Role of Quantitative Plaque Analysis and Fractional Flow Reserve Derived From Coronary Computed Tomography Angiography to Assess Plaque Progression : Journal of Thoracic Imaging

Secondary Logo

Journal Logo

Original Articles

Role of Quantitative Plaque Analysis and Fractional Flow Reserve Derived From Coronary Computed Tomography Angiography to Assess Plaque Progression

Qiao, Hong Yan MD*; Wu, Yong MSc*; Li, Hai Cheng BD; Zhang, Hai Yan BD; Wu, Qing Hua MD, PhD*; You, Qing Jun MD, PhD; Ma, Xin MD, PhD§; Hu, Shu Dong MD, PhD*

Author Information
Journal of Thoracic Imaging 38(3):p 186-193, May 2023. | DOI: 10.1097/RTI.0000000000000697



To explore the role of quantitative plaque analysis and fractional flow reserve (CT-FFR) derived from coronary computed angiography (CCTA) in evaluating plaque progression (PP).


A total of 248 consecutive patients who underwent serial CCTA examinations were enrolled. All patients’ images were analyzed quantitatively by plaque analysis software. The quantitative analysis indexes included diameter stenosis (%DS), plaque length, plaque volume (PV), calcified PV, noncalcified PV, minimum lumen area (MLA), and remodeling index (RI). PP is defined as PAV (percentage atheroma volume) change rate >1%. CT-FFR analysis was performed using the cFFR software.


A total of 76 patients (30.6%) and 172 patients (69.4%) were included in the PP group and non-PP group, respectively. Compared with the non-PP group, the PP group showed greater %DS, smaller MLA, larger PV and non-calcified PV, larger RI, and lower CT-FFR on baseline CCTA (all P<0.05). Logistic regression analysis showed that RI≥1.10 (odds ratio [OR]: 2.709, 95% CI: 1.447-5.072), and CT-FFR≤0.85 (OR: 5.079, 95% CI: 2.626-9.283) were independent predictors of PP. The model based on %DS, quantitative plaque features, and CT-FFR (area under the receiver-operating characteristics curve [AUC]=0.80, P<0.001) was significantly better than that based rarely on %DS (AUC=0.61, P=0.007) and that based on %DS and quantitative plaque characteristics (AUC=0.72, P<0.001).


Quantitative plaque analysis and CT-FFR are helpful to identify PP. RI and CT-FFR are important predictors of PP. Compared with the prediction model only depending on %DS, plaque quantitative markers and CT-FFR can further improve the predictive performance of PP.

The progression, erosion, and rupture of coronary plaque are closely related to cardiovascular events, which still remain the leading cause of death worldwide.1 In clinical practice, most patients with acute coronary syndrome are found only when atherosclerosis has advanced, which makes it difficult to implement highly targeted treatment and leads to major adverse cardiovascular events (MACE). Therefore, early and accurate identification of plaque progression (PP) is a great clinical unmet need. Serial imaging assessment of PP and its predictors is helpful to understand the natural history of coronary artery disease and guide patient-specific treatment to prevent future MACE.2–4 Invasive techniques such as intravascular ultrasound (IVUS) and optical coherence tomography (OCT) provide the ability to visualize and quantify plaque. However, these modalities are only recommended for patients with typical anginal symptoms and suspected or confirmed acute coronary syndrome because of their high cost and potential complications.5 Coronary computed tomography angiography (CCTA) is the preferred noninvasive imaging tool for the diagnosis and follow-up of coronary artery disease.6 With the rapid development of technology, the advanced post-processing functions provided by CCTA, such as quantitative plaque analysis, are highly consistent and correlated to IVUS.7

Recently, CCTA-derived fractional flow reserve (CT-FFR), as a technique for evaluating changes in coronary hemodynamics, can noninvasively identify lesion-specific ischemia and has become a hot spot in clinical research on cardiovascular diseases.8–10 The EMERALD study confirmed that CT-FFR was related to high-risk plaques and subsequent MACE.11 However, these studies on CCTA plaque or CT-FFR analysis are mostly cross-sectional.12,13 Longitudinal studies combining the above methods to evaluate risk predictors of PP are insufficiently studied.

In this study, we investigated the temporal changes of plaque features and the role of baseline CCTA plaque and CT-FFR analysis in assessing PP.


Patient Population

This retrospective study was approved by the local Institutional Review Board, and informed consent was waived. The study was compliant with the Health Insurance Portability and Accountability Act (HIPAA). We retrospectively analyzed consecutive patients who underwent serial CCTA due to worsening symptoms or routine follow-up of plaque status from January 2016 to December 2019. Patients with visible plaque on serial CCTA images were enrolled for this study. The exclusion criteria included the following: (1) coronary revascularization was performed between serial CCTA examinations, (2) the image quality of any CCTA could not meet the diagnostic requirements, (3) interval between two CCTA examination time was <1 y, and (4) patients who had incomplete clinical information. Patient characteristics (clinical risk factors, symptoms, serological examination, and medications) were collected by reviewing their electronic medical records and telephone interviews at each CCTA examination. The flowchart of this study is provided in Figure 1.

