Efficacy of Coronary Computed Tomography Angiography for the De Novo Detection of Chronic Total Occlusion Prior to Coronary Angiography: A Preliminary and Retrospective Study

Abstract

Coronary computed tomography angiography (CCTA) offers high-resolution anatomic characterization of the coronary vasculature but may be suboptimal for lesions dependent on real-time visualization of flow including chronic total occlusion (CTO). In CTOs, heavy calcification and distal vessel opacification from collateralization may confound luminal assessment. Several studies have examined the role of CCTA in characterizing known CTOs to guide percutaneous coronary intervention (PCI). However, the efficacy of CCTA in the de novo diagnosis of CTOs prior to coronary angiography (CAG) has not been demonstrated. A total of 233 consecutive patients who presented for CAG within a 3-month period of having CCTA were retrospectively reviewed. Those patients with prior diagnosis of CTO or prior bypass of the occluded vessels were excluded. Sensitivity and specificity analysis of CCTA in identifying CTOs using CAG as the gold standard was performed. The prevalence of CTO was 21.11% in the population that met criteria for analysis ( n  = 199). The sensitivity of CCTA in predicting CTO was 57.1%, while the specificity was 96.8%. The positive predictive value and negative predictive value of CCTA in detection of CTO were 82.8 and 89.4%, respectively. Our study shows that CCTA has excellent specificity but poor sensitivity in the detection of CTO thus limiting its clinical use in de novo diagnosis. Further studies to determine the effect of de novo CTO diagnosis on clinically important procedural factors, such as radiation exposure, contrast use, and need for repeat procedures, are warranted and may implicate a role for CCTA in this setting.

Coronary computed tomography angiography (CCTA) offers high-resolution anatomic characterization of the coronary vasculature. However, it may be suboptimal for lesions dependent on real-time visualization of flow including chronic total occlusions (CTO). ¹ Chronic total occlusion is defined as 100% occlusion with TIMI (Thrombolysis in myocardial infarction) 0 flow for at least 3 months ^{2 3} and is seen in 20 to 25% of patients undergoing diagnostic coronary angiography (CAG). ⁴ Heavy calcification and distal vessel opacification from collateralization may confound luminal assessment of CTO by CCTA, which is subject to blooming artifact ( Fig. 1 ). Studies have demonstrated that CCTA accurately characterizes CTO when the presence of such an occlusion has already been established by CAG. ⁵ As a guide to percutaneous coronary intervention (PCI) of CTO, CCTA has been shown to not only predict outcomes, ^{6 7 8} but improve procedural success rate, ^{9 10} clinical outcomes, ^{11 12} and help limit complications. While assessment of anatomic features germane to successful CTO-PCI may be difficult to assess by conventional CAG, ⁸ CCTA provides reliable three-dimensional measurement of length and calcification of occluded segments. By delineation of coronary calcification, CCTA offers characterization of CTO morphology and trajectory to facilitate crossing of occlusive plaque. ¹³ Longitudinally, features on initial CCTA can help distinguish acute versus chronic total occlusion ² and predict the natural progression of untreated CTO lesions, suggesting that some lesions may be amenable to future revascularization. ¹⁴ CTO diagnosis by CCTA also appears to have prognostic implications compared with less severe grades of coronary obstruction. ¹⁵ However, there is no current literature on the efficacy of CCTA in the de novo diagnosis of CTO compared with CAG. Additionally, it is not known in this context whether CTO may alter procedural characteristics during CAG, such as procedure time, contrast exposure, and radiation exposure. Therefore, the aim of our study is to assess the de novo diagnostic efficacy of CCTA in the assessment of CTO prior to CAG.

Fig. 1.

An example of coronary computed tomography angiography (CCTA) images demonstrating chronic total occlusion (CTO). A 3D reconstruction through CCTA demonstrates both the CTO and flow in distal vessels ( A ). Maximum intensity axial window ( B ), reconstruction MIP ( C and D ) are shown. However, on conventional angiography, the vessel is not seen distal to the CTO ( E and F ) in LAO-Cranial angiography ( E ) and LAO- caudal angiography ( F ).

