ, et al. Comparison of quantitative stenosis characteristics at routine coronary computed tomography angiography with invasive fractional flow reserve for assessing lesion-specific ischemia. J Cardiovasc Comput Tomogr. 2015;9:546e552. 15. Budoff MJ, Bloom SA, Chow BJ, Chandler Jr AB, Cole JH. Coronary computed tomography angiography predicts subsequent cardiac outcome events: results of the Visipaque CCTA registry study. Coron Artery Dis. 2015;26:301e307. 16. Leipsic J, Abbara S, Achenbach S, et al. SCCT guidelines for the interpretation and reporting of coronary CT angiography: a report of the Society of Cardiovascular Computed Tomography Guidelines Committee. J Cardiovasc Comput Tomogr. 2014;8:342e358. 17. Jacobs JE, Boxt LM, Desjardins B, et al. ACR practice guideline for the performance and interpretation of cardiac computed tomography (CT). J Am Coll Radiol. 2006;3:677e685. 18. Voros S, Rinehart S, Qian Z, et al. Coronary atherosclerosis imaging by coronary CT angiography: current status, correlation with intravascular interrogation and meta-analysis. JACC Cardiovasc Imaging. 2011;4:537e548. 19. Achenbach S, Ropers D, Hoffmann U, et al. Assessment of coronary remodeling in stenotic and nonstenotic coronary atherosclerotic lesions by multidetector spiral computed tomography. J Am Coll Cardiol. 2004;43:842e847. 20. DeLong ER, DeLong DM, Clarke-Pearson DL. Comparing the areas under two or more correlated receiver operating characteristic curves: a nonparametric approach. Biometrics. 1988;44:837e845. 21. Arakawa K, Ishibashi-Ueda H, Hao H, Ikeda Y, Kawamura A. Plaque tissue components obtained from de novo lesions may predict restenosis after directional coronary atherectomy. Ann Vasc Dis. 2010;3:52e59. 22. Prati F, Di Mario C, Moussa I, et al. In-stent neointimal proliferation correlates with the amount of residual plaque burden outside the stent: an intravascular ultrasound study. Circulation. 1999;99:1011e1014. 23. Rathore S, Terashima M, Katoh O, et al. Predictors of angiographic restenosis after drug eluting stents in the coronary arteries: contemporary practice in real world patients. EuroIntervention. 2009;5:349e354. 24. Kini A, Marmur JD, Dangas G, Choudhary S, Sharma SK. Angiographic patterns of in-stent restenosis and implications on subsequent revascularization. Catheter Cardiovasc Interv. 2000;49:23e29. 25. Goldberg SL, Loussararian A, De Gregorio J, Di Mario C, Albiero R, Colombo A. Predictors of diffuse and aggressive intra-stent restenosis. J Am Coll Cardiol. 2001;37:1019e1025. 26. van Velzen JE, de Graaf MA, Ciarka A, et al. Non-invasive assessment of atherosclerotic coronary lesion length using multidetector computed tomography angiography: comparison to quantitative coronary angiography. Int J Cardiovasc Imaging. 2012;28:2065e2071. 27. Chang SH, Chen CC, Hsieh MJ, Wang CY, Lee CH, Hsieh IC. Lesion length impacts long term outcomes of drug-eluting stents and bare metal stents differently. PLoS One. 2013;8:e53207. 28. Kunce JT. Random Variables and Correlational Overkill. Educational and Psychological Measurement. 29. Singh M, Gersh BJ, McClelland RL, et al. Clinical and angiographic predictors of restenosis after percutaneous coronary intervention: insights from the Prevention of Restenosis with Tranilast and its Outcomes (PRESTO) trial. Circulation. 2004;109:2727e2731.
