Author + information
- Received November 4, 2014
- Revision received March 23, 2015
- Accepted April 8, 2015
- Published online July 1, 2015.
- Chun Luo, MD∗,†,
- Meiping Huang, MD†∗ (, )
- Jinglei Li, MD, PhD∗,
- Changhong Liang, MD, PhD∗,
- Qun Zhang, MD†,
- Hui Liu, MD, PhD∗,
- Zaiyi Liu, MD, PhD∗,
- Yanji Qu, MD, PhD‡,
- Jun Jiang, MD§ and
- Jian Zhuang, MD, PhD‖
- ∗Department of Radiology, Guangdong General Hospital, Guangdong Academy of Medical Sciences, Guangzhou, China
- †Department of Catheterization Lab, Guangdong Cardiovascular Institute, Guangdong Provincial Key Laboratory of Coronary Heart Disease Prevention, Guangdong General Hospital, Guangdong Academy of Medical Science, Guangzhou, China
- ‡Epidemiology Division, Department of Cardiac Surgery, Guangdong Cardiovascular Institute, Guangdong General Hospital, Guangdong Academy of Medical Science, Guangzhou, China
- §Department of Radiology, Shenzhen Second People’s Hospital, Shenzhen, China
- ‖Department of Cardiac Surgery, Guangdong Cardiovascular Institute, Guangdong General Hospital, Guangdong Academy of Medical Science, Guangzhou, China
- ↵∗Reprint requests and correspondence:
Dr. Meiping Huang, Department of Catheterization Lab, Guangdong Cardiovascular Institute, Guangdong Provincial Key Laboratory of Coronary Heart Disease Prevention, Guangdong General Hospital, Guangdong Academy of Medical Science, 96 Dongchuan Road, Guangzhou 510100, China.
Objectives This study aimed to identify significant lesion features of chronic total occlusions (CTOs) that predict failure of antegrade (A) percutaneous coronary intervention (PCI) using pre-procedure coronary computed tomography angiography (CTA) combined with conventional coronary angiography (CCA).
Background The current predictors of successful A-PCI in the setting of CTOs are uncertain. Such knowledge might prompt early performance of a retrograde (R)-PCI approach if predictors of A-PCI failure are present.
Methods Consecutive patients confirmed to have at least 1 CTO of native coronary arteries underwent coronary CTA- and CCA-guided PCI in which computed tomography and fluoroscopic images were placed side by side before or during PCI.
Results The study included 103 patients with 108 CTOs; 80 lesions were successfully treated with A-PCI and 28 lesions failed this approach, for an A-PCI success rate of 74%. A total of 15 of 28 failed cases underwent attempted R-PCI. Only 1 case also failed R-PCI; thus, the total PCI success rate was 87%. By multivariable analysis, the factors significantly predictive of failed A-PCI included negative remodeling (odds ratio [OR]: 137.82) and lesion length >31.89 mm on coronary CTA (OR: 7.04), and ostial or bifurcation lesions on CCA (OR: 8.02). R-PCI was successful in 14 of 15 patients (93.3%), in whom good appearance of the occluded distal segment and well-developed collateral vessels were present.
Conclusions Morphologic predictors of failed A-PCI on the basis of pre-procedure coronary CTA and CCA imaging may be identified, which may assist in determining which patients with CTO lesions would benefit from an early R-PCI strategy.
Chronic total occlusion (CTO) of a coronary artery is defined as an obstruction of a native coronary artery exhibiting no luminal continuity for at least 3 months (1–5). CTO occurs in approximately 15% to 23% of patients who undergo conventional coronary angiography (CCA), for whom revascularization remains challenging, with success rates of only 55% to 80% (4–8). CTO recanalization is followed by transient impairment of vasomotor function at distal coronary segments, which could lead to underestimation of the required stent size (9). However, successful recanalization for CTO is associated with improved patient outcomes; therefore, percutaneous coronary intervention (PCI) remains a valid therapeutic option (10,11). Repeated attempts may be required for patients with chronic CTO (12,13), which leads to cicatrix formation caused by iatrogenic vascular injury and repair in association with inflammation and fibrosis (8). Thus, patients experiencing failed PCIs have higher complication rates and lower success rates for further PCIs (8). Therefore, achieving a successful PCI with minimal attempts is optimal for CTOs.
