Patients with CAD and clinically relevant coronary artery stenosis assessed by CCTA at Gifu Heart Center (Gifu, Japan) were screened from October 2016 to January 2020. In addition, patients who were diagnosed with non-ST-elevation ACS (NSTE-ACS), stable angina pectoris (SAP), and silent myocardial ischemia (SMI) and scheduled for PCI with NIRS–IVUS (LipiScan, InfraReDx, Burlington, MA, USA) within 90 days after CCTA were prospectively enrolled. Detailed diagnosis definitions are given in the Supplementary Methods. We decided the target lesions according to the evidence of myocardial ischemia by electrocardiogram (ECG), echocardiographic, and physiological findings such as fractional flow reserve (FFR) or stress myocardial perfusion imaging before PCI in correspondence with the most narrowing lesion on coronary angiogram (CAG). The lesion with FFR < 0.80 was diagnosed as positive functional ischemia and considered for PCI.

Patient-based exclusion criteria were (1) simultaneous treatment with multivessel PCI at the initial procedure, (2) chronic kidney disease manifested by an estimated glomerular filtration rate of <30 ml/min/1.73 m², (3) presentation with hemodynamic instability, and (4) severe valvular heart diseases. Lesion-based exclusion criteria were (1) lesions that have already been stented before and those undergoing stent placement during the index PCI, including 5-mm proximal/distal borders; (2) lesions belonging to a vessel that had undergone coronary artery bypass grafting; (3) left main trunk lesions; (4) ostial lesions (<10 mm from the left main bifurcation or coronary ostia to the lesion edge); (5) long lesions requiring two or more stents per lesion; (6) severe calcified lesions requiring atheroablative device prior to NIRS-IVUS imaging; (7) severe stenotic lesions for which difficulty was anticipated in advancing the NIRS-IVUS catheter without plain balloon angioplasty of ≥2.0-mm diameter; (8) poor image quality for which the technician could not perform either CCTA or NIRS-IVUS analyses; (9) lesions located in a small vessel [with an external elastic membrane (EEM) diameter of <2.0 mm]; and (10) use of a distal protection device during PCI.

In accordance with the above criteria, we enrolled 142 consecutive patients in the study after CCTA. Of these, we excluded 17 patients who met patient exclusion the criteria before undergoing PCI and 11 patients who met the lesion exclusion criteria at the time of PCI. Additionally, 5 patients who were unable to provide informed consent and 2 patients who did not undergo PCI were excluded. Finally, we included 107 patients (98 with SAP/SMI and 9 with NSTE-ACS) in the study.

Written informed consent was given by each patient. The study was conducted in accordance with the Declaration of Helsinki and authorized by the Institutional Review Board (2015015). This study was registered with the University Medical Information Network Clinical Trial Registry (UMIN000042883).

The CCTA protocol is provided in the Supplementary Methods. Briefly, ECG-gated CT angiography was performed using SOMATOM Definition AS+ (Siemens Healthcare, Fordchheim, Germany). Iopamiron 370 mg iodine/ml (Bayer Health Care, Osaka, Japan) or Omnipaque 350 mg iodine/ml (Daiichi Sankyo, Tokyo, Japan) were used in a contrast-enhanced ECG gating scan to collect CCTA data.

Experienced technicians blinded to quantitative coronary angiography (QCA) and NIRS–IVUS performed the scan analysis using the Aquarius iNtuition edition version 4.4.8 (TeraRecon, Frankfurt, Germany) three-dimensional workstation. Cross-sectional images were used to evaluate the outer vessel area and the remodeling index (RI) of the vessel. The RI was calculated as the EEM cross-sectional area (CSA) of the target lesion divided by the average of the EEM-CSAs of the proximal and distal reference segments; positive remodeling was defined as a RI of ≥1.10 (4).