Study flowchart.

CCTA Studies

All CCTA scans were performed using dual-source CT (Somatom Definition/Flash, Siemens Healthineers, Forchheim, Germany) with the following parameters: tube voltage, 100 to 120 kVp; effective tube current, 370 mAs; detector collimation, 64×0.6 or 64×2×0.6 mm; and temporal resolution, 83 ms or 75 ms. All patients received sublingual nitroglycerin (0.1 mg per dose, Nitroglycerin Inhaler; Jingwei Pharmacy Co. Ltd., Jinan, China) 3 to 5 minutes before CCTA acquisition, unless contraindicated. Oral β-blockers (Betaloc ZOK, AstraZeneca) were administered to patients whose heart rates were ≥75 bpm. CCTA was performed using prospectively ECG-triggered acquisition. An intravenous nonionic contrast (ioversol, 370 mg/mL, Jiangsu Hengrui) (50 to 70 mL) was injected followed by 30 mL of saline at 4 to 5 mL/s. The system automatically reconstructed the optimal data with iterative reconstruction using a medium soft-tissue convolution kernel (B26F).

CCTA Quantitative Plaque Analysis

Quantitative plaque analysis was performed on a dedicated semiautomatic software as previously described (coronary plaque analysis 4.2.1, Siemens Healthcare).14 The images were interpreted by 2 cardiovascular radiologists who had 10 (H.Y.Q.) and 8 (Y.W.) years of experience, respectively, and were blinded to all patients’ characteristics. The software can automatically extract the coronary artery centerline, identify the inner and outer lumen, and correct it manually when needed. Different plaque compositions were identified using predefined Hounsfield unit (HU) threshold values: lipid-rich plaque (−100 to 30 HU), fibrous plaque (30 to 150 HU), and calcified plaque (150 to 350 HU). The following quantitative parameters were calculated: percentage diameter stenosis (%DS), plaque length, plaque volume (PV), calcified PV, noncalcified PV, minimum lumen area (MLA), remodeling index (RI), and percentage atheroma volume (PAV). RI=the vessel area of the lesion/the proximal reference vessel area. Noncalcified PV=fibrous PV+lipid-rich PV, whereas PV=noncalcified PV+calcified PV. PAV=PV/vessel volume×100. Fiduciary anatomic landmarks such as distance from the ostium were used to compare plaques of the same coronary segments in the baseline (CCTA-1) and follow-up scan (CCTA-2). The rate of PAV change=[(PAV CCTA-2−PAV CCTA-1)/PAV CCTA-1)]×100. PP was defined as the PAV change >1%.15,16 Total PAV, calcified PAV, and noncalcified PAV were analyzed. Annualized changes of plaque characteristics (%/year)=[(value of parameter at CCTA-2)−(value of parameter at CCTA-1)]/interval between 2 CCTA examinations.4 All analyses were performed on per-lesion and per-patient level.

CT-FFR Studies

CT-FFR calculations were blindly and independently performed using a prototype software cFFR (cFFR, version 3.1.1, Siemens). The software is based on an artificial intelligence deep learning platform for the noninvasive computation of FFR values, which has been described previously.10,17,18 In short, the platform was trained using a synthetically generated database of 12,000 different anatomies of coronary models, for which a reduced-order CFD model was utilized to calculate the CT-derived FFR. After importing the CCTA images into the software, it will automatically identify the coronary artery centerline and lumen, which could be manually adjusted if they were not initially optimized. Then, a colored coronary artery tree is generated. For each coronary artery, 2 observers prepared the data for the CT-FFR program with values obtained within 2 to 4 cm distal to a discrete coronary lesion (H.Y.Q. and Y.W. with 8 and 5 years of experience in CT-FFR analysis, respectively). When serial lesions were observed, only the lesion with the most severe stenosis was used. The lowest per-patient and per-lesion CT-FFR values were provided to the investigators.


Interobserver agreement in CCTA plaque and CT-FFR analysis was evaluated on serial CCTA images selected at random from 50 patients in the study population. Two readers (S.D.H. and H.Y.Q.) independently made measurements on the CCTA images, and they also performed CT-FFR analysis and measurement independently. To assess intraobserver variability, 30 cases were analyzed 8 weeks later by 1 observer blinded to the first results.