Despite the lack of visualization of the occluded CTO segments, CCA remains the reference standard for evaluating coronary anatomy (2,14,15). Recently, the superiority of coronary computed tomography angiography (CTA) for directly observing morphological characteristics of occluded lesions, especially calcification, was highlighted (14,15). Rolf et al. (3) found that pre-procedural coronary CTA was associated with a higher PCI success rate than lack of pre-procedural coronary CTA. Therefore, studies have focused on morphological characteristics observed on CCA and/or coronary CTA that could predict successful PCI, such as blunt stump, tortuous course, and calcification (6,11). Unfortunately, predictors of successful PCI on either coronary CTA or CCA were not consistent in those studies (6,11,15).
The purpose of the present study was to investigate whether morphological characteristics on pre-procedural coronary CTA and CCA may serve as predictors of failed antegrade (A)-PCI, which might prompt a switch to an early retrograde approach for appropriate patients, thus enabling successful revascularization.
All patients who underwent PCI of a CTO (not including in-stent occlusion) at our hospital between January 2012 and December 2013 were screened for enrollment in this retrospective study. Patients who underwent coronary CTA prior to their staged coronary CTA- and CCA-guided PCI were included. Exclusion criteria included allergy to contrast media, previous coronary artery bypass graft (CABG) surgery, and a history of acute myocardial infarction within 1 month.
Indications for PCI included angina or evidence of myocardial ischemia in the CTO region of the artery, the presence of viable myocardium supplied by the occluded artery as revealed by stress echocardiography and/or single-photon emission computed tomography (CT), lack of cardiac and renal insufficiency, and moderate-to-high (60%) confidence of successful PCI without major adverse cardiovascular events.
Indications for pre–coronary CTA included lack of allergy to contrast media, no acute or chronic renal insufficiency (serum creatinine >1.5 g/dl [132.6 mmol/l]), no atrial fibrillation or other heart rhythm irregularity, and no clinical history of uncontrolled hyperthyroidism or multiple myeloma.
The interval between CCA and/or staged PCI and coronary CTA ranged from 24 h to 1 month. All inpatients received intravenous hydration with 1 ml/kg/h saline the day before and after coronary CTA and PCI. The duration of CTO occlusion was estimated on the basis of evidence from prior coronary angiography or clinical symptoms. The CTOs were divided into 2 groups: successful A-PCI and failed A-PCI. Successful PCI was defined as the attainment of a residual diameter stenosis <20% and Thrombolysis In Myocardial Infarction flow grade 3 without major adverse cardiovascular events during hospitalization. All patients signed their informed consent for each examination and treatment; the study protocol was approved by the Guangdong General Hospital ethics committee.
Coronary CTA protocol
A wide detector 256-slice CT scanner (Brilliance iCT, Philips Healthcare, Cleveland, Ohio) was used for coronary CTA. A metoprolol tartrate tablet (25 to 50 mg) was sublingually administered to patients with a heart rate (HR) >70 beats/min. The antecubital vein was injected with a bolus of contrast medium (Ultravist 370, Bayer Schering Pharma, Berlin, Germany) 50 to 70 ml (depending on the patient's weight) followed by 30 ml of saline solution at a rate of 5 ml/s. A real-time bolus tracking technique (BolusPro, Philips Healthcare) was used to synchronize contrast medium injection and scanning; the region of interest was placed on the ascending aorta root, and image acquisition was initiated 5 s after signal density reached a threshold of 180 Hounsfield units. Prospective electrocardiographically-gated coronary CTA was performed using the following parameters: tube voltage 120 kV; tube current 120 mAs; collimation 128 × 0.625 mm; rotation time 270 ms; pitch 0.18; and field of view 250 mm. The CT scan was performed from 1 cm below the tracheal bifurcation to 2 cm below the diaphragm. Electrocardiographically-gated datasets were reconstructed at 70%, 75%, and 80% of the cardiac cycle if the HR was <70 beats/min; additional datasets were reconstructed at 40%, 45%, and 50% of the cardiac cycle if the HR was ≥70 beats/min.
