Accuracy of J-CTO Score Derived From Computed Tomography Versus Angiography to Predict Successful Percutaneous Coronary Intervention
Author + information
- Akiko Fujino, MDa,∗ (kakibouzu{at}gmail.com),
- Satoru Otsuji, MDa,
- Katsuyuki Hasegawa, MDa,
- Toyohiro Arita, RTa,
- Shin Takiuchi, MD, PhDa,
- Kenichi Fujii, MD, PhDa,
- Masanori Yabuki, MD, PhDa,
- Motoaki Ibuki, MDa,
- Shinya Nagayama, MDa,
- Kasumi Ishibuchi, MDa,
- Toshikazu Kashiyama, MDa,
- Rui Ishii, MD, PhDa,
- Hiroto Tamaru, MDa,
- Wataru Yamamoto, MDa,
- Masahiko Hara, MD, PhDb and
- Yorihiko Higashino, MDa
- aDepartment of Cardiology, Higashi Takarazuka Satoh Hospital, Takarazuka, Japan
- bDepartment of Clinical Epidemiology and Biostatics, Osaka University Graduate School of Medicine, Suita, Japan
- ↵∗Address for correspondence:
Dr. Akiko Fujino, Department of Cardiology, Higashi Takarazuka Satoh Hospital, 2-1, Nagaocho, Takarazuka, Hyogo 665-0873, Japan.
Graphical abstract
Abstract
Objectives The aim of this study was to compare the ability of conventional versus computed tomography angiography (CTA) to predict procedural success and 30-min wire crossing rates in percutaneous coronary intervention (PCI) for chronic total occlusion (CTO) lesions.
Background Coronary CTA can be used to assess the morphology of CTO lesions.
Methods We examined 205 consecutive patients (218 CTO lesions) who underwent coronary CTA pre-PCI. The J-CTO (Multicenter CTO Registry of Japan) score (the sum of the following 5 binary parameters: blunt proximal cap, calcification, bending >45°, and length of occluded segment >20 mm plus previously failed PCI attempt) was calculated using both CTA and conventional coronary angiography and compared.
Results The median patient age was 69 years (interquartile range: 62 to 75 years), 82.4% were male, and a retrograde approach was attempted in 72 (33.0%) cases. The procedural success rate of the CTO-PCI procedures was 82.6%, and 29.4% of cases achieved 30-min wire crossing. The areas under the curve of the CTA-derived J-CTO score for predicting procedural success and 30-min wire crossing were significantly greater than those derived from conventional angiography (0.855 vs. 0.698; p < 0.001 for procedural success and 0.812 vs. 0.692; p < 0.001, for 30-min wire crossing). In addition, the areas under the curve of CTA-derived evaluations of calcification, bending, and occlusion length were significantly higher than those of derived from angiography for predicting procedural success.
Conclusions The CTA-derived J-CTO score was a more useful predictor of both procedural success and 30-min wire crossing than the J-CTO score derived from conventional angiography.
Despite notable progress in the development of technology, percutaneous coronary intervention (PCI) of coronary chronic total occlusion (CTO) lesions remains technically challenging; and the success rate of a CTO-PCI is lower than PCI of a nonocclusive coronary artery stenosis (1). As an objective index of procedural difficulty, the J-CTO (Multicenter CTO Registry of Japan) score has been introduced using 4 morphologic characteristics of a CTO lesion including entry shape, calcification, bending, and occlusion length (2). Although the utility of the J-CTO score has been confirmed externally by several previous reports with excellent predictive accuracy of 30-min wire crossing (3–5), recent advancements in coronary computed tomography angiography (CTA) may provide additional information compared to conventional angiographic findings when predicting the difficulty of the CTO-PCI procedure (6,7). For example, it has been reported that lesion length may be assessed more accurately with coronary CTA than conventional angiography (8–10). However, there have been few reports that have directly compared the ability of conventional angiography versus CTA to predict CTO-PCI procedural difficulty (11). The objective of this study was to compare the discriminating accuracy of the CTA-derived J-CTO score with that of the conventional coronary angiography–derived J-CTO score for predicting procedural success and 30-min wire crossing during a CTO-PCI procedure and to clarify the performance of CTA to predict procedural difficulties when performing a CTO-PCI.