Contents lists available at ScienceDirect
Journal of Cardiovascular Computed Tomography journal homepage: www.JournalofCardiovascularCT.com
Research paper
Coronary CT angiography-derived quantitative markers for predicting in-stent restenosis Christian Tesche a, b, Carlo N. De Cecco a, c, Rozemarijn Vliegenthart a, d, Taylor M. Duguay a, Andrew C. Stubenrauch a, Russell D. Rosenberg a, e, Akos Varga-Szemes a, Richard R. Bayer 2nd a, e, Junjie Yang a, f, Ullrich Ebersberger a, b, Moritz Baquet g, David Jochheim g, Ellen Hoffmann b, Daniel H. Steinberg e, Salvatore A. Chiaramida e, U. Joseph Schoepf a, e, * a
Division of Cardiovascular Imaging, Department of Radiology and Radiological Science, Medical University of South Carolina, Charleston, SC, USA Department of Cardiology and Intensive Care Medicine, Heart Center Munich-Bogenhausen, Munich, Germany c Department of Radiological Sciences, Oncology and Pathology, University of Rome “Sapienza”, Rome, Italy d University of Groningen, University Medical Center Groningen, Department of Radiology, Groningen, The Netherlands e Division of Cardiology, Department of Medicine, Medical University of South Carolina, Charleston, SC, USA f Department of Cardiology, People’s Liberation Army General Hospital, Beijing, China g Department of Cardiology, Hospital of the Ludwig-Maximilians-University, Munich, Germany b
a r t i c l e i n f o
a b s t r a c t
Article history: Received 30 June 2016 Accepted 3 July 2016 Available online 6 July 2016
Objective: To evaluate quantitative markers derived from coronary CT angiography (coronary CTA) performed prior to percutaneous coronary intervention (PCI) with stent placement for predicting in-stent restenosis (ISR) as defined by quantitative coronary angiography (QCA). Materials and methods: We retrospectively analyzed the data of 74 patients (60 ± 12 years, 72% male) who had undergone dual-source coronary CTA within 3 months prior to a PCI procedure that included stent placement. Quantitative markers of the target vessel were derived from coronary CTA: Total plaque volume (TPV), calcified and non-calcified plaque volumes (CPV and NCPV), plaque burden (PB in %), remodeling index (RI), and lesion length (LL). Marker performance for predicting ISR, as defined by QCA at follow-up, was assessed. Results: Twenty-one of 74 stented lesions showed ISR on follow-up (mean 616 ± 447 days). When comparing stent length and LL in patients with ISR, a trend towards less complete stent coverage of the target lesion was observed in cases with ISR (17/21 vs. 4/53 cases, p ¼ 0.07). In multivariate analysis (corrected for dyslipidemia), the following markers showed predictive value for ISR (odds ratio [OR]): NCPV (OR 1.08, p ¼ 0.045), LL (OR 1.38, p ¼ 0.0024), and RI (OR 1.13, p ¼ 0.0019). Sensitivity and specificity for ISR were: NCPV 65% and 80%, LL 74% and 74%, and RI 71% and 78%. At receiver-operating characteristics analysis, NCPV (0.72, p ¼ 0.001), LL (0.77, p < 0.0001), and RI (0.79, p < 0.0001) showed discriminatory power for predicting ISR. A combination of these markers showed incremental predictive value (AUC 0.89, p < 0.0001) with sensitivity and specificity of 90% and 84%, respectively. Conclusion: Coronary CTA-derived NCPV, LL, and RI portend predictive value for ISR with incremental predictive value when combining these parameters. © 2016 Society of Cardiovascular Computed Tomography. Published by Elsevier Inc. All rights reserved.
Keywords: Coronary artery disease Quantitative coronary angiography In-stent restenosis Coronary computed tomography angiography
1. Introduction
* Corresponding author. Heart & Vascular Center, Medical University of South Carolina, Ashley River Tower, 25 Courtenay Drive, Charleston, SC 29425-2260, USA. E-mail address: [email protected] (U.J. Schoepf).
Percutaneous coronary intervention (PCI) with stent placement is a standard therapy for myocardial revascularization of hemodynamically significant coronary artery disease (CAD).1 However, instent restenosis (ISR) after stent placement is possible and clinically relevant. Newer generation drug-eluting stents (DES) have
http://dx.doi.org/10.1016/j.jcct.2016.07.005 1934-5925/© 2016 Society of Cardiovascular Computed Tomography. Published by Elsevier Inc. All rights reserved.
378
C. Tesche et al. / Journal of Cardiovascular Computed Tomography 10 (2016) 377e383
Abbreviations BMI CAD CTA CI CPV DSCT ICA ISR IVUS LAD
Body mass index Coronary artery disease CT angiography Confidence interval Calcified plaque volume Dual-source CT Invasive catheter angiography In-stent restenosis Intravascular ultrasound Left anterior descending
decreased the incidence of ISR as compared to bare metal stents (BMS), with ISR rates of up to 5% and 30%, respectively.2e5 The development of ISR can be attributed to procedural factors, lesionrelated characteristics, and patient factors.6,7 Previous studies using pre-interventional angiographic and intravascular ultrasound (IVUS) measurements during invasive coronary angiography (ICA) have shown that lesion-related markers such as lesion length, plaque burden, plaque composition, and positive remodeling were independent predictors of ISR.8e10 Recent studies using coronary CT angiography (CTA) for plaque quantification have proposed that quantitative morphological and functional markers derived from coronary CTA for coronary artery stenosis characterization may enhance the performance of coronary CTA for predicting hemodynamically significant stenosis, acute coronary syndrome, and future cardiac events.11e15 However, the potential role of coronary CTA-derived lesion-specific markers to predict ISR has been insufficiently investigated to date. Thus, we sought to evaluate the performance of coronary CTAderived quantitative morphological and functional markers for predicting ISR as defined by quantitative catheter angiography (QCA).