The acquired datasets were post-processed and analyzed using a Philips workstation (Extended Brilliance Workspace, Philips Healthcare). The post-processing technique included volume rendering, maximum intensity projection, multiplanar reformation, and curved planar reconstruction.
A digital CCA system (AlluraXper FD10, Philips Medical System, Best, the Netherlands) was used to obtain an average of 5 and 2 cineangiograms for the left and right coronary arteries, respectively, at 15 frames/s. Some patients simultaneously underwent bilateral coronary artery angiography. The patients were placed in a supine position, and coronary artery angiography was performed through the radial or femoral artery. The gantry angles of the projections were selected at the cardiologist’s discretion.
Assessment of parameters
The following characteristics were evaluated on coronary CTA and CCA: lesion site (left anterior descending artery/right coronary artery/left circumflex artery/left main coronary artery), the calcification degree of the lesion (none/slight/severe: calcified area was 0%, <50%, or ≥50% on coronary CTA or present/absent on CCA), stump morphology (sharp or blunt), lesion length (mm), ostial or bifurcation lesions, tortuous course (≥45° or <45°); remodeling type, the presence of microchannels (linear or dot intrathrombus enhancement on coronary CTA and slender blood flow on CCA), bridging vessels (present or absent), the appearance of the occluded distal segment (good/bad: vessels filled well with contrast medium and without significant stenosis ≥50%/poor visualization or stenosis ≥50%), and the score of the collateral vessels (good/bad: present/absent on coronary CTA or score of 3/score of 0 to 2 on CCA using the classification of Rentrop ). The diameter of the occluded vessel (DO) and the adjacent normal vessel (DN) were measured by coronary CTA to determine the type of vascular remodeling (DO/DN ≥1 or <1 representing positive or negative remodeling, respectively) (Figure 1) (11). The CTO length on coronary CTA was measured from the proximal margin to the distal margin of the total occluded segment without enhancement (compared to the plain scan). The CTO origin could be estimated according to the angiogram if either the proximal or distal end of the occlusion had 360° calcification. All characteristics were independently evaluated by 2 experienced observers (M.H. and Q.Z.) who were blinded to the examination findings.
The coronary CTA- and CCA-guided PCI procedure involved placing the CT and fluoroscopic images side by side before or during PCI. The coronary CTA images were reconstructed using post-processing software to show images with a similar closed angle to that of the CCA image (Figure 2). The procedure was abandoned in case of complications, such as coronary dissection or perforation, that could not be ameliorated after treatment.
Each CTO lesion was regarded as an independent observation for statistical analysis, which was performed using commercially-available software (SPSS Statistics version 13.0, IBM Corporation, Armonk, New York). The quantitative variables were expressed as mean ± SD or median (interquartile range [IQR]). Interobserver and intraobserver agreement were expressed as percentages of agreement and as Cohen k values for categorical variables. The Student t test was used for normally distributed data, whereas the Mann-Whitney U test was used for non-normally distributed data; the chi-square test was used for categorical variables. Univariate statistical tests were first performed using a binary logistic regression model to identify variables of coronary CTA and CCA that were associated with failed A-PCI. A multivariate model for the prediction of A-PCI failure was generated by including the variables that were significant (p < 0.05) in the univariate analysis. For multivariate analysis, continuous variables were treated as categorical variables using the cutoff values determined by Youden's criterion after receiver-operating characteristic (ROC) curve analysis. When the same variables of both CCA and coronary CTA were predictive of A-PCI failure on the basis of univariate analysis, the better predictor (such as blunt stump on CCA and tortuous course and lesion length on coronary CTA) was entered into the multivariate analysis. A p value <0.05 was considered to be statistically significant.