Method

Study Design and Population

The Johns Hopkins University CAG database was retrospectively queried over the 10-year period, from 2004 to 2014, for patients over the age of 18 years, who had undergone CCTA within 3 months prior to CAG. The angiographic definition of CTO is complete interruption of the contrast-enhanced lumen along with luminal opacification of the distal to the occlusion site. Given the study design of de novo diagnosis of CTO by CCTA prior to CAG, CCTA diagnosis was based on the clinical expertise of a CCTA board-certified reader. Patients with previously known CTO (including small vessel CTO) or prior bypass with graft anastomosis distal to the CTO (unless the graft itself was totally occluded) were excluded. Of the 233 patients initially screened, 199 patients met all criteria and were selected for analysis. Baseline clinical characteristics are based on the active problem listed in the medical record of the patient including hypertension, hyperlipidemia, and other diseases.

CCTA and Coronary Angiography Protocol and Analysis

All CCTA imaging was performed on Toshiba Aquilon 64 and 320 slice scanners (Canon Medical Systems Corp., Tochigi, Japan). Scans were conducted in accordance with Society of Cardiovascular Computed Tomography (SCCT) guidelines for scan performance and interpretation. ^{16 17} Oral β-blockers (50–100 mg metoprolol tartrate) were administered 1 hour prior to imaging to achieve a target heart rate less than 65 beats per minute. As CAG was performed after CCTA, CCTA results were inherently blinded to results of CAG. The reverse cannot be ascertained. All CCTA scans were reviewed prior to the CAG by board-certified specialist. Calcium score analysis was not performed in these patients. Procedure data including procedure time, contrast volume, and radiation exposure time were collected.

Ethical Approval

Informed consent was not necessary due to the retrospective nature of this work and deidentified chart review. The methods of this study were approved by the Johns Hopkins, School of Medicine Institutional Review Board (Approval no.: IRB0062857). The study has been performed in accordance with the 1964 Declaration of Helsinki and its lateral amendments.

Statistical Analysis

Data are presented as mean ± standard deviation (SD) for continuous variables and frequencies with percentage for categorical variables. Unpaired Student's t -test (for continuous variables) was used to compare groups. Categorical variables were compared using the Chi-square test. If one of the cells had an expected count of less than 5, Fisher's exact test was used. For linear categorical values (number of vessels involved), linear by linear association was used. Sensitivity, specificity, positive predictive value and negative predictive value analysis, and likelihood ratio of CCTA in identifying CTO using CAG as the gold standard was calculated and expressed with 95% confidence interval (CI) of the mean. Statistical significance was defined as a two-tailed p -value of less than 0.05. Statistical analyses were performed with SPSS (Version 24.0, SPSS Inc., Chicago, IL).

Results

Clinical characteristics of our study population are summarized in Table 1 . Of the 199 patients reviewed, 42 patients (21.1%, CI: 0.158–0.276) were diagnosed with CTO by CAG. There was no difference in age, gender, or prevalence of comorbid conditions. Also, there was no difference in clinical characteristics between CCTA positive (true positive) and CCTA negative (false negative) CTO patients.

Table 1. Basic characteristics of patient population in study.

Variable	Patients with CTO ( n  = 42)			Patients without CTO ( n  = 157)
	True positive CTO on CCTA ( n  = 24)	False negative CTO on CCTA ( n  = 18)	p -Value a	True negative CTO on CCTA ( n  = 152)	False positive CTO on CCTA ( n  = 5)	p -Value b
Age c	61 ± 13	66 ± 9	0.18	60 ± 12	53 ± 19	0.18
Creatinine	0.90 ± 0.21	0.96 ± 0.18	0.36	1.13 ± 1.10	1.87 ± 2.14	0.20
Hypertension	13 (54.2%)	14 (77.8%)	0.19	99 (65.1%)	5 (100%)	0.86
Diabetes	4 (16.7%)	4 (22.2%)	0.71	34 (22.4%)	3 (60%)	0.68
Hyperlipidemia	16 (66.7%)	13 (72.2%)	0.75	95 (62.9%)	5 (100%)	0.59
Smoking	8 (33.3%)	3 (16.7%)	0.30	43 (28.3%)	0 (0%)	1.00
Family history of premature CAD	7 (29.2%)	3 (16.7%)	0.47	37 (24.3%)	1 (20%)	1.00
Systolic heart failure	5 (20.8%)	3 (16.7%)	1.00	18 (11.8%)	0 (0%)	0.20
Diastolic heart failure	1 (4.2%)	4 (22.2%)	0.146	23 (15.1%)	1 (20%)	0.81
Atrial fibrillation	4 (16.7%)	2 (11.1%)	0.69	28 (18.4%)	1 (20%)	0.65
Anemia	5 (20.8%)	4 (22.2%)	1.00	40 (26.3%)	2 (40%)	0.56
Thyroid disease	4 (16.7%)	3 (16.7%)	1.00	20 (13.2%)	3 (60%)	0.81
Dialysis	0 (0%)	0 (0%)	N/A d	6 (3.9%)	1 (20%)	0.35

Abbreviations: CAD, coronary artery disease; CTO, coronary total occlusion; CCTA, coronary computed tomography angiography; N/A, not available.

