Author + information
- Received December 30, 2014
- Revision received March 19, 2015
- Accepted March 26, 2015
- Published online August 17, 2015.
- Jin-Ho Choi, MD, PhD∗,†∗ (, )
- Eun-Kyoung Kim, MD∗,
- Sung Mok Kim, MD, PhD‡,
- Hyungyoon Kim, MD∗,
- Young Bin Song, MD, PhD∗,
- Joo-Yong Hahn, MD, PhD∗,
- Seung Hyuk Choi, MD, PhD∗,
- Hyeon-Cheol Gwon, MD, PhD∗,
- Sang-Hoon Lee, MD, PhD∗,
- Yeon Hyeon Choe, MD, PhD‡ and
- Jae K. Oh, MD∗,§
- ∗Department of Medicine, Samsung Medical Center, Sungkyunkwan University School of Medicine, Seoul, Republic of Korea
- †Department of Emergency Medicine, Samsung Medical Center, Sungkyunkwan University School of Medicine, Seoul, Republic of Korea
- ‡Department of Radiology, Cardiovascular Imaging Center, Samsung Medical Center, Sungkyunkwan University School of Medicine, Seoul, Republic of Korea
- §Department of Internal Medicine, Mayo Clinic College of Medicine, Rochester, Minnesota
- ↵∗Reprint requests and correspondence:
Dr. Jin-Ho Choi, Department of Emergency Medicine, Samsung Medical Center, Sungkyunkwan University School of Medicine, 81 Irwon-Ro Gangnam-gu, Seoul 135-710, Republic of Korea.
Objectives The aim of this study was to investigate whether noninvasive discrimination of chronic total occlusion (CTO), a complete interruption of coronary artery flow, and subtotal occlusion (STO), a functional total occlusion, is feasible using coronary computed tomography angiography (CTA).
Background CTO and STO may be different in pathophysiology and clinical treatment strategy.
Methods We included 486 consecutive patients (median age 63 years, 82% male) who showed a total of 553 completely occluded coronary arteries in coronary CTA. The length of occlusion, side branches, shape of proximal stump, and collateral vessels were measured as anatomical findings. Transluminal attenuation gradient, which reflects intraluminal contrast kinetics and functional extent of collateral flow, was measured as a physiological surrogate. All patients were followed by invasive coronary angiography.
Results Coronary arteries with CTO showed longer occlusion length (cutoff ≥15 mm), higher distal transluminal attenuation gradient (cutoff ≥−0.9 Hounsfield units [HU]/10 mm), more frequent side branches, blunted stump, cross-sectional calcification ≥50%, and collateral vessels compared with arteries with STO (p < 0.001, all). The combination of these findings could distinguish CTO from STO (c-statistics = 0.88 [95% confidence interval: 0.94 to 0.90], sensitivity 83%, specificity 77%, positive predictive value 55%, negative predictive value 93%; p < 0.001). Percutaneous coronary intervention (PCI) was attempted in 342 arteries and was successful in 279 arteries (82%). The computed tomography findings could predict the unsuccessful PCI (c-statistics = 0.70 [95% confidence interval: 0.65 to 0.75], sensitivity 63%, specificity 73%, positive predictive value 91%, negative predictive value 31%; p < 0.001).
Conclusions Noninvasive coronary CTA could discern CTO from STO, and also could predict the success of attempted PCI.
Chronic total occlusion (CTO), a complete interruption of coronary arterial flow caused by plaque obstructing arterial lumen, is not uncommon and is found in 18% to 31% of invasive coronary angiography (CAG) cases (1,2). CTO is known to be frequently accompanied by microchannels smaller than the spatial resolution of angiography, which corresponds to less than complete luminal occlusion histologically (3). Therefore, CTO and subtotal occlusion (STO), which is a “functional” total occlusion or a slow contrast penetration through the occluded segment, are not distinguished strictly in clinical practice (4–6).
Discrimination of CTO from STO before cardiac catheterization is of clinical value because percutaneous coronary intervention (PCI) is attempted less frequently for CTO lesions and has a higher rate of unsuccessful procedures and late restenosis compared with non-CTO lesions (7,8). However, noninvasive discrimination of CTO from STO remains a challenge, because both show complete interruption of contrast-enhanced arterial lumen in anatomical imaging tests and myocardial ischemia in functional imaging tests (9–11).