During the study period, 658 patients with 702 CTO lesions were screened, and 103 patients with 108 CTO lesions who underwent pre-procedural coronary CTA were enrolled in the study. The study population flow chart, PCI success rate, and indications for pre–coronary CTA are shown in Figure 3. The total success rate was higher for the CTO lesions in which pre–coronary CTA was performed versus those where it was not performed (94 of 108 [87.0%] vs. 454 of 594 [76.4%]; p = 0.016). However, the differences in procedural success with pre–coronary CTA for both A-PCI (80 of 108 [74.0%] vs. 401 of 594 [67.5%]; p = 0.215) and retrograde (R)-PCI (14 of 15 [93.3%] vs. 53 of 61 [86.9%]; p = 0.678) did not reach statistical significance. Significant differences in lesion duration, procedure time, and times of PCI attempts were observed between the failed and successful A-PCI groups (p < 0.05). The cutoff value for CTO duration determined by ROC curve analysis was 6.5 months. The baseline clinical characteristics and the PCI results of the pre-procedural coronary CTA patients are reported in Table 1. Five cases of coronary artery perforation, 1 case of coronary dissection, and 1 case of cardiac tamponade occurred. All of these complications were resolved after treatment.
The median CTO lesion length by coronary CTA was 23.62 mm (IQR: 15.26 to 32.26 mm); the lesions with failed A-PCI (median 28.37 mm [IQR: 15.88 to 49.30 mm]) were longer than those with successful A-PCI (median 22.77 mm [IQR: 14.66 to 30.54 mm]; p = 0.025). The cutoff value determined by ROC curve analysis was 31.89 mm. Negative remodeling by coronary CTA (Figure 1) was observed in 14 cases in the failed A-PCI group (50.0%); this frequency was significantly higher than that in the successful A-PCI group (p < 0.001). Staged R-PCI was attempted for 5 of 14 negative remodeling lesions; 1 case of R-PCI failure showed a DO/DN of <0.5, whereas the other 4 cases showed 0.5 < DO/DN <1. Calcification in our cases was eccentric, scattered, or sandwich biscuit-like (Figure 2), and in all cases was <360° in circumference. There were 4 extremely severe calcification cases (defined as calcified area ≥90%), and 3 of them were successfully recanalized. Other coronary CTA characteristics are shown in Table 2. The interobserver agreement on coronary CTA was good for all lesions (k = 0.75) as was the intraobserver agreement for all lesions (k = 0.77 and 0.82, for observer 1 and 2 respectively).
Long lesion length, blunt stump, ostial or bifurcation lesions, and a tortuous course were more frequently observed in the failed A-PCI group than in the successful A-PCI group. No difference in other characteristics on CCA was observed between these 2 groups. Detailed data are shown in Table 3. The 15 patients who underwent attempted R-PCI showed good collateral vessels (score of 3 in 14 cases, score of 2 in 1 case) and good visualization of the occluded distal segment (14 of 15); PCI was successful in 14 of these 15 cases (93.3%). The interobserver agreement on CCA was good for all lesions (k = 0.73), as was the intraobserver agreement for all lesions (k = 0.76 and 0.79 for observer 1 and 2, respectively).
In the multivariate model, the independent negative predictor identified was ostial or bifurcation lesions on CCA (odds ratio [OR]: 8.02; 95% confidence interval [CI]: 1.90 to 35.36; p = 0.005), and the presence of negative remodeling (OR: 137.82; 95% CI: 11.69 to 1,624.36; p < 0.001) and a lesion length >31.89 mm on coronary CTA (OR: 7.04; 95% CI: 1.72 to 28.86; p = 0.007). Stump morphology, tortuous course on CCA and coronary CTA, and a CTO duration >6.5 months were not independent predictors on the basis of multivariate analysis (p > 0.05 for each) (Table 4). The success rate of A-PCI in cases without negative remodeling was 85.0% (79 of 93), and the overall success rate of PCI in these cases was 96.8% (90 of 93). Two or more risk factors for A-PCI failure were present in 20 of 108 patients (18.5%). Compared with patients with 0 or 1 risk factor, those with 2 or 3 risk factors had a markedly lower A-PCI success rate (4 of 20 [20.0%] vs. 76 of 88 [86.4%]; p < 0.001). The A-PCI success rates according to the number of risk factors are shown in Table 5.
In our study, the success rate of A-PCI was 74.1% and the total PCI success rate was 87.0%. Both are higher than the rates reported in previous studies (1,12,17). Negative remodeling and lesion length >31.89 mm on coronary CTA, and ostial or bifurcation lesions on CCA were independent predictors of failed A-PCI guided by coronary CTA and CCA. In the absence of these 3 risk factors, the success rate of A-PCI was 98.0%. In contrast, the success rate of A-PCI was especially low (20.0%) if 2 or more risk factors were present. Other signs, such as calcification and tortuous course, were not significantly different between the successful and failed A-PCI groups. R-PCI during the same procedure after failed A-PCI in patients with a satisfactory appearance of the occluded distal segment on coronary CTA and well-developed collaterals (score of 3) on CCA was feasible.