Methods
Study population
This was a single-center, retrospective, observational study enrolling 205 consecutive patients with 218 lesions who underwent coronary CTA before PCI for a CTO at Higashi Takarazuka Satoh Hospital, Takarazuka, Japan, between January 2012 and April 2016. The exclusion criteria included a history of acute myocardial infarction within 3 months, patients not undergoing a coronary CTA within 1 month before the procedure, and patients with previously implanted stents located at the occlusion site because in-stent occlusions were not factored into the original J-CTO score (Figure 1) (2). A CTO was defined as a coronary lesion with TIMI (Thrombolysis in Myocardial Infarction) flow grade 0 for at least 3 months as estimated using clinical information or the results of previous angiography (1). All variables shown in Tables 1 to 3⇓⇓ were retrospectively obtained from patient records. The study protocol complied with the Helsinki Declaration standards and was approved by the Ethical Committee of Higashi Takarazuka Satoh Hospital that waved the requirement for written informed consent because this study used retrospective data obtained from hospital records and there were no interventions in the study patients. Several cases have previously been reported as part of a series in another journal, and we obtained reprint permission for the modified use of different images compared to the originals (12).
Patient Selection Flow
AMI = acute myocardial infarction; CT = computed tomography; CTO = chronic total occlusion; PCI = percutaneous coronary intervention.
Patient and Lesion Characteristics
Comparison of CTA-Derived and Angiography-Derived J-CTO Score
Comparison of AUC Between CTA-Derived and Angiography-Derived J-CTO Score and Each Constituent Element
CTA protocol and analysis
We routinely performed coronary CTA before CTO-PCI except for patients with renal insufficiency with estimated glomerular filtration rate <45 ml/min/1.73 m2. All patients were scanned on a 320-detector Toshiba Aquilion system (Toshiba Corporation, Tokyo, Japan) within 1 month before PCI. Intravenous landiolol was administrated to achieve heart rates of <65 beats/min, and sublingual nitroglycerin was administered before the CTA scan. During acquisition, 0.8 ml/kg of contrast (Iopamiron 370, Bayer Ltd., Osaka, Japan) was injected followed by a saline flush. The scan parameters were as follows: collimation, 320 × 0.5 mm; tube voltage, 120 kV; tube current, 450 mA; rotation time, 350 ms, 375 ms, or 400 ms; slice thickness, 0.5 mm; and reconstructed increment, 0.25 mm. The contrast transit time was estimated by the injection of a test bolus or by using a real-time bolus tracking technique.
CTA was analyzed on a 3-dimensional workstation (ZIO Station, ZIO Soft, Tokyo, Japan) by 2 experienced cardiologists who were blinded to the clinical data and the results of the conventional coronary angiography. The CTA was visualized using axial, multiplanar reformation, and/or maximum intensity projections and cross-sectional analysis. Reconstructed CTA data were used to evaluate the morphology of the CTO lesion, and the J-CTO score was calculated as defined previously (2). The J-CTO score includes 4 morphologic characteristics of a CTO: blunt proximal cap; calcification; bending >45°; and length of occluded segment >20 mm (Figure 2). The morphology of the occlusion entry point was classified as tapered or blunt on the maximum intensity projection image. As defined in the conventional angiography–derived J-CTO score, tapered morphology was an entry point with a funnel-shaped form, whereas blunt morphology lacked a funnel-shaped entry point. A calcified lesion on CTA had a calcified area >50% of the vessel cross-sectional area; this was different from conventional angiography that merely reported the presence or absence of calcification. The length and bending of a CTO segment were evaluated 3-dimensionally using CTA. Bending was defined as the presence of at least one bend of >45° within the length of the occlusion, and occlusion length was measured from the proximal margin to the distal margin of the total occluded segment. Length and bending CTA definitions were also consistent with definitions used in conventional coronary angiography (2).