Events that occurred after the CCTA-2 were identified as MACE including all-cause death, nonfatal myocardial infarction, unstable angina requiring hospitalization, and urgent coronary revascularization. Data of MACE were obtained through medical record reviews, clinic visits, and telephone. Event adjudication was conducted by an independent Clinical Events Committee blinded to clinical and computed tomographic data using standard definitions.19

Statistical Analysis

Continuous variables were displayed as mean±SDs or median (interquartile range), as appropriate. Categorical data were presented as numbers or proportions. The χ2 test, Mann-Whitney U test, and Student t test were used as appropriate. Interobserver and intraobserver agreement of CCTA plaque and CT-FFR analysis was analyzed with intraclass correlation coefficient (ICC) analysis. The independent predictors of PP were assessed using multivariate logistic regression analysis. The risk factors analyzed in multivariate analysis were selected when the P value was ≤0.05 in the univariate analysis. The relative risks were expressed as odds ratios (OR) with 95% CIs. The area under the receiver-operating characteristic curve (AUC) using the method of Delong was used to compare the performance of the prediction models. MedCalc 3.0 (MedCalc Software, Ostend, Belgium) and IBM SPSS version 25.0 (SPSS, Chicago, IL) were used for the statistical analyses. A 2-sided level of <0.05 was considered statistically significant.


Patient Characteristics

Over a 4-year period from January 2016 to December 2019, a total of 506 patients who underwent serial CCTA with visible plaque were included. Among them, 258 patients were excluded for the following reasons: prior coronary revascularization (n=83), missing or insufficient CCTA images (n=56), interval between 2 CCTA exams <1 year (n=59), incomplete clinical information (n=27), and inadequate for CT-FFR or plaque analysis (n=33). Finally, this study included 248 patients with 354 lesions who had complete CCTA plaque and CT-FFR analysis (mean age 66.9±8.1 y, 49.2% male). The mean duration between the 2 CCTA examinations was 26±9 months ranging from 15 months to 48 months. Among them, 76 patients (30.6%) demonstrated PP during the follow-up period. The baseline clinical characteristics, symptoms, and medications of the 2 groups are presented in Table 1. The patients in the PP group were older (69.8 vs. 66.0 y, P=0.018), had a higher incidence of hyperlipidemia (43.4% vs. 29.7%, P=0.035), and had taken a smaller proportion of statins on admission (25.0% vs. 38.4%, P=0.041). No differences were observed in other baseline characteristics between the 2 groups (all P>0.05).

TABLE 1 - Patient Characteristics
Variables PP group (n=76) Non-PP group (n=172) P
Characteristics at CCTA-1
 Age, y 69.8±8.7 66.0±8.6 0.018
 Sex (male), n (%) 41 (53.9) 81 (47.1) 0.320
 BMI, mean±SD (kg/m2) 24.7±2.5 25.1±3.3 0.129
 Hypertension, n (%) 46 (60.5) 87 (50.6) 0.148
 Hyperlipidemia, n (%) 33 (43.4) 51 (29.7) 0.035
 Diabetes, n (%) 28 (36.8) 47 (27.3) 0.133
 Smoking, n (%) 20 (26.3) 40 (23.3) 0.604
 Family history of CAD, n (%) 25 (32.9) 43 (25.0) 0.199
 Total cholesterol, mg/dL 199.5±48.1 186.5±43.4 0.068
 TG, mg/dL 134.1±38.6 125.3±32.1 0.922
 LDL, mg/dL 115.9±27.7 112.1±18.6 0.210
 CRP, mg/dL 2.4 (0.7-3.7) 2.2 (0.8-3.9) 0.717
Symptoms at CCTA-1 0.393
 Typical angina, n (%) 25 (32.9) 53 (30.8)
 Atypical angina, n (%) 29 (38.2) 51 (29.7)
 Nonanginal chest pain, n (%) 9 (11.8) 25 (14.5)
 Dyspnea/palpitation, n (%) 13 (17.1) 43 (25.0)
Symptoms at CCTA-2 0.164
 Typical angina, n (%) 24 (31.6) 50 (29.1)
 Atypical angina, n (%) 31 (40.8) 53 (30.8)
 Nonanginal chest pain, n (%) 10 (13.2) 23 (13.4)
 Dyspnea/palpitation, n (%) 11 (14.5) 46 (26.7)
Medications at CCTA-1
 Aspirin, n (%) 25 (32.9) 51 (29.7) 0.609
 Second antiplatelet agent, n (%) 22 (28.9) 43 (24.4) 0.452
 Statins, n (%) 19 (25.0) 66 (38.4) 0.041
Medications at CCTA-2
 Aspirin, n (%) 28 (36.8) 55 (32.0) 0.454
 Second antiplatelet agent, n (%) 25 (32.9) 45 (26.2) 0.278
 Statins, n (%) 24 (31.6) 67 (39.0) 0.267
Values are mean±SD, n (%), or median (interquartile range).
BMI indicates body mass index; CAD, coronary artery disease; CCTA-1, baseline CCTA; CCTA-2, follow-up CCTA; CRP, C-reactive protein; LDL, low-density lipoprotein; TG, triglyceride.