LCX LL LM NCPV NPV OR PPV QCA RCA RI SD TPV
Left circumflex artery Lesion length Left main Non-calcified plaque volume Negative predictive value Odds ratio Positive predictive value Quantitative coronary angiography Right coronary artery Remodeling index Standard deviation Total plaque volume
voltage; tube current, 75 mA; 3-mm slice thickness with 1.5 mm increment). For the subsequent contrast-enhanced coronary CTA, image acquisition parameters were as follows: a retrospectively ECG-gated protocol for the 1st generation DSCT scanner and a prospectively ECG-triggered sequential scan protocol with padding window for the 2nd generation DSCT scanner; tube voltage of 100e120 kV, tube current of 320e412 mA, temporal resolution of 83 or 75 ms, and 2 32 0.6 mm or 2 64 0.6 mm collimation with z-flying focal spot. Contrast enhancement was achieved by injecting 50e80 mL iopromide (Ultravist 370mgI/mL, Bayer, Wayne, NJ) at 4e6 mL/s followed by a 30 mL saline bolus chaser. Beta-blockers and nitroglycerine were used at the discretion of the attending physician. Weighted filtered back projection image reconstruction was performed in the cardiac phase with the least motion unsing a section thickness of 0.75 mm, reconstruction increment of 0.5 mm and a smooth convolution kernel (B26f). Effective radiation dose was measured by multiplying the doselength-product by the chest-specific conversion coefficient k ¼ 0.014 mSv/Gy/cm.
2.3. Analysis of coronary CTA data and quantitative markers 2. Material and methods 2.1. Study population This study was approved by the local Institutional Review Board with a waiver of informed consent. We retrospectively analyzed the data of a patient cohort with suspected or known CAD who had undergone dual-source coronary CTA within 3 months prior to a PCI procedure that included stent placement between March 2007 and May 2012. Another inclusion criterion was the availability of ICA follow-up for the evaluation of ISR. Patients were excluded if they had a prior percutaneous coronary stent implantation or coronary artery bypass grafting (CABG) in the vessel of interest at the time of coronary CTA. In addition, patients with stent placement in bifurcation stenoses, ostial lesions, bypass grafts, and/or with nondiagnostic CT image quality were excluded. Covariates, including cardiac risk factors and patient baseline characteristics were obtained from medical records. 2.2. Coronary CTA data acquisition All CT examinations were performed by using either a 1st or 2nd generation dual-source CT (DSCT) system (Somatom Definition or Somatom Definition Flash, Siemens Healthineers, Forchheim, Germany). All patients initially underwent a non-contrast enhanced calcium scoring scan (collimation, 32 1.2 mm; 120 kV tube
CT data were transferred to a post-processing workstation (syngo.via VA30, Siemens, Forchheim, Germany) for further analysis. For this retrospective study, two observers who were blinded to the patients’ history independently analyzed the lesion characteristics. Discordant cases were resolved by a consensus reading. Transverse sections and automatically generated curved multiplanar reformations were available for assessment. In accordance with current guidelines,16,17 the target lesion on ICA where the stent was eventually deployed was evaluated on coronary CTA using the 18-segment coronary artery model. As the reference for diameter and area stenosis determination, average dimensions of nonaffected vessel segments immediately proximal and distal to the lesion of interest were measured at sections free of atherosclerotic plaque. Coronary artery stenoses were graded as mild (25e49% stenosis), moderate (50e69% stenosis), severe (70e99% stenosis), and totally occluded. For the analysis of morphological characteristics a dedicated semi-automatic software prototype was used (Coronary Plaque Analysis 2.0.3 syngo.via FRONTIER, Siemens). This analysis software uses automated segmentation based on attenuation values prevailing in the target anatomy within user-defined proximal and distal boundaries to compute a comprehensive array of quantitative atherosclerosis lesion descriptors. The start and end of a target lesion was defined as the proximal and distal non-diseased section with absence of atherosclerotic changes, adjacent to the atheroma shoulders. Lesion length (LL), total plaque volume (TPV), calcified
C. Tesche et al. / Journal of Cardiovascular Computed Tomography 10 (2016) 377e383