We evaluated the following plaque characteristics: (1) CT density, (2) spotty calcification, and (3) napkin-ring sign. The CT Hounsfield unit (HU) values of the culprit plaque were determined from at least three rounded regions of interest of 0.5 mm², and the unit score at each region of interest was measured and then averaged. Low-attenuated plaque was defined as any voxel of <30 HU within a coronary plaque (10). Spotty calcification was classified as <3 mm in size on curved multiplanar reformation images and unilateral on cross-sectional images (4). The following criteria were used to define a napkin-ring sign: (1) existence of high attenuation ring around certain coronary artery plaques or (2) CT attenuation of the ring, presenting higher than those of the adjacent plaque and not >130 HU (11).

In this study, cTnT was detected by HISCL™ Troponin-T hs assay kit (Sysmex, Kobe, Japan), with a suggested upper limit of normal (ULN) cutoff value of 0.016 ng/ml; thus, periprocedural myocardial injury was defined as a cTnT level of >5 times the ULN at either 6 or 18–24 h after PCI, according to the fourth universal definition of MI (12). In patients with NSTE-ACS who had elevated preprocedural cTnT levels, the postprocedural cTnT must rise by >20% from the preprocedural value. However, the absolute postprocedural value must still be >5 times our institution's ULN.

The Supplementary Methods provide a description of the preoperative antiplatelet therapy and perioperative anticoagulants. CAG and PCI were performed using standard techniques, and the selection of devices and pharmacotherapy was at the operator's discretion.

Following intracoronary injection of 2 mg of isosorbide dinitrate, NIRS–IVUS images were obtained using a commercially available system (LipiScan, InfraReDx, Burlington, MA, USA). The NIRS-IVUS system used a 40-MHz (TVC Insight™, model TVC-MC8) or 50-MHz transducer (Dualpro™, model TVC-MC10) in a 3.2 Fr catheter. NIRS–IVUS imaging was performed before either direct stent implantation or balloon predilatation. An automatic mechanical pullback was used to bring the IVUS catheter back to the aorto-ostial junction at a speed of 0.5 mm/s. The system demonstrates grayscale IVUS images and color-coded NIRS data in a single pullback at the same time, as described in Supplementary Methods. All image data were stored on a DVD and forwarded to the local core laboratory for analysis.

A verified automated edge-detection software (CCIP-310/W, Cathex, Tokyo, Japan) was used to perform QCA analysis. Using orthogonal projections, the minimum lumen diameter, reference diameter, and lesion length were assessed in diastolic frames.

Off-line gray-scale IVUS and NIRS analyses were performed using QIvus 3.1 (Medis Medical Imaging Systems, Leiden, The Netherlands). Experienced investigators blinded to the clinical findings, angiography, and CCTA characteristics analyzed these data.

On the basis of the American College of Cardiology Clinical Expert Consensus Document on Standards for Acquisition, Measurement, and Reporting of IVUS Investigations, whole lesion segments were cross-analyzed (13).

Measurements of the regions containing the minimal luminal area (MLA) were acquired every 0.5 mm. The slice with the largest lumens within 10 mm of the MLA site and prior to any significant side branches was referred to as the proximal or distal reference segment. These measurements included EEM CSA and lumen CSA. EEM CSA minus lumen CSA was used to determine plaque plus media CSA, and percent plaque burden (PB) was calculated as (plaque plus media CSA/EEM CSA) × 100. The average area of the proximal/distal reference segment EEM CSAs was divided by the EEM CSA at the MLA site to determine the RI.

Vessel, lumen, and plaque volume analysis was calculated using Simpson's rule. Normalized total atheroma volume (TAV) was estimated using the following formula (14):

Normalized TAV = [Σ (EEM CSA−Lumen CSA)/Number of images in pullback] × median number of images in cohort.

In order to account for analysis length, the other volumetric parameters were also normalized.