Similar to our results, Rolf et al. (3) reported that performing coronary CTA before PCI of CTO lesions was associated with a higher success rate. In our study, the difference in the success rate between the A- and R-PCI procedures in patients with versus without pre–coronary CTA was not significant, possibly due to an insufficient sample size in each group or selection bias.
Negative remodeling by coronary CTA, which we found in 14 lesions (50.0%) in the failed A-PCI group, was the strongest predictor of failed A-PCI (OR: 137.82). If negative remodeling was absent, the success rate of A-PCI was 85.0% (79 of 93). This predictor was also found by Ehara et al. (11). Negative remodeling occurs in the chronic phase after arterial occlusion (18).
Several factors can result in the failure of the guidewire to cross a negatively remodeled CTO. First, negative remodeling reflects vessel diameter. A histological study of rabbit CTOs demonstrated that the arterial occlusion segment reduced overall vessel size by approximately 80% within 6 weeks and that this reduction increased over time (12). Autopsies of human coronary arteries also show that negative remodeling occurs during the late stage of CTO (12,18). Second, the negatively-remodeled tissue represents a form of wound healing characterized by the replacement of an initially proteoglycan-rich extracellular matrix with a collagen-rich scar, producing a stiff vessel lumen (12,18). Furthermore, anatomic evidence of a “proximal fibrous cap,” a thickened structure at the entrance of the CTO containing particularly densely-packed collagen, makes it more difficult for a guidewire to cross the occlusion (13) and easier to perforate the vessel. By avoiding tough fibrous caps at the proximal end and by utilizing subintimal dissection techniques, some studies have reported that R-PCI improved the total success rate of PCI (17). In our series, R-PCI was successful in 14 of 15 cases (93.3%) after failed A-PCI. These CTOs were characterized by good collateral vessels and good visualization of the occluded distal segment. In addition, the DO/DN of the single failed R-PCI case was <0.5, representing severe vessel sclerosis and fibrosis (12), which may have contributed to the unsuccessful R-PCI.
CTO calcification has classically been associated with revascularization failure (6,7). However, no significant difference in calcification was observed between the failed and successful A-PCI groups in our study, although the proportion of CTOs with severe calcification was nearly identical to that in previous reports (3). Several explanations may underlie this result. First, calcification in our cases was commonly scattered, eccentric, or sandwich biscuit-like (Figure 2Eb to 2Ed). Even the extremely-calcified CTOs had a slender portion of noncalcified residue in the lumen (Figure 2Ed), which is an important factor for successful recanalization. Using the increased resolution of 256-slice CT for coronary CTA, the noncalcified residue lumens of occluded arteries are more sensitively identified (6,7), and the operator could easily find the optimal anchor point and path to avoid calcification (Figure 2) (3).
Additionally, we defined severe calcification as a calcified area >50%, and only 4 patients had extremely severe calcification (calcified area >90%) in our study. This may have been due to exclusion of patients without a previous CABG procedure. Because no previous study has reported on the proportion of patients with extremely severe calcification, it remains difficult to compare the proportion of such cases in our study with the results of others. The pathological study by Sakakura et al. (18) found that the calcified area was greater in CTO patients with CABG (duration of graft >2 years) than in those without CABG (9). Nevertheless, 3 of our 4 extremely severe calcification cases achieved successful A-PCI. Advancements of equipment and enhancements in operator expertise also have contributed to incremental improvements in the PCI success rates (4,19). However, due to a lack of cases with 360° calcified lesions in our study, we cannot exclude calcification as an independent negative predictor for those cases.
Ostial or bifurcation lesions were another negative predictor of successful A-PCI in our study, as also reported in the study from Galassi et al. (20). The presence of these lesions led to difficulty in precisely positioning the guidewire at the CTO entry point, reduced guide catheter support, and an increased likelihood that the guidewire would enter the subintimal space.