The pathogenesis of CTO is presumed to be thrombotic occlusion in the context of nonfatal myocardial infarction (MI) followed by progression of organized thrombus filling up to the ostium of side branches. Organized thrombus is replaced by collagen-rich fibrous tissue or calcification. Collateral vessel develops to supply myocardium subtended by occluded vessels (12–14). The angiographic characteristics of CTO would be summarized by the long totally occluded segment between adjacent side branches, blunt proximal stump, and collateral vessels flowing into the distal segment or side branches (Figure 1).
We reasoned that these anatomical and physiological findings can be identified by coronary computed tomography angiography (CTA), a noninvasive modality that enables investigation of both the arterial lumen and the obstructive plaque. It would be helpful to the decision of the revascularization strategy and the prediction of PCI procedural success. We investigated anatomical and physiological findings in the coronary CTA of patients with totally occluded coronary arteries and compared it with the invasive CAG and the result of PCI.
From June 2006 to November 2013, we consecutively screened 578 patients who showed at least 1 totally occluded coronary artery by clinically indicated coronary CTA and were validated by subsequent diagnostic CAG within 12 weeks (median 14 days). Patients with recent MI within 90 days by the computed tomography acquisition time were not enrolled, nor were those with uncompensated heart failure, left main disease, creatinine ≥2.0 mg/dl, or estimated glomerular filtration rate ≤30 ml/min. Patients with an occlusion in the small branch, vasospastic angina, and poor image quality were excluded. Patients who underwent bypass surgery or PCI for occluded arteries were also excluded due to the potential confounding on the coronary flow physiology. The remaining 486 patients comprised the study cohort (Figure 2). Our prior studies enrolled patients from January 2007 to April 2009, and 208 patients were included in both prior studies and the current study (15,16). The institutional review board committee approved the study protocol.
Clinical history and risk factors were investigated by research nurses. The prevalence of prior MI in patients with total occlusion is known to be underestimated. Therefore, prior MI was defined by the patient’s history of ischemic symptoms consistent with MI, Q waves involving ≥2 contiguous electrocardiography (ECG) leads, or evidence of left ventricular (LV) wall motion abnormality consistent with the territory of obstructed coronary artery (12).
Acquisition of coronary CTA
A 64-slice (Aquilion 64, Toshiba Medical Systems, Tokyo, Japan) or 128-slice scanner (SOMATOM Definition, Siemens Medical Solution, Forchheim, Germany) was used in 236 and 250 patients, respectively. Oral metoprolol and nitroglycerin were used before image acquisition for optimal heart rate and vasodilation. Contrast injection consisted of a bolus of 80 ml of nonionic contrast medium (Iomeron 400 mg iodine/ml, Bracco, Milan, Italy) at a flow rate of 4 ml/s followed by an injection of 30 ml of 30:70 contrast medium-saline mixtures and an injection of saline 30 ml. The scan mode of the 64-slice scanner was a retrospective gated helical mode with the following parameters: 120 kV tube voltage, 400 mA tube current, 350 ms gantry rotation time, 0.22 to 0.25 pitch, and 64 × 0.5-mm detector collimation. For the 128-slice scanner, the retrospective ECG-gated helical technique was used with the following parameters: 120 kV tube voltage, 400 mA tube current, 280 ms gantry rotation time, 0.22 to 0.24 pitch, and 2 × 64 × 0.6-mm detector collimation resulting in 2 × 128 × 0.6-mm sections. The image dataset was reconstructed by 0.5- or 0.6-mm slices. The axial scan time was 6 to 8 s in both scanners.
Assessment of totally occluded segment and side branches
Cardiovascular radiologists interpreted coronary CTA using a dedicated workstation (iNtuition, Terarecon, Foster City, California). Totally occluded vessel segment was defined by complete absence of luminal enhancement. Axial length of artery and occluded segment was measured in curved multiplanar reformatted view to avoid shortening caused by angulation or viewing angle. The luminal size, vessel area in the midst of occluded segment, remodeling index, and semiquantitative extent of calcification were assessed in cross-sectional view. Side branches adjacent to the total occlusion was defined when both proximal and distal side branches having diameter ≥1.5 mm were identified adjacent to the totally occluded segment. In case of ambiguity, another radiologist or imaging specialist was invited to reach an agreement.