Percent atheroma volume (PAV) was calculated according to this equation:PAV(%)=[Σ(EEMCSA−LumenCSA)ΣEEMCSA]×100.Gray-scale IVUS derived attenuated plaque was defined as IVUS images with backward signal attenuation of ≥180 degree behind plaque without dense calcium (15). The maxLCBI4mm was calculated as the maximum value of the lipid-core burden index (LCBI) for any 4-mm segment in the target lesion, including the MLA (Supplementary Methods).

Figure 1 presents an example of the patient who underwent CCTA and NIRS-IVUS.

The normality of the variables was assessed using the Shapiro–Wilk test. Continuous variables were compared using the Student's t-test or the Mann–Whitney U test, and were presented as mean ± SD or median and interquartile range (IQR). Chi-squared or Fisher's exact test was used to compare categorical variables that were presented as numbers and percentages. Differences in cTnT from the baseline value within each group were analyzed using Friedman test, and Bonferroni correction was added for post-hoc analysis. Spearman's correlation was utilized to analyze the relationship between the CCTA–measured parameters and the NIRS–IVUS–derived parameters. Inter- and intra-observer reproducibility on CCTA- and IVUS parameters were assessed using intraclass correlation coefficient. The interobserver reproducibility for CT density and maxLCBI4mm were 0.81 and 0.90 as Kappa coefficients, respectively. Moreover, the intraobserver Kappa coefficients for CT density and maxLCBI4mm were 0.86 and 0.92, respectively.

Multivariable logistic regression analysis was utilized to evaluate the relationship between periprocedural myocardial injury and clinical characteristics, angiographic characteristics, CCTA/NIRS-IVUS findings, and other potential confounders. Given the limited number of subclinical cTnT in this study, variables with p < 0.05 in the univariable analysis and clinically meaningful associations with periprocedural myocardial injury were included in the multivariable logistic regression model. To determine the best cutoff threshold of the independent risk factors for incidence of periprocedural myocardial injury, the receiver-operating characteristic (ROC) analysis was calculated. In the present study, the Youden Index was defined as the best cut-off threshold.

The statistical significance level was established at p < 0.05. All statistical analyses were performed using IBM SPSS version 28.0 (IBM Corp., Armonk, NY, United States).

A total of 107 patients were divided into two groups according to the maxLCBI4mm, using 400 as the cutoff value based on previous investigation (Figure 2) (8, 16). Of these patients, 48 (45.8%) were allocated to the LRP group (maxLCBI4mm ≥ 400) and 59 (54.2%) to the no-LRP group (maxLCBI4mm < 400).

Baseline characteristics and procedural findings are shown in Table 1. Between the groups, there were no significant differences in patient characteristics except for percentage of angiotensin-converting enzyme inhibitor/angiotensin II receptor blocker. Otherwise, larger number of patients with left anterior descending lesion was included in the LRP group according to the baseline lesion characteristics.

In terms of procedural findings, the LRP group received significantly longer stents (p = 0.043), whereas no significant differences in other parameters were observed between the groups.

Comparable baseline biomarker levels including cTnT (0.006 vs. 0.005 ng/ml; p = 0.86) were found for both groups (Table 1 and Figure 3). However, during the study period, the LRP group showed a significantly greater and faster increase in cTnT levels (0.019 vs. 0.010 ng/ml at 6 h, 0.035 vs. 0.016 ng/ml at 18–24 h; p = 0.026 for both). Consequently, there was a significantly higher proportion of periprocedural myocardial injury in the LRP group (13 patients [27.1%] vs. 6 patients [10.2%]; p = 0.023).

QCA and CCTA measurements are summarized in Table 2. The LRP group was characterized by larger RI and lower CT density than those in the no-LRP group (1.30 vs. 1.18, p = 0.020; 52.9 vs. 96.7, p < 0.001, respectively). Moreover, the prevalence of positive remodeling and low attenuated plaque were significantly greater in the LRP group than in the no-LRP group (87.5% vs. 66.1%, p = 0.010; 22.9% vs. 4.1%, p = 0.002, respectively). In terms of CCTA-derived plaque morphology, the LRP group had a higher rate of lesions with the napkin-ring sign (50.0% vs. 28.8%; p = 0.025).