Baseline and Follow-up Lesion Characteristics

The baseline and follow-up CCTA quantitative plaque parameters and CT-FFR between the 2 groups are presented in Table 2. The interobserver and intraobserver reproducibility was good for CCTA plaque (with ICC ranging from 0.739 to 0.977 for all variables) and CT-FFR analysis (ICC: 0.971 and 0.991) (Supplemental Table S1, Supplemental Digital Content 1,, all P<0.05). In all, 86 lesions (24.3%) demonstrated PP during the follow-up period. Compared with non-PP lesions, PP lesions had a greater %DS (43.3 vs. 40.0, P=0.039), smaller MLA (5.9 vs. 6.8 mm2, P=0.018), larger PV (272.9 vs. 228.6 mm3, P=0.008) and noncalcified PV (157.2 vs. 146.7 mm3, P=0.031), greater RI (1.12 vs. 1.02, P=0.001), and lower CT-FFR value (0.84 vs. 0.89, P<0.001) at baseline CCTA. No significant differences were observed in plaque length, PAV, and calcified PV on baseline CCTA between the 2 groups (all P>0.05). Similar results were observed in terms of annualized change of plaque characteristics between the 2 groups (Table 2). A representative case is illustrated in Figure 2.

TABLE 2 - Lesion Characteristics of the Baseline and Follow-up CCTA
Variables PP lesions (n=86) Non-PP lesions (n=268) P
 %DS 43.3±13.2 40.0±11.8 0.039
 Plaque length (mm) 26.9 [20.4, 42.2] 28.5 [21.1, 37.8] 0.755
 MLA (mm2) 5.9±2.9 6.8±3.4 0.018
 PV (mm3) 272.9 [208.3, 368.2] 228.6 [171.9, 316.1] 0.008
 Calcified PV (mm3) 96.8 [60.8, 140.2] 88.5 [55.8, 134.5] 0.107
 Noncalcified PV (mm3) 157.2 [121.1, 214.8] 146.7 [102.9, 190.3] 0.031
 RI 1.12 [1.04, 1.22] 1.02 [0.89, 1.19] 0.001
 Total PVA (%) 49.4±7.7 47.7±8.1 0.075
 Calcified PVA (%) 17.5±3.4 16.9±2.7 0.168
 Noncalcified PVA (%) 31.9±4.6 30.7±5.1 0.083
 CT-FFR 0.84 [0.78, 0.92] 0.89 [0.85, 0.93] <0.001
 %DS 47.2±11.1 39.7±11.5 <0.001
 Plaque length (mm) 27.8. [19.8, 43.7] 28.3 [20.9, 38.3] 0.818
 MLA (mm2) 5.7±2.8 6.8±3.3 0.006
 PV (mm3) 292.9 [214.1, 388.5] 230.6 [170.5, 322.5] 0.002
 Calcified PV (mm3) 108.2 [62.7, 152.9] 82.5 [54.9, 139.5] 0.060
 Noncalcified PV (mm3) 176.8 [127.6, 234.9] 152.6 [101.6, 194.2] 0.010
 RI 1.14 [1.05, 1.26] 1.01 [0.89, 1.19] <0.001
 Total PVA (%) 52.5 [46.7, 57.1] 48.2 [42.7, 54.5] 0.009
 Calcified PVA (%) 19.3 [14.5, 21.1] 17.2 [15.2, 20.5] 0.064
 Noncalcified PVA (%) 33.1 [29.8, 36.5] 30.9 [27.4, 35.0] 0.019
 CT-FFR 0.82 [0.76, 0.90] 0.89 [0.83, 0.93] <0.001
Annualized change of plaque characteristics
 MLA (mm2) −0.1±0.1 0.0±0.1 0.003
 PV (mm3) 10.3 [2.6, 18.9] 2.2 [−0.4, 3.2] 0.001
 Calcified PV (mm3) 1.4 [0.2, 2.9] 1.2 [−0.4, 2.5] 0.052
 Noncalcified PV (mm3) 7.8 [3.7, 10.5] 1.0 [−0.6, 2.0] 0.008
 Total PVA (%) 1.5 [0.2, 2.5] 0.1 [−0.3, 0.4] <0.001
 Calcified PVA (%) 0.7 [0.1, 1.2] 0.0 [−0.3, 0.2] 0.003
 Noncalcified PVA (%) 0.9 [0.1, 1.8] 0.1 [−0.2, 0.3] <0.001
 CT-FFR −0.01 [−0.02, 0.0] 0.0 [0.0, 0.0] 0.102
CCTA-1 indicates baseline CCTA; CCTA-2, follow-up CCTA; PAV, percentage atheroma volume; PV, plaque volume.

Representative case of PP. A–D, A 58-year-old man with typical chest pain found to have mild stenosis in the proximal mid-segment of the left anterior descending artery (LAD) on CCTA-1 (white arrow) (A). (B) and (C) show quantitative plaque analysis based on CCTA. (D) shows that the CT-FFR value at a distance of 4 cm from the plaque was 0.85. E–H, CCTA-2 examination (26 mo follow-up) showed significant plaque progress (yellow arrow in E). (F) and (G) show significant PP on quantitative plaque analysis. (H) shows that the CT-FFR value at a distance of 4 cm from the plaque was 0.70.