Transluminal attenuation gradient
Transluminal attenuation gradient (TAG)all and TAGdistal, which reflect the kinetics of intraluminal flow and collateral circulation, respectively, were assessed as described previously (15,17–20). TAG (HU/10 mm) was defined by the linear regression coefficient between intraluminal radiological attenuations and vessel length from the ostium to distal vessel (TAGall) or distal to occlusion (TAGdistal).
Collateral vessel and LV mass index
The presence of collateral vessel was determined by complete vascular connection between donor and recipient coronary arteries, which was defined by continuous connection in all curved multiplanar reformatted images rotated by 1° each, as shown in Online Videos 1, 2, 3, 4, 5, 6, 7, 8, and 9. The LV mass was assessed (Vitrea, Toshiba Medical Systems, Otawara-shi, Japan) and was adjusted with body surface area (m2).
Invasive CAG and PCI
CAG was performed using standard techniques with intracoronary nitroglycerin, and was assessed by interventional cardiologists blinded to the other imaging data. CTO was defined by completely interrupted arterial lumen with absence of antegrade flow through the lesion (TIMI [Thrombolysis In Myocardial Infarction] flow grade 0) (5). STO was defined by diameter stenosis ≥95% and significantly decreased antegrade flow (TIMI grade 1 to 2) (6,21). Angiographic success of PCI was defined by achievement of a minimum diameter stenosis <50% and TIMI grade 3 flow.
The diameter and angiographic flow of collateral vessels was semiquantitatively assessed using Collateral Connection grade (0 to 2) and Rentrop grade (0 to 3), which represent the size and functional extent of collateral vessels, respectively (22,23). Quantitative coronary angiography was performed by technicians who were blinded to the other imaging data and was supervised by an interventional cardiologist. The angiographic length of the occluded segment was measured between the proximal and distal segment of occlusion or visual stenosis >95%.
Analysis was done on a per-vessel basis unless indicated otherwise. Data were not normally distributed, and nonparametric statistics were applied. Continuous variables are shown as median with first and third quartiles. Continuous and categorical variables were compared using the Mann-Whitney U or Fisher exact test. Predictive performance of individual models was compared using receiver-operating characteristic analysis with DeLong’s method. The inter-rater agreement and reproducibility of the morphological assessment were validated in 20 randomly selected vessels using Cohen’s weighted kappa. The multivariate logistic model included optimal cut-offs of independent predictors (p < 0.05) determined by Youden’s J statistics, and was validated by bootstrap with 10,000 iterations (24). SPSS version 19.0 (IBM, Armonk, New York) and R version 3.1.0 (R Foundation for Statistical Computing, Vienna, Austria) with boot package was used. A 2-tailed p < 0.05 was considered to be statistically significant.
There was no difference in clinical risk factors between patients with CTO and patients with STO by CAG (p = NS, all). Patients with CTO showed higher frequency of Q-wave on ECG, lower LV ejection fraction, and higher LV mass index compared with patients with STO (p < 0.05, all). Silent ischemia was more common in patients with CTO compared with patients with STO (23% vs. 13%; p = 0.020). Prevalence of multivessel disease was not different between the 2 groups (p = 0.18) (Table 1).
Reproducibility and agreement of measurements
For the presence of side branch and collateral visualization, the weighted Kappa was 0.90 (95% confidence interval [CI]: 0.70 to 1.00) in both intrarater and inter-rater agreement. For the Rentrop or Collateral connection scores, the weighted Kappa was 0.83 (95% CI: 0.64 to 1.00) in intrarater agreement and 0.70 (95% CI: 0.45 to 0.94) in inter-rater agreement (Table 2).