plaque volume (CPV), and non-calcified plaque volume (NCPV) were automatically measured by the software with the following cut-off values (HU) for the different plaque characteristics: lipidrich (17-70HU), fibrotic (71-124HU), vessel lumen (125-511HU), and calcified (>511HU).18 Plaque burden (PB in %) was measured as follows: PB ¼ [plaque area/vessel area] x 100. On vessel crosssections, remodeling index (RI) was measured as the ratio of the vessel area of the lesion over the proximal luminal reference area.19 Coverage of the atheromatous plaque by the implanted stent was measured as the ratio of LL derived from coronary CTA and stent length. 2.4. Invasive procedures PCI, stent placement, and QCA were performed using standard techniques in our catheterization laboratory.1 At least two views were obtained of each major epicardial vessel in the same projections both during PCI and ICA/QCA follow-up. Additional IVUS to guide stent placement was at the discretion of the interventional cardiologist. ISR on QCA follow-up was defined as luminal stenosis >50% within the previously implanted stent. Balloon angioplasty was performed for focal ISR, while repeated stent placement was used for diffuse ISR or total stent occlusion. 2.5. Statistical analysis MedCalc (MedCalc Software, version 15, Ostend, Belgium) was used for all statistical analyses. Continuous variables were presented as mean ± standard deviation or median with interquartile range when not normally distributed. Student t-test and MannWhitney U test were used for parametric or non-parametric data. Receiver-operating characteristics (ROC) analysis was used to characterize predictors of ISR. For the evaluation of discriminatory power, the area under the ROC curve (AUC) was measured according to the method of DeLong.20 Multivariate logistic regression analysis was performed with angiographic ISR as a dichotomous outcome, with coronary CTA-derived lesion-related markers as independent variables to identify possible independent predictors. Correction for clinical risk factors was performed for significant (p < 0.05) univariable variables. Using the Youden index for the optimal threshold, sensitivity, specificity, positive predictive value, and negative predictive value from ROC curve analysis were measured and presented with a 95% confidence interval. Statistical
379
significance was assumed with a p-value 0.05. 3. Results 3.1. Patient characteristics In this single-center retrospective study, a total of 74 stents in 74 patients (60 ± 12 years, 72% male) who had undergone dual-source coronary CTA within 3 months prior to PCI with stent placement were included. Clinical indications for initial ICA/QCA and follow up catheterization included abnormal exercise or nuclear stress test results, recurring chest pain, and/or routine follow-up. A flow diagram of the study is shown in Fig. 1. Further patient demographics and baseline characteristics are presented in Table 1. 3.2. QCA with stent placement Mean time to the initial ICA/PCI after coronary CTA was 14.7 ± 28.8 days (ISR mean 15.3 ± 31.7 days, no ISR mean 13.4 ± 26.7 days, p ¼ 0.247). A total of 74 lesions in 74 patients were treated with stent placement: 47 in the LAD (64%), 9 in the LCX (12%), and 18 in the RCA (24%). Forty-six (62%) DES and 28 (38%) BMS were deployed. Mean stent length was 18.3 ± 5.8 mm without significant differences between patients with ISR (18.7 ± 5.9 mm) and without ISR (17.9 ± 5.8 mm, p ¼ 0.55). Mean stent diameter was 3.0 ± 0.5 mm with no significant differences between patients with ISR (3.1 ± 0.5 mm) and without (2.9 ± 0.4 mm, p ¼ 0.21). On ICA/ QCA follow-up, 21 stents showed significant ISR (14 DES, 7 BMS). Of these, 11 were in the LAD, 3 in the LCX, and 7 in the RCA. The mean time to ICA/QCA follow-up was 617 ± 447 days with no statistical differences between the groups (ISR mean 484 ± 511 days, no ISR mean 692 ± 395 days, p ¼ 0.09). Plaque coverage was sufficient in only 26 patients (35%) and insufficient in 48 patients (65%), with a considerable trend towards a mismatch (17/21 vs. 4/53, p ¼ 0.07) between stent length and LL for patients with ISR versus patients without ISR. Location of stents with ISR throughout the vessel tree (proximal, medial, distal) showed no statistically significant (p ¼ 0.06) differences between groups. Further procedural results are shown in Tables 2 and 3. 3.3. Association of CT-derived quantitative markers and ISR Mean TPV was 126.9 ± 81.6 mm3 with significant differences
Fig. 1. Flow diagram of the study.
380
C. Tesche et al. / Journal of Cardiovascular Computed Tomography 10 (2016) 377e383
Table 1 Patient demographics and baseline characteristics of the study population (n ¼ 74). Parameter
All patients (n ¼ 74)
Patients with ISR (n ¼ 21)
Patients without ISR (n ¼ 53)
p value
Age (years) Male sex n (%) Body-mass-index (kg/m2) Cardiovascular risk factors Hypertension n (%) Diabetes n (%) Dyslipidemia n (%) Current smoker n (%) CAD family history n (%)
60 ± 12 53 (72%) 29.8 ± 6.4
59 ± 13 16 (30%) 28.2 ± 5.0
60 ± 11 37 (70%) 28.8 ± 6.7
0.65 0.48 0.74
53 24 49 36 42
17 10 19 15 13
36 14 30 21 29
0.23 0.09 0.006 0.07 0.57
(72%) (32%) (66%) (49%) (57%)
(32%) (42%) (39%) (42%) (31%)
(68%) (58%) (61%) (58%) (69%)
CAD ¼ coronary artery disease. Data presented as mean ± standard deviation or numbers with percentages (%).