Association of CCTA Plaque Markers, CT-FFR, and PP

In a univariate analysis, age (OR: 1.036, 95% CI: 1.001-1.073, P=0.046), hypercholesterolemia (OR: 1.821, 95% CI: 1.041-3.185, P=0.036), statins (OR: 0.535, 95% CI: 0.292-0.979, P=0.042), %DS (OR: 1.082, 95% CI: 1.006-1.050, P=0.011), MLA (OR: 0.892, 95% CI: 0.816-0.976, P=0.013), PV (OR: 1.004, 95% CI: 1.001-1.006, P=0.002), noncalcified PV (OR: 1.005, 95% CI: 1.002-1.009, P=0.002), PAV (OR: 2.000, 95% CI: 1.152-3.472, P=0.014), RI (OR: 3.025, 95% CI: 1.730-5.291, P<0.001), and CT-FFR (OR: 6.459, 95% CI: 3.558-11.725, P<0.001) were all associated with PP. However, after multivariate adjustment, RI (OR: 2.709, 95% CI: 1.447-5.072, P=0.002) and CT-FFR (OR: 5.079, 95% CI: 2.626-9.283, P<0.001) were independent predictors of PP (Table 3).

TABLE 3 - Predictors of Coronary PP
Univariate analysis Multivariate analysis
Variables OR 95% CI P OR 95% CI P
Age 1.036 1.001-1.073 0.046 1.022 0.982-1.065 0.288
Hypercholesterolemia 1.821 1.041-3.185 0.036 1.603 0.831-3.093 0.159
Statins 0.535 0.293-0.979 0.042 0.620 0.312-1.231 0.172
%DS 1.082 1.006-1.050 0.011 1.011 0.982-1.040 0.458
MLA 0.892 0.816-0.976 0.013 0.938 0.839-1.049 0.261
PV 1.004 1.001-1.006 0.002 0.999 0.993-1.005 0.813
Noncalcified PV 1.005 1.002-1.009 0.002 1.004 0.996-1.012 0.297
Total PAV≥49* 2.001 1.152-3.472 0.014 0.879 0.438-1.767 0.718
RI≥1.10* 3.025 1.730-5.291 <0.001 2.709 1.447-5.072 0.002
CT-FFR≤0.85* 6.459 3.558-11.725 <0.001 5.079 2.626-9.283 <0.001
*Optimal cut-off value of PAV, RI, and CT-FFR determined by the Yorden index of ROC curve analysis.
PAV indicates percentage atheroma volume; PV, plaque volume.

Prediction Models of Coronary PP

Three models for predicting PP were established based on different baseline CCTA parameters (Fig. 3): model 1: based on %DS, AUC=0.61, 95% CI: 0.54-0.67, P=0.007; model 2: based on %DS+quantitative plaque characteristics, AUC=0.72, 95% CI: 0.67-0.78, P<0.001; model 3: based on %DS+quantitative plaque characteristics+CT-FFR, AUC=0.80, 95% CI: 0.74-0.95, P<0.001. Comparison of the 3 models showed that model 1 versus model 2, P=0.002; model 1 versus model 3, P<0.001; model 2 versus model 3, P=0.006.

Comparison of different models for PP. Model 1: ROC curve based on CCTA stenosis degree (blue line); model 2: ROC curve based on CCTA stenosis degree and quantitative plaque characteristics (green line); model 3: ROC curve based on CCTA stenosis degree, quantitative plaque characteristics, and CT-FFR (orange line).


During a median follow-up of 30 months after CCTA-2, MACEs occurred in 12.1% (30/248) of patients, urgent coronary revascularization was the most components with 40% (12/30) of the observed events (Table 4). Patients in the PP group had a higher rate of MACE than those in the non-PP group (18.4% vs. 9.3%, P=0.042), which was mainly driven by urgent coronary revascularization (9.2% vs. 2.9%, P=0.033).

TABLE 4 - MACE at Follow-up According to PP
Variables PP group (n=76) Non-PP group (n=172) P
MACE, n (%) 14 (18.4) 16 (9.3) 0.042
All-cause death, n (%) 1 (1.3) 2 (1.2) 0.919
Nonfatal myocardial infarction, n (%) 2 (2.6) 3 (1.7) 0.647
Hospitalization for unstable angina, n (%) 4 (5.3) 6 (4.5) 0.512
Urgent coronary revascularization 7 (9.2) 5 (2.9) 0.033


In this longitudinal observational study, we found that baseline quantitative plaque analysis and CT-FFR were used to identify PP. RI and CT-FFR were independent predictors of PP. Furthermore, compared with the prediction model that only depends on stenosis degree, the model that combined CT-FFR and quantitative plaque parameters could further improve the predictive value of PP.