A greater occlusion length has been associated with lower PCI success rates in previous studies (2), and our results confirmed this finding. However, our study’s lesion length cutoff value of 31.89 mm is much longer than the 15- to 20-mm cutoff value reported previously (21,22). Other signs were not significant predictors of antegrade CTO success in our study, in contrast to previous studies (6,7). In addition to the fact that our study utilized coronary CTA and CCA, these findings might be partly explained by the recent development of interventional devices used exclusively for CTO lesions (6,21,22).
The principal limitation of this study is its retrospective nature, resulting in possible bias in the patients selected for pre-PCI coronary CTA and in those excluded from the study. Additionally, the decision to perform A-PCI versus R-PCI was operator-dependent and was not pre-specified. Second, too few study patients underwent R-PCI to determine the predictors of procedural failure for this technique. Third, the exact location of the proximal or distal margin of the CTO may be hidden on coronary CTA if heavy calcification is present, obscuring the lumen. Therefore, it should be recognized that if an extended length of 360° calcification is present at the proximal or distal CTO margin, lesion length determination by coronary CTA may be inaccurate. (No CTO in the present series had 360° calcification.) Finally, with 6 candidate predictors, the multivariable model may be over-fit. Whether the 3 risk factors of A-PCI failure identified in our study are indeed the most important procedural predictors and whether the routine performance of coronary CTA improves the overall PCI success rate for CTOs requires confirmation in a larger multicenter study. Similarly, the cutoff values for the continuous variables identified by post hoc ROC curve analysis in the present study require validation in future investigations.
The negative predictors of successful A-PCI were negative remodeling and CTO lesion length on coronary CTA and ostial or bifurcation lesions on CCA. For appropriate patients after failed A-PCI, early R-PCI may be a suitable choice for minimal iatrogenic vessel injury and radiation dosage, thereby contributing to rapid, successful revascularization. Additionally, an initial R-PCI (rather than A-PCI) attempt might be considered in patients with 2 or more indicators of high risk for A-PCI failure.
COMPETENCY IN MEDICAL KNOWLEDGE: Negative predictors for successful A-PCI can be useful in guiding clinical management for CTOs to achieve rapid successful revascularization with minimum iatrogenic vessel injury and radiation dosage. Additionally, an initial R-PCI (rather than A-PCI) attempt might be considered in appropriate patients.
TRANSLATIONAL OUTLOOK: Additional clinical, multicenter studies with a large sample and more events are warranted to determine whether the predictors we found are indeed the most important procedural predictors and whether the routine performance of coronary CTA might improve the overall PCI success rate of CTOs.
This study was supported by the National Key Technology R&D Program of China (No. 2011BAI11B22), the Science and Technology Planning Project of Guangdong Province (No. 2009B030801257), and the Guangdong Province Medical Research Foundation (No. A2013036). The authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- coronary artery bypass graft
- conventional coronary angiography
- computed tomography angiography
- chronic total occlusion
- right coronary artery
- Received November 4, 2014.
- Revision received March 23, 2015.
- Accepted April 8, 2015.
- American College of Cardiology Foundation
- Olivari Z.,
- Rubartelli P.,
- Piscione F.,
- et al.
- Fefer P.,
- Knudtson M.L.,
- Cheema A.N.,
- et al.
- Patel V.G.,
- Brayton K.M.,
- Tamayo A.,
- et al.
- Galassi A.R.,
- Tomasello S.D.,
- Crea F.,
- et al.
- Jaffe R.,
- Leung G.,
- Munce N.R.,
- et al.
- Sumitsuji S.,
- Inoue K.,
- Ochiai M.,
- Tsuchikane E.,
- Ikeno F.
- Choi J.-H.,
- Song Y.B.,
- Hahn J.-Y.,
- et al.
- Werner G.S.,
- Ferrari M.,
- Heinke S.,
- et al.
- Sakakura K.,
- Nakano M.,
- Otsuka F.,
- et al.
- Morino Y.,
- Kimura T.,
- Hayashi Y.,
- et al.
- Martín-Yuste V.,
- Barros A.,
- Leta R.,
- et al.
- Morino Y.,
- Abe M.,
- Morimoto T.,
- et al.