between patients with ISR (143.9 ± 105.5 mm3) and patients without (101.9 ± 51.1 mm3, p ¼ 0.045). Plaque composition revealed a mean CPV of 5.8 ± 3.1 mm3 with lower CPV in patients with ISR (4.9 ± 3.4 mm3) compared to patients without (6.9 ± 4.1 mm3, p < 0.001). Conversely, NCPV (mean 109.1 ± 62.0 mm3) was significantly (p ¼ 0.013) higher in ISR patients (136.0 ± 80.7 mm3) compared to patients with no ISR (92.2 ± 50.3 mm3) (Table 4). Mean LL was 21.2 ± 6.9 mm with a significant differences in patients with ISR (23.9 ± 6.0 mm) and patients without ISR (19.8 ± 6.5 mm, p < 0.001). Mean RI was 0.96 ± 0.17 with a significantly higher RI in patients with ISR (1.02 ± 0.10) compared to patients without (0.91 ± 0.19, p ¼ 0.0002) (Table 4). A representative example of the semi-automatic software prototype is shown in Fig. 2. At multivariate analysis (corrected for dyslipidemia), NCPV (OR 1.08 per mm3, p ¼ 0.045), LL (OR 1.38 per mm, p ¼ 0.0024), and RI (OR 1.13, p ¼ 0.0019) were significant predictors of ISR, whereas TPV and CPV were not (OR 1.01 per mm3, p ¼ 0.12 and OR 0.91 per mm3, p ¼ 0.36) (Table 5). ROC curve analysis showed discriminatory
Table 2 Procedural results. Quantitative coronary angiography No. of lesions with stent placement n Left anterior descending artery n (%) Left circumflex artery n (%) Right coronary artery n (%) Stent placement after CTA (days) Fluoroscopy time (min) Contrast agent (cc) ICA/QCA mean follow-up time (days) No. of lesions with in-stent restenosis n Left anterior descending artery n (%) Left circumflex artery n (%) Right coronary artery n (%) Coronary CT Angiography Heart rate (bpm) Agatston coronary artery calcium score Dose-length-product (mGy*cm) Estimated effective radiation dose (mSv)
74 47 (64%) 9 (12%) 18 (24%) 14.7 ± 28.8 18.9 ± 12.5 154.3 ± 65.7 616.9 ± 447.4 21 11 (52%) 3 (14%) 7 (33%) 69.8 ± 12.6 348.2 ± 484.6 481 ± 54.8 6.7 ± 0.4
Data presented as mean ± standard deviation or numbers with percentages (%).
Table 3 Analysis of stent characteristics. Parameter
All lesions (n ¼ 74)
Lesions with ISR (n ¼ 21)
Lesions without ISR (n ¼ 53)
p value
DES BMS Stent length Stent diameter
46 (62%) 28 (38%) 18.3 ± 5.8 3.0 ± 0.5
12 (26%) 9 (32%) 18.7 ± 5.9 3.1 ± 0.5
34 (74%) 19 (68%) 17.9 ± 5.8 2.9 ± 0.4
0.56 0.57 0.55 0.21
Data are presented as mean ± standard deviation or numbers (%); DES ¼ drug eluting stent, BMS ¼ bare metal stent.
power for RI (AUC 0.79, p < 0.0001), LL (AUC 0.77, p < 0.0001), and NCPV (AUC 0.72, p ¼ 0.001). The combination of these markers (RI þ LL þ NCPV) yielded an AUC of 0.89 (p < 0.0001) with a corresponding sensitivity and specificity of 90% and 84%, respectively. ROC curve analyses and the calculated sensitivity, specificity, PPV, and NPV for the markers are shown in Fig. 3 and Table 6. 4. Discussion This study evaluated coronary lesion characteristics derived from coronary CTA regarding their ability to predict restenosis after treatment by stent placement. Our results show that non-calcified plaque volume (NCPV) portends predictive value for ISR, resulting in an AUC value of 0.72 (p ¼ 0.001) and a corresponding sensitivity and specificity of 65% and 80%, respectively. RI and LL as markers also showed predictive power with AUC values of 0.79 (p < 0.0001) and 0.77 (p < 0.0001). Sensitivity and specificity were 71% and 78% (RI) and 74% and 74% (LL), respectively. These results, obtained by non-invasive coronary CTA, are in accordance with previous studies based on invasive angiography and IVUS, which showed a significant association of anatomic and morphological markers with ISR.6,8,9 We demonstrated a discriminatory power of NCPV (AUC 0.72, p ¼ 0.001) for predicting ISR. Farb et al. showed that stent strut penetration into the plaque lipid core of non-calcified plaques increases inflammatory endothelial reaction and is associated with higher neointimal thickening, in turn leading to more frequent ISR.8 Our findings appear to confirm these results and further support the prognostic value of NCPV as a morphological marker. Contrary to previous reports based on invasive testing,9,21 TPV and CPV failed to show discriminatory power to predict ISR in our investigation (p ¼ 0.12 and p ¼ 0.36). Furthermore, total plaque burden yielded no significant difference between lesions with or without ISR (p ¼ 0.16), which is different from previous invasive studies demonstrating a significant impact of total plaque burden on ISR.9,22 Table 4 Analysis of coronary CTA-derived lesion-related markers. Parameter