Clinical Value of Identifying PP Based on CCTA

The progression of coronary atherosclerosis is a complex process affected by numerous factors such as local microenvironment, clinical comorbidity, and medication use.20–22 PP is a necessary and modifiable step in the development of atherosclerotic plaque, so early identification of PP that be assessed by imaging has important clinical implications for identifying high-risk patients and intensifying medical treatment.23 Our follow-up data showed that the MACE rate was high in the PP group (18.4% vs. 9.3%), indicating the need to analyze baseline characteristics to accurately predict PP and therefore to improve the outcome. It remains a challenge by invasive modalities such as IVUS, OCT. CCTA can not only noninvasively evaluate coronary artery stenosis degree, but also reasonably evaluate plaque composition, remodeling, and characteristics and identify high-risk plaque signs such as the napkin-ring sign.24 CCTA plaque assessment has high consistency with IVUS and is a feasibly repeatable method with low interobserver variability.7,25 Our study also showed good intraobserver and interobserver agreement for CT-FFR and plaque measurements. However, most previous studies have been cross-sectional and focused on differentiating at-risk groups.26–28 To our knowledge, this study is the first to combine CT-FFR and quantitative plaque features to predict the future risk of PP. In addition, recent studies have shown that total PAV is a powerful tool for predicting PP.16,29,30 Thus, we used the patient-level PAV change rate to measure PP, which is helpful to comprehensively evaluate patient’s treatment efficacy, adjust management strategy, and improve clinical prognosis.

CCTA Plaque Quantitative Parameters as Predictors of PP

In accordance with the study by Nakanishi et al,31 our study found that plaque quantitative parameters based on baseline CCTA, including DS%, MLA, PV, noncalcified PV, and RI, were significantly different in the PP group. Moreover, RI ≥1.10 (positive remodeling) is an independent risk factor for PP. Previous studies have also confirmed that positive remodeling is one of the high-risk plaque signs, and the possible mechanism is that oxidative stress, inflammation, and other factors cause changes in the coronary microenvironment such as endothelial shear stress, which leads to lumen remodeling and PP.32 However, PAV was not a significant predictor of PP in our study. The possible reason was that the baseline PAV was relatively low (<50%), which was not statistically different between the two groups. In addition, our results further showed that patients in the PP group were older and had more hyperlipidemia. Studies have shown that statin therapy can reduce the lipid PV, slow down and halt PP, and reduce the incidence rate of MACE.33,34 This study supports the necessity of early monitoring of blood lipids and statin treatment in elderly patients, so as to reduce the incidence rate of MACE.

CT-FFR as a Predictor of PP

CT-FFR can accurately assess the hemodynamic relevance of coronary stenosis validated against invasive FFR.35 A CT-FFR value ≤0.80 is generally considered an ischemic lesion. Recently, CT-FFR based on machine learning algorithm can be used to evaluate lesion-specific ischemia with a short calculation time, which is shorter than the CFD algorithms varying from 10 minutes to several hours.35 CCTA can evaluate the stenosis degree of coronary plaque, but it is often inconsistent in gauging the hemodynamic significance of lesions.36 Mild or moderate stenosis plaques can progress and further cause MACE.37 Therefore, the impact of patients’ hemodynamic information on PP needs to be analyzed. Lee et al11 reported that hemodynamic parameters based on CT-FFR can improve the efficiency of predicting high-risk culprit plaque only by CCTA anatomical parameters. Our study extends a previous study to explore the predictive value of baseline CT-FFR in the process of PP. The results show that CT-FFR is lower in the PP group, and CT-FFR≤0.85 can independently predict PP. The possible mechanism is that the reduction of CT-FFR may cause the instability of endothelial function, change the microenvironment, and promote the progression and rupture of plaque.38 The threshold of CT-FFR for predicting PP in this study was 0.85, which was different from predicting ischemic lesions. A possible explanation is that the DS% of baseline CCTA plaque is low (<50%) in our study. In addition, adding CT-FFR functional indicators can improve the prediction performance relying only on DS% (AUC: 0.61 vs. 0.72). The CCTA multiparameter model that combined CT-FFR and plaque quantitative analysis is a comprehensive means to evaluate PP, which can improve risk stratification and promote clinicians to carry out timely and effective preventive treatment.


Our study had several limitations: First, this was a single-center retrospective study, with a relatively small sample size, and some selective bias may be present. Second, there is a relatively big range of duration between serial CT examinations, which could be a potential time-dependent bias. Third, invasive FFR, IVUS, and OCT were not performed in this study. However, the CCTA plaque quantitative software and CT-FFR have been well verified in previous studies.7,39 Fourth, some clinical risk factors, such as lipid levels, were not included in the prediction model, which may be relevant to the retrospective observational design; future prospective studies are needed to further evaluate the role of these factors in PP. Fifth, our study mainly analyzed the clinical risk factors and drug treatment at baseline CCTA. We would not have fully accurately assessed the effect of drug changes on PP, which will be further explored in a future work. Sixth, although CT-FFR is proven to have high reproducibility, a small margin of baseline CT-FFR value between the two groups (0.84 vs. 0.89) needs to be further verified by prospective large sample studies to enhance its clinical application.