There was more right coronary artery occlusion in CTO compared with STO (46% vs. 32%; p = 0.003). CTO showed longer occlusion length, large proximal reference luminal area, and larger vessel area in occluded segment compared with STO (p < 0.005, all). CTO also showed more side branches, blunt stump, calcifications, and noninvasively visualized collateral vessels compared with STO (p < 0.005, all). Both TAGall and TAGdistal were higher in CTO compared with STO (p < 0.01, all) (Table 2). Representative cases are shown in Figure 3.
Invasive coronary angiography
CTO showed longer occlusion length and more frequent side branches compared with STO (p < 0.001, all). Collateral vessel was found in 100% of CTO and in 68% of STO. Also, collaterals supplying arteries with CTO were larger, had higher flow, and had more retrograde flow compared with collaterals supplying arteries with STO (p < 0.001, all).
PCI was attempted less frequently (57% vs. 76%) with much lower procedural success rate (75% vs. 95%) in CTO compared with STO (p < 0.001), which was mostly driven by inability to perform wire passage. Coronary perforation had developed only in the PCI for CTO (Table 3).
Prediction of the presence of CTO and the procedural success of PCI
The predictive performance of each coronary CTA finding for the presence of CTO by CAG was moderate (c-statistics = 0.55 to 0.74) (Table 4). A predictive model using the combination of optimal cutoff values of coronary CTA findings, including occlusion length ≥15 mm, side branches, blunt stump, visualized collateral vessel, cross-sectional calcification ≥50%, and TAGdistal ≥−0.9 HU/10 mm, showed much higher discriminative performance compared with the other predictive models on the basis of single CTA findings: c-statistics = 0.88 (95% CI: 0.85 to 0.90) sensitivity 85% (95% CI: 78% to 90%); specificity 75% (95% CI: 71% to 80%); positive predictive value (PPV), 55% (95% CI: 48% to 61%); negative predictive value (NPV), 93% (95% CI: 90% to 96%); p < 0.001. This model could also predict the unsuccessful PCI procedural result better than single CTA findings: c-statistics = 0.70 (95% CI: 0.65 to 0.75); sensitivity 63% (95% CI: 57% to 69%); specificity 83% (95% CI: 60% to 83%); PPV 91% (95% CI: 86% to 95%); NPV 31% (95% CI: 24% to 39%); p < 0.005 (Figure 4). The c-statistics of the Bootstrap-refined model were 0.88 (95% CI: 0.85 to 0.91) and 0.70 (95% CI: 0.63 to 0.77), respectively.
In our study, comprehensive assessment of anatomical and physiological findings in coronary CTA could discern CTO from STO. These findings could also predict the result of PCI.
The success rate of PCI is >95% for non-CTO, whereas it is still 70% to 80% for CTO despite the introduction of novel techniques and new devices (25). The most common cause of PCI failure is the inability to pass a guidewire through the occluded segment. Hence, the most reasonable approach might be the selection of appropriate cases in which procedural success is highly expected (5). Besides unsuccessful PCI, intraprocedural complications and the need for specialized devices are also more common in CTO compared with STO. Therefore, noninvasive discrimination of CTO from STO may be useful for deciding upon a revascularization strategy or estimating the procedural difficulty.
Two prior studies attempted computed tomography–based discrimination of CTO from STO (9,21). In both studies, occlusion length was highly specific, but the specificity of reverse attenuation gradient was not consistent. Our prior study has shown that reverse attenuation gradient corresponds to TAGdistal >0, depends on the direction and extent of collateral flow, and is moderately specific (15). Our results were derived from a much larger study of 553 vessels, and are determined on the basis of both anatomical and physiological findings. Importantly, our results were validated with the attempt of a guidewire passage through the occlusion, which may be the best method of discriminating CTO from STO. CTO with minimal contrast penetration through the occluded segment has been frequently referred as “functional” total occlusion and has been regarded as CTO (4–6). Based on our results, the current definition of CTO might be redefined by CTA.
Significant coronary calcification has been shown to be a predictor of successful CTO PCI. In our study, the calcification did not affect the procedural result, which is similar to our prior study (16). This finding is partially explained by the poor performance of CTA for discrimination of hard fibrous tissue from calcified tissue, which shows distinct HU values but blocks guidewire advancement, or for detection of small calcification (13,26).