All lesions (n ¼ 74)
Lesions with ISR (n ¼ 21)
Lesions without ISR (n ¼ 53)
p value
TPV (mm3) CPV (mm3) NCPV (mm3) LL (mm) PB (%) RI
126.9 ± 81.6 5.8 ± 3.1 109.1 ± 62.0 21.2 ± 6.9 63.8 ± 12.3 0.96 ± 0.17
143.9 ± 105.5 4.9 ± 3.4 136.0 ± 80.7 23.9 ± 6.0 66.8 ± 10.2 1.02 ± 0.10
101.9 ± 51.1 6.9 ± 4.1 92.2 ± 50.3 19.8 ± 6.5 62.1 ± 13.3 0.91 ± 0.19
0.045 <0.001 0.013 <0.001 0.16 0.0002
Data presented as mean ± standard deviation; LL ¼ lesion length, TPV ¼ total plaque volume, CPV ¼ calcified plaque volume, NCPV ¼ non-calcified plaque volume, PB ¼ plaque burden, RI ¼ remodeling index.
C. Tesche et al. / Journal of Cardiovascular Computed Tomography 10 (2016) 377e383
381
Fig. 2. 55-year old man presenting with chest pain. (A) Coronary CTA demonstrates severe stenosis of the distal LAD (arrow). (B) Color-coded automated plaque evaluation of the causative lesion by the analysis software quantitates (C) the predominantly non-calcified composition of the underlying atheroma. Vessel RI is measured at 1.2. Subsequent ICA confirms severe stenosis (C, arrow) which was successfully treated by stent placement (D, arrow). The patient returned 14 months after PCI with recurrent angina. (F) Repeat ICA shows severe ISR (arrow), which was subsequently treated with balloon angioplasty and repeated stent placement.
LL and RI as functional markers showed significant discriminatory power and emerged as predictors for ISR. LL demonstrated an AUC of 0.77 (p < 0.0001) and thus provided incremental diagnostic value. The validity of LL as an independent predictor of ISR has been demonstrated in several studies based on invasive testing.23,24 Goldberg et al. demonstrated that a longer baseline lesion length on ICA is a strong predictor for diffuse ISR.25 Van Velzen et al. showed that mean LL on coronary CTA was significantly longer compared to QCA measurements and also longer than the mean length of the stents eventually deployed, resulting in insufficient plaque coverage.26 However, the authors of this previous investigation did not systematically analyze the prognostic implications of this observation for ISR. Analogously, in our study LL analysis of atheromatous plaque extent on coronary CTA yielded a strong trend (17/21 vs. 4/53, p ¼ 0.07) linking a mismatch between stent length
and LL, resulting in insufficient coverage of the atheromatous area, with the eventual development of ISR. However, it is worth noting that other ICA studies reported no significant role of LL for predicting ISR.27 Positive vessel remodeling on QCA evaluation has previously shown an association with ISR.8,9 Sahara et al. demonstrated that on IVUS evaluation, RI was the most powerful predictor for ISR, showing that positively remodeled lesions tend to develop into ISR after stenting.10 In our study, lesions with ISR also showed a higher RI compared to lesions without ISR with RI demonstrating significant prognostic power (AUC 0.79, p < 0.0001). The combination of anatomic and morphological markers (RI þ LL þ NCPV) yielded incremental predictive power (AUC of 0.89, p < 0.0001) and further corroborate the prognostic value of these coronary CTA-derived parameters for predicting ISR.
382
C. Tesche et al. / Journal of Cardiovascular Computed Tomography 10 (2016) 377e383
Table 5 Multivariate logistic regression analysis of coronary CTA-derived lesion-related markers (corrected for dyslipidemia). Parameter
Odds ratio
95% CI of Odds ratio
p value
TPV (mm3) CPV (mm3) NCPV (mm3) LL (mm) RI
1.01 0.91 1.08 1.38 1.13
0.73e1.31 0.59e1.40 0.99e1.21 1.12e1.71 1.03e1.22
0.12 0.36 0.045 0.0024 0.0019
Data presented as mean ± standard deviation; CI ¼ confidence interval, LL ¼ lesion length, TPV ¼ total plaque volume, CPV ¼ calcified plaque volume, NCPV ¼ noncalcified plaque volume, RI ¼ remodeling index.