Representative Cases Showing Discrepancies Between CTA and Conventional Angiography Regarding 4 Morphologic Characteristics of J-CTO Score
(A) Although conventional angiography demonstrated a blunt entry shape, CTA revealed a microchannel. A guidewire easily crossed the lesion. (B) An unsuccessful CTO case in which the guidewire did not pass the lesion because of severe calcification throughout the entire circumference as revealed by CTA. (C) A CTO lesion that appeared straight with conventional angiography. However, CTA clearly showed bending of the CTO segment. (D) Even though the occlusion length of the CTO measured using conventional angiography was 22 mm, the length measured by CTA was only 10 mm. Modified with permission from Fujino (12). CTA = computed tomographic angiography; J-CTO = Multicenter CTO Registry of Japan score; LAD = left anterior descending artery; other abbreviation as in Figure 1.
Conventional coronary angiography and intervention
Conventional coronary angiography and PCI were performed in all 205 patients with 218 CTO lesions. All of the enrolled patients underwent CTO-PCI as a staged procedure (i.e., not ad hoc immediately after coronary angiography); this enabled operators to consider the strategy using both CTA and conventional coronary angiography findings. When contralateral collaterals were present, arterial access was established via a bilateral approach; and dual injection was performed. Two experienced cardiologists independently calculated the J-CTO score using conventional coronary angiography data. Examiners were blinded to the clinical data or the J-CTO scores calculated by CTA.
CTO-PCI was performed by 2 experienced operators. In line with previously published articles, the wire-crossing time was defined as the time from initial insertion of the guidewire into the coronary lumen to the time of successful wire crossing beyond the lesion into the distal vessel (2). The retrograde (vs. the antegrade) approach or intravascular ultrasound-guidance was performed according to the lesion characteristics and operator preference. Procedural success was defined as post-PCI TIMI flow grade 3 and residual stenosis <30%.
Statistical analysis
We set the primary endpoint as procedural success and the secondary endpoint as 30-min wire crossing because procedural success rate has been more attractive for physicians in recent studies, although the 30-min wire-crossing rate has been a more historically objective index of procedural difficulty (2–4,11,13,14). Binary variables were presented as n (%), and continuous variables were presented as median (interquartile range). Continuous variables were compared with the Wilcoxon signed rank test, and categorical variables were compared with the McNemar test. The J-CTO score is the sum of the following 5 binary parameters: blunt proximal cap, calcification, bending >45°, length of occluded segment >20 mm, and previously failed PCI attempt. Each of these independent variables was assigned a value of 1 when present. The discriminating accuracy of each endpoint with either the CTA- or angiography-derived J-CTO score was evaluated using the area under the receiver operator characteristics curve (AUC) with the binary regression model and compared with Delong’s method. To calculate AUC, we used the J-CTO score as a continuous variable. Intraobserver and interobserver agreements were assessed using Lin’s concordance correlation coefficient with data from 35 randomly selected patients. Statistical analyses were performed with R software packages version 3.1.1 (R Development Core Team, Vienna, Austria). Statistical significance was defined as a 2-sided p value <0.05.
Results
A total of 205 patients with 218 CTO lesions were enrolled. The patient and lesion characteristics of the total cohort have been described in Table 1. Median patient age was 69 years (range 62 to 75 years), 82.4% of the patients were male, 29.8% had prior myocardial infarction, and 66.3% had prior PCI. Procedural success was obtained in 180 (82.6%) lesions, and 30-min wire crossing was achieved in 64 (29.4%) lesions.
Overall, 10.6% of the CTO lesions had a previously failed PCI. Regarding procedural techniques, 33.0% PCIs used the retrograde approach; and 13.8% used the controlled antegrade and retrograde tracking (CART) or reverse-CART technique. The median radiation dose was 3,778 mGy (range 2,653 to 5,625 mGy), and the median contrast volume was 327 ml (range 260 to 420 ml).
Intraobserver and interobserver variabilities of both CTA-derived and conventional angiography–derived J-CTO scores showed excellent intraobserver (Lin’s coefficient; 0.93 for CTA and 0.93 for conventional angiography) and interobserver reliability (Lin’s coefficient; 0.91 for CTA and 0.92 for conventional angiography).