In conclusion, our findings revealed that plaque quantitative analysis based on serial CCTA is helpful in identifying patients with PP, and the combination of CT-FFR with quantitative CCTA indicators can further improve the predictive ability of PP.


1. Tsao CW, Aday AW, Almarzooq ZI, et al. Heart disease and stroke statistics-2022 update: a report from the American Heart Association. Circulation. 2022;145:e153–e639.
2. Stone GW, Maehara A, Lansky AJ, et al. A prospective natural-history study of coronary atherosclerosis. N Engl J Med. 2011;364:226–235.
3. Thomsen C, Abdulla J. Characteristics of high-risk coronary plaques identified by computed tomographic angiography and associated prognosis: a systematic review and meta-analysis. Eur Heart J Cardiovasc Imaging. 2016;17:120–129.
4. Lee SE, Chang HJ, Sung JM, et al. Effects of statins on coronary atherosclerotic plaques: the PARADIGM study. JACC Cardiovasc Imaging. 2016;11:1475–1484.
5. Yamamoto MH, Yamashita K, Matsumura M, et al. Serial 3-vessel optical coherence tomography and intravascular ultrasound analysis of changing morphologies associated with lesion progression in patients with stable angina pectoris clinical perspective. Circ Cardiovasc Imaging. 2017;10:e006347.
6. Knuuti J, Wijns W, Saraste A, et al. 2019 ESC Guidelines for the diagnosis and management of chronic coronary syndromes. Eur Heart J. 2020;41:407–477.
7. Graaf MA, Broersen A, Kitslaar PH, et al. Automatic quantification and characterization of coronary atherosclerosis with computed tomography coronary angiography: cross-correlation with intravascular ultrasound virtual histology. Int J Cardiovascr Imaging. 2013;29:1177–1190.
8. Zhuang B, Wang S, Zhao S, et al. Computed tomography angiography-derived fractional flow reserve (CT-FFR) for the detection of myocardial ischemia with invasive fractional flow reserve as reference: systematic review and meta-analysis. Eur Radiol. 2020;30:712–725.
9. Benton SM Jr, Tesche C, De Cecco CN, et al. Noninvasive derivation of fractional flow reserve from coronary computed tomographic angiography: a review. J Thorac Imaging. 2018;33:88–96.
10. Qiao HY, Tang CX, Schoepf UJ, et al. One-year outcomes of CCTA alone versus machine learning-based FFRCT for coronary artery disease: a single-center, prospective study. Eur Radiol. 2022;32:5179–5188.
11. Lee JM, Choi G, Koo BK, et al. Identification of high-risk plaques destined to cause acute coronary syndrome using coronary computed tomographic angiography and computational fluid dynamics. JACC Cardiovasc Imaging. 2019;12:1032–1043.
12. Velangi PS, Maharaj V, Athwal SS, et al. Computed tomography coronary plaque characteristics predict ischemia detected by invasive fractional flow reserve. J Thorac Imaging. 2021;36:360–366.
13. Yin P, Dou G, Yang X, et al. Noninvasive quantitative plaque analysis identifies hemodynamically significant coronary arteries disease. J Thorac Imaging. 2021;36:102–107.
14. Yang L, Xu PP, Schoepf UJ, et al. Serial coronary CT angiography–derived fractional flow reserve and plaque progression can predict long-term outcomes of coronary artery disease. Eur Radiol. 2021;31:7110–7120.
15. Papadopoulou SL, Neefjes LA, Garcia-Garcia HM, et al. Natural history of coronary atherosclerosis by multislice computed tomography. JACC Cardiovasc Imaging. 2012;5:S28–S37.
16. van Rosendael AR, Lin FY, Ma X, et al. Percent atheroma volume: optimal variable to report whole-heart atherosclerotic plaque burden with coronary CTA, the PARADIGM study. J Cardiovasc Comput Tomogr. 2020;14:400–406.
17. Qiao HY, Tang CX, Schoepf UJ, et al. Impact of machine learning-based coronary computed tomography angiography fractional flow reserve on treatment decisions and clinical outcomes in patients with suspected coronary artery disease. Eur Radiol. 2020;30:5841–5851.
18. Tang CX, Qiao HY, Zhang XL, et al. Functional CAD-RADS using FFRCT on therapeutic management and prognosis in patients with coronary artery disease. Eur Radiol. 2022;32:5210–5221.
19. Hicks KA, Tcheng JE, Biykem B, et al. 2014 ACC/AHA key data elements and definitions for cardiovascular endpoint events in clinical trials: a report of the American College of Cardiology/American Heart Association Task Force on Clinical Data Standards. J Nucl Cardiol. 2015;22:1041–1144.
20. Sakakura K, Nakano M, Otsuka F, et al. Pathophysiology of atherosclerosis plaque progression. Heart Lung Circ. 2013;22:399–411.