Several limitations of this study need to be taken into consideration. We performed a retrospective investigation which is therefore subject to limitations inherent in this study design. A relatively small number of patients were included; therefore, larger, possibly prospective studies will be necessary to validate our findings. Our results on multivariate analysis may be underpowered by the limited number of observations per variable included.28 Therefore, this data generated in this study should only be hypothesis generating. We only evaluated one lesion in each patient, where stent placement was performed. Furthermore, we excluded bypass grafts, bifurcation lesions, left main revascularization, and ostial lesions, which are prone to higher restenosis rates.9 Additionally, we did not evaluate the influence of procedural results
Table 6 Diagnostic performance of coronary CTA-derived lesion-related markers for the prediction of ISR according to QCA. Parameter
LL > 21.1 (95%CI)
RI > 1.03 (95%CI)
NCPV > 131.7 (95%CI)
Sensitivity
74% (55e88%) 74% (59e86%) 66% (48e81%) 81% (66e91%)
71% (52e86%) 78% (64e89%) 69% (50e84%) 80% (65%e90%
65% (45e81%) 80% (65e90%) 69% (49e85%) 77% (62e88%)
Specificity PPV NPV
CI ¼ confidence interval, LL ¼ lesion length, NCPV ¼ non-calcified plaque volume, RI ¼ remodeling index. Data presented as percentages (%).
(e.g.: stent deployment, stent malapposition, stent fracture) and clinical parameters which are known to affect ISR.29 We did not correlate our findings on coronary CTA with intracoronary imaging techniques, as IVUS-guided stent placement had been performed in only 6 of the 21 cases with ISR and optical coherence tomography is not routinely performed at our institution. In conclusion this study demonstrates that coronary CTAderived NCPV, LL, and RI as quantitative markers of target lesion anatomy and CAD disease activity portend predictive value for ISR with incremental value of combination of these parameters. Obtaining these markers prior to PCI may guide the selection of an
Fig. 3. Diagnostic performance of Coronary CTA-derived lesion-related markers for the prediction of ISR.
C. Tesche et al. / Journal of Cardiovascular Computed Tomography 10 (2016) 377e383
appropriate revascularization strategy, e.g., selection of stents that cover the entire atheroma or DES to treat lesions with high disease activity by coronary CTA. Disclosures Dr. Schoepf is a consultant for and/or receives research support from Astellas, Bayer, Bracco, GE, Guerbet, Medrad, and Siemens. Drs. De Cecco and Varga-Szemes are consultants for Guerbet. All other authors have no conflicts of interest to disclose. Christian Tesche is an exchange visiting scholar supported by a grant from the Fulbright Visiting Scholar Program of the U.S. Department of State, Bureau of Educational and Cultural Affairs (ECA). References 1. Smith Jr SC, Dove JT, Jacobs AK, et al. ACC/AHA guidelines for percutaneous coronary intervention (revision of the 1993 PTCA guidelines)-executive summary: a report of the American College of Cardiology/American Heart Association task force on practice guidelines (Committee to revise the 1993 guidelines for percutaneous transluminal coronary angioplasty) endorsed by the Society for Cardiac Angiography and Interventions. Circulation. 2001;103: 3019e3041. 2. Palmerini T, Biondi-Zoccai G, Della Riva D, et al. Clinical outcomes with bioabsorbable polymer- versus durable polymer-based drug-eluting and baremetal stents: evidence from a comprehensive network meta-analysis. J Am Coll Cardiol. 2014;63:299e307. 3. Vink MA, Van Nooijen FC, Laarman GJ, et al. Patency of paclitaxel-eluting versus bare metal stents long term after implantation in acute ST-segment elevation myocardial infarction. Am J Cardiol. 2011;108:1214e1219. 4. Cohen DJ, Van Hout B, Serruys PW, et al. Quality of life after PCI with drugeluting stents or coronary-artery bypass surgery. N Engl J Med. 2011;364: 1016e1026. 5. Stenestrand U, James SK, Lindback J, et al. Safety and efficacy of drug-eluting vs. bare metal stents in patients with diabetes mellitus: long-term follow-up in the Swedish Coronary Angiography and Angioplasty Registry (SCAAR). Eur Heart J. 2010;31:177e186. 6. Holmes Jr DR, Kereiakes DJ, Garg S, et al. Stent thrombosis. J Am Coll Cardiol. 2010;56:1357e1365. 7. Mehran R, Dangas G, Abizaid AS, et al. Angiographic patterns of in-stent restenosis: classification and implications for long-term outcome. Circulation. 1999;100:1872e1878. 8. Farb A, Weber DK, Kolodgie FD, Burke AP, Virmani R. Morphological predictors of restenosis after coronary stenting in humans. Circulation. 2002;105: 2974e2980. 9. Hoffmann R, Mintz GS, Mehran R, et al. Intravascular ultrasound predictors of angiographic restenosis in lesions treated with Palmaz-Schatz stents. J Am Coll Cardiol. 1998;31:43e49. 10. Sahara M, Kirigaya H, Oikawa Y, et al. Arterial remodeling patterns before intervention predict diffuse in-stent restenosis: an intravascular ultrasound study. J Am Coll Cardiol. 2003;42:1731e1738. 11. Bittencourt MS, Hulten E, Ghoshhajra B, et al. Prognostic value of nonobstructive and obstructive coronary artery disease detected by coronary computed tomography angiography to identify cardiovascular events. Circ