Note: Data are summarized as mean ± standard deviation for age and creatinine or number of patients (percent %) for other medical conditions.

^a p -Value for comparing true positive CTO and false negative CTO.

^b p -Value for comparing patients with CTO and patients without CTO.

^cAge is expressed in years. Creatinine is expressed as mg/dL.

^d p -value is not calculable due to no cases.

The diagnostic accuracy of CCTA in diagnosing CTO is summarized in Table 2 . The sensitivity of CCTA in predicting CTO was 57.1% (95% CI: 47.1–71.9%), while the specificity was 96.8% (95% CI: 92.3–98.8%). The positive predictive value and negative predictive value of CCTA in detection of CTO were 82.8% (95% CI: 63.5–93.4%) and 89.4% (95% CI: 83.5–93.4%), respectively. The positive likelihood ratio was 17.94 (95% CI: 7.29–44.19) and the negative likelihood ratio was 0.44 (95% CI: 0.31–0.63).

Table 2. Diagnostic accuracy of CCTA in diagnosing CTO.

Prevalence of CTO ( n  = 42)	21.1% (95% CI: 15.8–27.6%)
Sensitivity of CCTA	57.1% (95% CI: 41.1–71.9%)
Specificity of CCTA	96.8% (95% CI: 92.3–98.8%)
Positive predictive value of CCTA	82.8% (95% CI: 63.5–93.5%)
Negative predictive value of CCTA	89.4% (95% CI: 83.5–93.43%)
Positive likelihood ratio	17.94 (95% CI: 7.29–44.19)
Negative likelihood ratio	0.44 (95% CI: 0.31–0.63)

Abbreviations: CI, confidence interval; CCTA, coronary computed tomography angiography; CTO, chronic total occlusion.

Note: Data are expressed as diagnostic accuracy % (95% CI of mean).

The procedural characteristics of CAG after CCTA are summarized in Table 3 . Having CCTA positive for CTO did not result in a statistically significant difference in procedure time (mean, 80.9 vs. 76.9 minutes), contrast use (mean, 166.8 vs. 139.0 cc), and radiation exposure (mean, 11.3 vs. 9.5 minutes) during CAG, compared with patients with negative CTO, although there was a trend showing reduced procedure time, contrast use, and radiation exposure when CTO was not diagnosed. Given the high specificity and negative predictive value, we made following three additional exploratory comparisons: (1) the true negative group versus all other patients, (2) all true (positive and negative) patients versus all false patients, and (3) all negative patients (true and false) versus all positive patients. In all three cases, there was a nonsignificant, but consistent trend in the same direction for all three procedural variables.

Table 3. Characteristics of coronary angiography.

Variable	Patients with CTO ( n  = 42)			Patients without CTO ( n  = 157)
	True Positive CTO on CCTA ( n  = 24)	False Negative CTO on CCTA ( n  = 18)	p -Value a	True Negative CTO on CCTA ( n  = 152)	False Positive CTO on CCTA ( n  = 5)	p -Value b
Time to CAG (d)	14.30 ± 17.89	10.74 ± 15.99	0.51	9.93 ± 11.09	12.46 ± 17.07	0.21
CAG procedure time (min)	81.08 ± 36.80	80.72 ± 37.87	0.98	75.56 ± 32.80	117.00 ± 22.86	0.49
CAG contrast use (cc)	157.08 ± 102.28	179.72 ± 163.73	0.59	137.36 ± 85.59	189.00 ± 71.10	0.10
CAG radiation exposure (min)	10.28 ± 9.23	12.54 ± 12.22	0.50	9.07 ± 8.37	21.044 ± 10.93	0.26
Repeat angiography	0 (0%)	1 (5.6%)	0.43	3 (2.0%)	1 (20%)	1.00
Number of vessels involved			0.557			<0.001
0 vessel	0 (0%)	0 (0%)		52 (34.2%)	0 (0%)
1 vessel	6 (25%)	4 (22.2%)		44 (28.9%)	1 (20%)
2 vessels	7 (29.2%)	9 (50%)		31 (20.4%)	4 (80%)
3 vessels	11 (45.8%)	5 (27.8%)		25 (16.4%)	0 (0%)

Abbreviations: CAG, coronary angiography; CTO, coronary total occlusion; CCTA, coronary computed tomography angiography.