Gray-scale IVUS and NIRS measurements are summarized in Table 3. At the MLA site, although there was no difference among groups in terms of gray-scale IVUS-measured lumen, vessel, and plaque areas, the LRP group had a larger percent PB and RI at the MLA site (79.9% vs. 77.2%, p = 0.029; 1.06 vs. 1.00, p < 0.001, respectively). Additionally, the percentage of patients with IVUS-attenuated plaques was significantly higher in the LRP group (54.2% vs. 15.3%, p < 0.001). Volumetric analysis also showed that PAV was significantly greater in the LRP group (58.4% vs. 54.9%; p = 0.036).

Figure 4 illustrates a significant negative linear correlation between CT density and maxLCBI4mm (rho = −0.552, p < 0.001).

Periprocedural myocardial injury was observed in 19 of the 107 patients (17.8%). According to the univariable analysis, QCA-measured lesion length, CT density, NIRS–IVUS-derived PAV, and maxLCBI4mm were significantly associated with periprocedural myocardial injury (Table 4). The strongest correlation between each of these modalities was between maxLCBI4mm and CT density, but there was also a significant correlation between the other factors, except for lesion length and CT density (Supplementary Table S1). Multivariable logistic regression analysis showed that the maxLCBI4mm (odds ratio [OR] 1.006, 95% confidence interval [CI] 1.002–1.011, p = 0.003) and PAV (OR 1.125, 95% CI, 1.025–1.235; p = 0.014) were independent predictors of periprocedural myocardial injury, whereas CT density was not an independent predictor (OR 1.010, 95% CI, 0.993–1.027; p = 0.25). CT density was not an independent predictor even when maxLCBI4mm was excluded from the model (OR 0.991, 95% CI, 0.977–1.005; p = 0.22).

We compared the incidence of periprocedural myocardial injury by classifying each case according to the maxLCBI4mm of 435 and the PAV of 58.4%, calculated by the ROC analysis [area under the curve (AUC) 0.751, sensitivity 68.4%, specificity 76.1%; AUC 0.763, sensitivity 78.9%, specificity 67.0%; respectively]. Figure 5 indicates periprocedural myocardial injury was most likely to occur in patients with higher maxLCBI4mm and greater PAV compared with those without (50.0% vs. 10.3% of frequency; OR 8.67; p < 0.001).

Periprocedural myocardial injury has been correlated to worse short- and long-term clinical outcome; therefore, various studies have previously identified high-risk plaques causing this “nightmare” by IVUS and CCTA (5, 17). Some studies have also revealed that the correlation between CCTA and gray-scale IVUS for LRP detection as well as consequently lower CT density was shown in the plaque with attenuation evaluated by IVUS (3, 18). Moreover, other CCTA studies have suggested that the napkin-ring sign was presented more frequently in high-risk plaques with a higher risk of ACS events (11).

A recent study showed that the group with a larger maxLCBI4mm had a smaller CT density and a higher proportion of napkin-ring signs (19). In addition, our results revealed that maxLCBI4mm has a linear inverse correlation with CT density.

Our study indicated that NIRS–IVUS-assessed maxLCBI4mm and PAV were significant risk factors for periprocedural myocardial injury. Previous studies clarified that the lesions with higher maxLCBI4mm were associated with a higher occurrence of periprocedural myocardial infarction (20, 21). Besides, previous optical coherence tomography study demonstrated that thin-cap fibroatheroma correlates well with lesions with large percentage PB, and an autopsy study clarified that the ability to identify fibroatheroma was significantly increased with the addition of NIRS-derived LCBI to gray-scale IVUS-derived PAV (22, 23). As a result of these findings, the combination of NIRS–IVUS-derived LRP and PAV might be a stronger predictor of periprocedural myocardial injury. In fact, our study revealed that the combination of increased plaque volume and a high maxLCBI4mm is a high-risk plaque with a higher incidence of periprocedural myocardial injury (Figure 5).