383
Cardiovasc Imaging. 2014;7:282e291. 12. Bauer RW, Thilo C, Chiaramida SA, Vogl TJ, Costello P, Schoepf UJ. Noncalcified atherosclerotic plaque burden at coronary CT angiography: a better predictor of ischemia at stress myocardial perfusion imaging than calcium score and stenosis severity. AJR Am J Roentgenol. 2009;193:410e418. 13. Wang R, Renker M, Schoepf UJ, et al. Diagnostic value of quantitative stenosis predictors with coronary CT angiography compared to invasive fractional flow reserve. Eur J Radiol. 2015;84:1509e1515. 14. Wang R, Baumann S, Schoepf UJ, et al. Comparison of quantitative stenosis characteristics at routine coronary computed tomography angiography with invasive fractional flow reserve for assessing lesion-specific ischemia. J Cardiovasc Comput Tomogr. 2015;9:546e552. 15. Budoff MJ, Bloom SA, Chow BJ, Chandler Jr AB, Cole JH. Coronary computed tomography angiography predicts subsequent cardiac outcome events: results of the Visipaque CCTA registry study. Coron Artery Dis. 2015;26:301e307. 16. Leipsic J, Abbara S, Achenbach S, et al. SCCT guidelines for the interpretation and reporting of coronary CT angiography: a report of the Society of Cardiovascular Computed Tomography Guidelines Committee. J Cardiovasc Comput Tomogr. 2014;8:342e358. 17. Jacobs JE, Boxt LM, Desjardins B, et al. ACR practice guideline for the performance and interpretation of cardiac computed tomography (CT). J Am Coll Radiol. 2006;3:677e685. 18. Voros S, Rinehart S, Qian Z, et al. Coronary atherosclerosis imaging by coronary CT angiography: current status, correlation with intravascular interrogation and meta-analysis. JACC Cardiovasc Imaging. 2011;4:537e548. 19. Achenbach S, Ropers D, Hoffmann U, et al. Assessment of coronary remodeling in stenotic and nonstenotic coronary atherosclerotic lesions by multidetector spiral computed tomography. J Am Coll Cardiol. 2004;43:842e847. 20. DeLong ER, DeLong DM, Clarke-Pearson DL. Comparing the areas under two or more correlated receiver operating characteristic curves: a nonparametric approach. Biometrics. 1988;44:837e845. 21. Arakawa K, Ishibashi-Ueda H, Hao H, Ikeda Y, Kawamura A. Plaque tissue components obtained from de novo lesions may predict restenosis after directional coronary atherectomy. Ann Vasc Dis. 2010;3:52e59. 22. Prati F, Di Mario C, Moussa I, et al. In-stent neointimal proliferation correlates with the amount of residual plaque burden outside the stent: an intravascular ultrasound study. Circulation. 1999;99:1011e1014. 23. Rathore S, Terashima M, Katoh O, et al. Predictors of angiographic restenosis after drug eluting stents in the coronary arteries: contemporary practice in real world patients. EuroIntervention. 2009;5:349e354. 24. Kini A, Marmur JD, Dangas G, Choudhary S, Sharma SK. Angiographic patterns of in-stent restenosis and implications on subsequent revascularization. Catheter Cardiovasc Interv. 2000;49:23e29. 25. Goldberg SL, Loussararian A, De Gregorio J, Di Mario C, Albiero R, Colombo A. Predictors of diffuse and aggressive intra-stent restenosis. J Am Coll Cardiol. 2001;37:1019e1025. 26. van Velzen JE, de Graaf MA, Ciarka A, et al. Non-invasive assessment of atherosclerotic coronary lesion length using multidetector computed tomography angiography: comparison to quantitative coronary angiography. Int J Cardiovasc Imaging. 2012;28:2065e2071. 27. Chang SH, Chen CC, Hsieh MJ, Wang CY, Lee CH, Hsieh IC. Lesion length impacts long term outcomes of drug-eluting stents and bare metal stents differently. PLoS One. 2013;8:e53207. 28. Kunce JT. Random Variables and Correlational Overkill. Educational and Psychological Measurement. 29. Singh M, Gersh BJ, McClelland RL, et al. Clinical and angiographic predictors of restenosis after percutaneous coronary intervention: insights from the Prevention of Restenosis with Tranilast and its Outcomes (PRESTO) trial. Circulation. 2004;109:2727e2731.