Note: Data are summarized as mean ± standard deviation for continuous variables or number of patients (percent %) for other medical conditions.

^a p -Value for comparing true positive CTO and false negative CTO.

^b p -Value for comparing patients with CTO and patients without CTO.

Discussion

To our knowledge, this is the first study to evaluate the diagnostic efficacy of de novo CTO diagnosis with CCTA using CAG as a gold standard. Our study shows that CCTA has excellent specificity (96.8%) but poor sensitivity (57.1%) in the detection of CTO, thus limiting its clinical use as a screening tool. The prevalence of CTO in our population was consistent with previously described rates of CTO among patients undergoing CAG. Possible causes of low sensitivity of CCTA in diagnosis of CTO are collateralization and calcium deposition. Collateral flow allows for vessel opacification in a retrograde fashion, which generates the impression of a patent vessel on standard imaging protocols, when in reality, anterograde flow is impaired by CTO. ⁹ Calcium deposition at the site of stenosis blooms on CCTA and confounds our ability to know true luminal narrowing (severe stenosis versus total occlusion) at a site of heavy calcium deposition. Currently, the gold standard of diagnosis of CTO is CAG. However, compared with non–CTO-PCI, success rates of CTO-PCI are low, and incorporating other imaging techniques, such as CCTA, intravascular ultrasound, or optical coherence tomography, for increasing success rate is being investigated. ¹⁸ It is already known that CCTA aids in the management of CTO in multiple ways (i.e., increasing success rate and predicting complication and prognosis). For example, a study by von Erffa et al attempted to differentiate high-grade stenosis and complete occlusion through CCTA. They report that CTO lesions tend to have longer occlusion lengths compared with high-grade stenosis (16.6 vs. 4.6 mm), thus guiding in differentiation of complete coronary occlusion from high-grade coronary artery stenosis and informing treatment approach. ¹⁹ Another group from China has shown that combined imaging guidance with CT and SPECT (Single photon emission computed tomography) provides a more accurate assessment of CTO lesions and is beneficial compared with traditional CAG alone. ⁹ To date, there are no imaging modalities that directly compare the de novo diagnosis of CTO compared with CAG. However, there are imaging modalities known to assist in management of CTO such as cardiac MR imaging (MRI) which can identify myocardial viability, inducible perfusion defect of the CTO, as well as predict patients with successful intervention of CTO lesions. ^{20 21} In addition, novel echocardiographic techniques, such as three-dimensional systolic dyssynchrony index and global longitudinal strains, can be utilized to assess for success after recanalization of the CTO. ²²

Overall, patients had similar procedural time and contrast volume to other previous studies. ^{12 23} Among the CTO-diagnosed patient group, statistically significant differences between CCTA-positive and CCTA-negative patients were not seen with regard to procedural characteristics. However, there was a favorable trend toward reduction in contrast use, procedure time, and radiation exposure both in the primary analysis and across multiple exploratory analyses as described. As the use of CCTA in CTO-intervention planning is known to favorably impact procedural parameters, ¹¹ there is likely to be a true correlative effect of CCTA on procedural parameters in the de novo setting. Future studies with the benefit of additional power may reveal this effect.

Limitations of Study

This study has several limitations. First, it is a retrospective observational study. Standardized clinical imaging protocols described in the methods section were used but scanners varied from 64 to 320 slices. Second, characterization of individual lesions was not performed. Parameters, such as lesion length, serial occlusion, blunt stump, calcification, calcium arc, or vessel tortuosity, are criteria for the CT-RECTOR score ⁶ that predicts success of CTO PCI. ²⁴ Although our study was performed in a different clinical context and was meant to assess the binomial distribution of CTO diagnosis, assessment of these parameters may be informative. Future studies may also want to assess collateralization grade on conventional angiography, as it impacts the fidelity CTO diagnosis by CCTA.

Conclusion

In conclusion, this retrospective study shows that CCTA has excellent specificity but poor sensitivity in the de novo detection of CTO. Better-powered analyses are needed to assess the potential of CCTA in favorably effecting procedural variables during subsequent index CAG.