21. Mushenkova NV, Summerhill VI, Zhang D, et al. Current advances in the diagnostic imaging of atherosclerosis: insights into the pathophysiology of vulnerable plaque. Int J Mol Sci. 2020;21:2992.
22. Stone PH, Maehara A, Coskun AU, et al. Role of low endothelial shear stress and plaque characteristics in the prediction of nonculprit major adverse cardiac events: the PROSPECT study. JACC Cardiovasc Imaging. 2018;11:462–471.
23. Ahmadi A, Argulian E, Leipsic J, et al. From subclinical atherosclerosis to plaque progression and acute coronary events. J Am Coll Cardiol. 2019;74:1608–1617.
24. Maurovich-Horvat P, Ferencik M, Voros S, et al. Comprehensive plaque assessment by coronary CT angiography. Nat Rev Cardiol. 2014;11:390–402.
25. Fischer C, Hulten E, Belur P, et al. Coronary CT angiography versus intravascular ultrasound for estimation of coronary stenosis and atherosclerotic plaque burden: a meta-analysis. J Cardiovasc Comput Tomogr. 2013;7:256–266.
26. Han D, Starikov A, ÓH B, et al. Relationship between endothelial wall shear stress and high‐risk atherosclerotic plaque characteristics for identification of coronary lesions that cause ischemia: a direct comparison with fractional flow reserve. J Am Heart Assoc. 2016;5:e004186.
27. Bittner DO, Mayrhofer T, Puchner SB, et al. Coronary computed tomography angiography-specific definitions of high-risk plaque features improve detection of acute coronary syndrome. Circ Cardiovasc Imaging. 2018;11:e007657.
28. Obaid DR, Calvert PA, Brown A, et al. Coronary CT angiography features of ruptured and high-risk atherosclerotic plaques: correlation with intra-vascular ultrasound. J Cardiovasc Comput Tomogr. 2017;11:455–461.
29. Kim U, Leipsic JA, Sellers SL, et al. Natural history of diabetic coronary atherosclerosis by quantitative measurement of serial coronary computed tomographic angiography: results of the PARADIGM study. JACC Cardiovasc Imaging. 2018;11:1461–1471.
30. Won KB, Lee SE, Lee BK, et al. Longitudinal assessment of coronary plaque volume change related to glycemic status using serial coronary computed tomography angiography: a PARADIGM substudy. J Cardiovasc Comput Tomogr. 2019;13:142–147.
31. Nakanishi R, Ceponiene I, Osawa K, et al. Plaque progression assessed by a novel semi-automated quantitative plaque software on coronary computed tomography angiography between diabetes and non-diabetes patients: a propensity-score matching study. Atherosclerosis. 2016;255:73–79.
32. Tomaniak M, Katagiri Y, Modolo R, et al. Vulnerable plaques and patients: state-of-the-art. Eur Heart J. 2020;41:2997–3004.
33. Lee SE, Sung JM, Andreini D, et al. Differential association between the progression of coronary artery calcium score and coronary plaque volume progression according to statins: the PARADIGM study. JACC Cardiovasc Imaging. 2019;20:1307–1314.
34. Budoff MJ, Bhatt DL, Kinninger A, et al. Effect of icosapent ethyl on progression of coronary atherosclerosis in patients with elevated triglycerides on statin therapy: final results of the EVAPORATE trial. Eur Heart J. 2020;41:3925–3932.
35. Tang CX, Wang YN, Zhou F, et al. Diagnostic performance of fractional flow reserve derived from coronary CT angiography for detection of lesion-specific ischemia: a multi-center study and meta-analysis. Eur J Radiol. 2019;116:90–97.
36. Ciccarelli G, Barbato E, Toth GG, et al. Angiography versus hemodynamics to predict the natural history of coronary stenoses: fractional flow reserve versus angiography in multivessel evaluation 2 substudy. Circulation. 2018;137:1475–1485.
37. Imai S, Kondo T, Stone GW, et al. Abnormal fractional flow reserve in nonobstructive coronary artery disease. Circ Cardiovasc Interv. 2019;12:e006961.
38. Nozue T, Takamura T, Fukui K, et al. Changes in coronary atherosclerosis, composition, and fractional flow reserve evaluated by coronary computed tomography angiography in patients with type 2 diabetes. Int J Cardiol Heart Vasc. 2018;19:46–51.
39. Gaur S, Ovrehus KA, Dey D, et al. Coronary plaque quantification and fractional flow reserve by coronary computed tomography angiography identify ischaemia-causing lesions. Eur Heart J. 2016;37:1220–1227.

coronary artery disease; computed tomography angiography; fractional flow reserve; plaque progression

Supplemental Digital Content

Copyright © 2023 The Author(s). Published by Wolters Kluwer Health, Inc.