In Table 2 and Figure 3, comparisons of the J-CTO scores calculated using CTA versus conventional angiography are shown. In 45% of the lesions, the CTA-derived J-CTO score equaled that determined by conventional coronary angiography. However, 29.8% of the patients had a CTA-derived J-CTO score greater than that derived by conventional coronary angiography, whereas 25.2% of the patients had a CTA-derived J-CTO score less than the conventional angiography–derived J-CTO score.
Comparison of J-CTO Score and Bland-Altman Plots Between CTA and Conventional Angiography
(A) Comparison of J-CTO Score; CTA-derived J-CTO score was equal to conventional angiography-derived J-CTO score in 98 (45.0%) lesions, higher in 65 (29.8%) lesions, and lower in 55 (25.2%) lesions. (B) Bland-Altman Plots; the confidence band of the Bland-Altman plot has been shown in a dotted line and ranged from −2.04 to 2.08. Abbreviations as in Figures 1 and 2.
The AUC of the primary and secondary endpoints are shown in Table 3 and Figure 4 along with sensitivity, specificity, positive predictive value, and negative predictive value. The discriminating accuracies of both CTA-derived and angiography-derived J-CTO scores were acceptable with AUCs of 0.855 and 0.698, respectively, for procedural success of the CTO-PCI and AUCs of 0.812 and 0.692, respectively, for 30-min wire crossing. The AUC of the CTA-derived J-CTO score was significantly higher than that of the angiography-derived J-CTO score for predicting a successful CTO-PCI (Delong p < 0.001) and for 30-min wire crossing (Delong p < 0.001). Table 3 also showed that the AUC of each morphologic parameter of the J-CTO score tended to be higher when assessed using CTA than when assessed using conventional angiography, showing better accuracies of CTA for predicting: 1) procedural success in the evaluation of calcification, bending, and occlusion length; and 2) 30-min wire crossing in the evaluation of entry shape, calcification, and occlusion length. Representative cases showing discrepancies between CTA and conventional angiography regarding the 4 morphologic characteristics included in the J-CTO score are shown in Figure 2. In addition, procedural success and 30-min wire-crossing rates in each J-CTO score category for the entire cohort, antegrade approach group, and retrograde approach group are shown in Figure 5. Findings were consistent regardless of the procedural approach (i.e., antegrade or retrograde).
Receiver Operating Characteristics Curve for Predicting Procedural Success and 30-Min Wire Crossing
(A) Procedural success. (B) 30-min wire crossing. AUC = area under the curve; NPV = negative predictive value; PPV = positive predictive value; Sens = sensitivity; Spec = specificity; other abbreviations as in Figure 2.
Discussion
In this study, we compared the predictive accuracy of a CTA-derived J-CTO score with that derived from conventional angiography using AUC for predicting procedural success and 30-min wire crossing in 205 consecutive patients (218 lesions) who underwent CTO-PCI. The procedural success rate of the CTO-PCI was 82.6% overall, acceptable compared with other contemporary studies (Table 4) (15–17), although lower for the J-CTO score ≥3 subgroup, as compared to recent registries (4). In addition, we showed that the CTA-derived J-CTO score showed a significantly greater predictability of procedural success and 30-min wire crossing compared to that of conventional angiography.
Studies Investigating the Accuracy of Angiography-Derived J-CTO Score for Predicting Procedural Success or 30-Min Wire Crossing of CTO-PCI
Predictive accuracy of the J-CTO score
Several studies have investigated the accuracy of the conventional angiography–derived J-CTO score for predicting successful CTO-PCI or 30-min wire crossing (2–4,11,13,14,18). Table 4 summarizes the AUC of each study and the distribution of the angiography-derived J-CTO score. The AUCs for successful CTO-PCI varied widely among the studies ranging from 0.399 to 0.868, whereas those for predicting 30-min wire crossing varied less widely ranging from 0.692 to 0.820. Considering these results, we confirmed the J-CTO score to be a useful and powerful index for predicting a relatively easy CTO lesion defined as guidewire crossing within 30 min. On the other hand, predicting the procedural success of a CTO-PCI using the J-CTO score may be more challenging for difficult CTO lesions (2). Discrepancies among studies may occur because the technical success of a CTO-PCI can be influenced by other factors such as operator skill, device availability, and allocation of resources, especially in difficult cases.