The NIRS-IVUS-derived maxLCBI4mm can only measure lipid distribution. Therefore, this fact also supports the hypothesis that the combination of maxLCBI4mm and PAV is useful for predicting periprocedural myocardial injury.

Our study showed that PAV and maxLCBI4mm are associated with periprocedural myocardial injury when PCI is performed with NIRS-IVUS in patients who have undergone prior CCTA. In other words, only PAV was a predictor of periprocedural myocardial injury in patients undergoing PCI with gray-scale IVUS alone.

Previous studies showed different values for the CT density threshold as the cutoff for LRP. Motoyama et al. demonstrated that a CT density of <30 HU indicated low-attenuation plaque according to the comparison of CCTA with gray-scale IVUS, whereas a pathological study revealed a cutoff value of low-attenuation plaque of <75 HU (10, 24).

In the case of CCTA, partial-volume effects due to the limitation of spatial resolution and motion artifacts caused by heartbeats and respiration could affect the measurement of CT density (10). Furthermore, if the lesion contains a calcified component, it may be difficult to assess the lumen and CT density by CCTA due to calcification artifacts (25). These features make it difficult to analyze plaque morphology as well as the volumetric measurement of the plaque in contact with the calcification. Moreover, this study focused on the culprit lesion, which has a large degree of stenosis, possibly making it difficult to accurately measure the lumen, especially with CCTA.

The PROSPECT2 study demonstrated that the high-risk plaques for non–culprit lesion-related MACE was maxLCBI4mm ≥324.7 and larger percent plaque area (9). In our study, a larger maxLCBI4mm and greater PAV were associated with periprocedural myocardial injury, indicating that patients with larger LRP and greater PAV are at higher risk for complications associated with PCI. This demonstrates that individuals with larger LRP and greater PAV are at high-risk not only for long-term prognosis for non-culprit plaques but also for perioperative complications associated with PCI. Our results suggest that assessing coronary plaque volume and plaque characteristics using NIRS–IVUS may be useful for predicting event in various situations, such as with both culprit and non-culprit lesions.

This study elucidated that CT density correlated well with maxLCBI4mm even in culprit lesions, implying that CCTA before PCI is useful for evaluating plaque characteristics. Recent research showed that aggressive lipid-lowering therapy (LLT) with statins and proprotein convertase subtilisin-kexin type 9 inhibitor has plaque-stabilizing and plaque-reducing effects on coronary plaque (26, 27). Therefore, more potent LLT options before PCI might be provided if CCTA reveals LRP in the target lesion. Subsequently, NIRS–IVUS can be used to evaluate plaque characteristics during PCI to determine the efficacy of preoperative drug treatment and to predict perioperative adverse events. Consequently, preoperative CCTA combined with NIRS–IVUS plaque assessment may contribute to further reduction of not only non–culprit lesion-related but also culprit lesion-related adverse events.

Accordingly, assessing maxLCBI4mm using NIRS-IVUS at the time of PCI can help predict perioperative complications, especially in patients who underwent CCTA prior to PCI and had low CT density in the culprit lesions.

This study has several limitations. First, bias is possible due to the relatively small number of patients. A larger patient population is needed to confirm the results. Second, we excluded patients who were ineligible for CCTA before PCI (e.g., ST-elevation MI). Therefore, we could not assess all of the high-risk plaques in clinical practice. Third, patients in which imaging modality other than NIRS-IVUS was selected at the operator's discretion were excluded from this study. This may have involved the operator's selection bias. Fourth, primary prevention with oral medication before PCI might have affected the results. And finally, multivariable logistic regression was performed to analyze independent predictors for periprocedural myocardial injuries. Since the number of cases with periprocedural myocardial injury was relatively small in this study, it is possible that parameters other than those identified in this study may also play a role as the predictor of periprocedural myocardial injury.