The following inherent limitations should be recognized before translating current results to the other scenarios. The results were derived from single-center datasets. Our result may be biased because of case selection, operator expertise, and CTA imaging techniques. We did not enroll patients who showed CTO but did not undergo CAG. PCI was not attempted in all patients, which may be linked to selection bias or measures of outcome bias. Interpretation of small branches, calcification, interruption of vessel lumen, or distal intravascular attenuation may be affected by the spatial resolution of CTA and the extent of intraluminal opacification (27). Although TAG has been successfully studied from CTA with multiple cardiac cycles, the lack of temporal uniformity might affect the result of TAG (17). CTA showed high NPV (93%) but moderate PPV (55%) for the presence of CTO, which suggests that CTA may overestimate the complexity of occluded lesion. Our study is based on the presumed pathogenesis of CTO, a progression of organized thrombus and collagen tissues filling up to the ostium of side branches accompanied by collateral vessels supplying subtended myocardium, but does not explain or provide the pathogenesis of STO, which has not been investigated sufficiently (28).
Noninvasive coronary CTA could discern CTO from STO and could also predict unsuccessful PCI. Our results would be helpful for the evaluation of patients with totally occluded coronary arteries and guidance of revascularization strategy.
WHAT IS KNOWN? Discrimination of CTO from STO before cardiac catheterization is of clinical value because PCI is attempted less frequently for CTO, with a higher rate of unsuccessful procedures and late restenosis compared with non-CTO. However, noninvasive discrimination of CTO from STO remains a challenge.
WHAT IS NEW? This study showed that comprehensive assessment of anatomical and physiological findings in coronary CTA on the basis of the presumed pathogenesis of CTO, a progression of organized thrombus and collagen tissues filling up to the ostium of side branches accompanied by collateral vessels supplying subtended myocardium, could discern CTO from STO and could also predict the result of PCI.
WHAT IS NEXT? Noninvasive discrimination of CTO from STO before invasive catheterization may be useful for determining a revascularization strategy for patients with totally occluded coronary arteries.
For supplemental videos and their legends, please see the online version of this article.
This study was supported by a Samsung Biomedical Research Institute grant (PHO1140251), and by the Heart Vascular and Stroke Institute, Samsung Medical Center. The authors have reported that they have no relationships relevant to the contents of this paper to disclose. Drs. J.-H. Choi and E.-K. Kim contributed equally to this work.
- Abbreviations and Acronyms
- coronary angiography
- computed tomography angiography
- chronic total occlusion
- percutaneous coronary intervention
- subtotal occlusion
- transluminal attenuation gradient
- Received December 30, 2014.
- Revision received March 19, 2015.
- Accepted March 26, 2015.
- 2015 American College of Cardiology Foundation
- Fefer P.,
- Knudtson M.L.,
- Cheema A.N.,
- et al.
- Srivatsa S.S.,
- Edwards W.D.,
- Boos C.M.,
- et al.
- Stone G.W.,
- Kandzari D.E.,
- Mehran R.,
- et al.
- Prasad A.,
- Rihal C.S.,
- Lennon R.J.,
- Wiste H.J.,
- Singh M.,
- Holmes D.R. Jr..
- Choi J.H.,
- Chang S.A.,
- Choi J.O.,
- et al.
- Sumitsuji S.,
- Inoue K.,
- Ochiai M.,
- Tsuchikane E.,
- Ikeno F.
- Choi J.H.,
- Kim E.K.,
- Kim S.M.,
- et al.
- Wong D.T.,
- Ko B.S.,
- Cameron J.D.,
- et al.
- Choi J.H.,
- Koo B.K.,
- Yoon Y.E.,
- et al.
- Yoon Y.E.,
- Choi J.H.,
- Kim J.H.,
- et al.
- Werner G.S.,
- Ferrari M.,
- Heinke S.,
- et al.
- Rentrop K.P.,
- Cohen M.,
- Blanke H.,
- Phillips R.A.
- Patel V.G.,
- Brayton K.M.,
- Tamayo A.,
- et al.
- Kang S.J.,
- Nakano M.,
- Virmani R.,
- et al.