CTA- versus angiography–derived J-CTO score
To elucidate the reason for the greater predictive accuracy of the CTA-derived J-CTO score, we focused on and compared each of the 4 key morphologic characteristics. The AUC of each morphologic parameter of the J-CTO score tended to be higher in coronary CTA versus conventional angiography (Table 3). These results are consistent with previously published data. For example, Rolf et al. (10) reported that CTA had a superior ability to detect a blunt stump than conventional angiography because the time needed for coronary CTA may result in the retention of contrast medium in the microchannel within the CTO segment (10). With regards to calcification detection, CTA has been reported to be overly sensitive in some situations (9). To overcome this problem, recent studies have introduced the benchmark of the presence of calcified area >50% of the vessel cross-sectional area to indicate clinically meaningful calcification, a definition used in the present study (6–9,11,18). Regarding the assessment of bending and occlusion length, several previous studies reported that computed tomography had a notable advantage over conventional angiography because it provided a 3-dimensional depiction of the CTO segment (8–10). On the other hand, Li et al. (11) reported that J-CTO scores derived from CTA versus conventional angiography showed similar accuracy for predicting a successful CTO-PCI; this could be explained by the exceptionally high AUC of conventional angiography compared to studies, although the AUC of the CTA was similar to other reports. With the above-mentioned differences between CTA and conventional angiography for evaluating the 4 morphologic characteristics, it was reasonable for us to conclude that coronary CTA showed a better predictive accuracy both for successful CTO-PCI and 30-min wire crossing, especially for complex cases.
Clinical implications
First, CTA enables selection of appropriate CTO-PCI candidates pre-procedurally. Second, pre-procedural coronary CTA helps to develop a strategy for CTO-PCI especially in complex cases because of the morphological information provided. Moreover, new real-time fusion technology of coronary CTA has the potential to further contribute to safe treatment strategy planning (19). Additional studies are required to determine how the incremental information available from the CTA-derived J-CTO score should be used to guide patient selection and procedural technique.
Study limitations
First, this was a single-center, retrospective, observational study. Second, we did not consider any morphologic findings apart from the parameters included in the J-CTO score although several new scoring systems have recently been introduced to predict a successful CTO-PCI (13–15) by including additional information such as arterial remodeling, presence of a side branch, distal cap characteristics, and so on, or by weighing the factors differently. Third, lesion information available from the pre-procedural CTA might have contributed to the 30-min wire crossing and successful CTO-PCI. Fourth, we only enrolled elective CTO-PCI patients who underwent enhanced CTA before the procedure. Unfortunately, data regarding excluded patients (Figure 1) were not available because our study protocol was approved only for the final study population. Fifth, data regarding radiation exposure for CTA were not available.
Conclusions
The discriminating accuracy of the CTA-derived J-CTO score was significantly higher than that of the angiography-derived J-CTO score for predicting procedural success and 30-min wire crossing. Coronary CTA has the potential to provide more detailed information regarding CTO morphology compared to conventional angiography. These data need to be validated in a prospective, multicenter study; and if confirmed, a randomized trial examining whether procedural outcomes are improved by pre-procedural CTA imaging is warranted.
COMPETENCY IN PATIENT CARE AND PROCEDURAL SKILLS: Our study suggests that the assessment of CTO lesions with coronary CTA enables us to know more accurate morphological information before the PCI procedure.
TRANSLATIONAL OUTLOOK: Pre-procedural CTA has the potential to improve the technical success rate and shorten procedural time. A randomized study comparing the success rate with versus without pre-procedural CTA would clarify the efficacy of CTA in this setting.
Acknowledgment
The authors thank Dominic P. Francese, MPH for his help in editing this manuscript.
Footnotes
The authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- AUC
- area under the curve
- CTA
- computed tomography angiography
- CTO
- chronic total occlusion
- PCI
- percutaneous coronary intervention
- TIMI
- Thrombolysis in Myocardial Infarction
- Received October 6, 2016.
- Revision received January 23, 2017.
- Accepted January 25, 2017.
- 2017 American College of Cardiology Foundation
