Introduction

Coronary artery disease (CAD) is one of the leading causes of death worldwide, despite all scientific advances [1]. Therefore, accurate diagnosis and effective planning are crucial. Electrocardiogram (ECG)-gated coronary computed tomography (CT) angiography, widely used to minimize the disadvantages of invasive imaging methods, plays a crucial role in the diagnosis of CAD. It aids in disease evaluation, early diagnosis and improvement of treatment processes [1].

In recent years, efforts have been made to establish a common communication language between cardiologists and radiologists through the Coronary Artery Disease Reporting and Data System (CAD-RADS) [2]. The main goals of CAD-RADS are to enhance communication between interpretive and guiding clinicians and to ensure better quality patients management through collaboration. In this system, severe stenosis is defined as 70% or more. Medical treatment and additional evaluation methods are recommended for patients with 50–69% stenosis, while invasive treatment methods are recommended for those with 70% and greater stenosis [2].

This study aims to evaluate the accuracy of the CAD-RADS scoring system, the compatibility between coronary computed tomography angiography (CCTA) and invasive coronary angiography (ICA), and the treatments applied by examining patients who were evaluated with CCTA and subsequently underwent ICA.

Materials and methods

Patient selection

The study was approved by the local ethics committee. A retrospective analysis was conducted on 2,120 patients who underwent CCTA in our radiology clinic for suspected CAD between January 1, 2022, and January 15, 2024. Patients examined with both CCTA and invasive coronary angiography (ICA) were included in the study (n: 246). Twenty-two patients who had a gap of more than one month between CCTA and ICA, and ten patients with artifact-laden coronary artery segments in CCTA images due to movement, cardiac arrhythmia, and respiratory artifacts (CAD-RADS N) were excluded from the study (128 vessels). Consequently, 214 patients and 586 vessels were analyzed.

CCTA procedure

In our clinic, routine CCTA examinations are performed using a dual-source CT scanner with a 128 × 2-slice capability and two X-ray tubes positioned at a 95° angle (Somatom Definition Flash, Siemens Healthcare, Germany). Initially, non-contrast images with a slice thickness of 3 mm are obtained according to the Agatston classification used for calcium scoring (Ca score), using a standard prospective ECG-triggered protocol at 120 kVp. During this process, both the calcium score is determined and detailed images that encompass the entire thorax are obtained using a wide field of view (FOV). This also facilitates the acquisition of extracardiac findings. Subsequently, in the CCTA examination, the scanning area, starting from the level of the carina, extends to the diaphragm surface of the heart based on images taken for the Ca score. The contrast material is administered in a triphasic manner. Patients included in the study underwent these protocols and were evaluated using the electronic data system.

ICA procedure

Invasive coronary angiography was performed using the transfemoral Judkins approach. During this procedure, the RCA and left main coronary artery branches were imaged in different projections. The cardiologist compared the stenoses in two different planes with a proximal normal segment and recorded the ratios.

Image analysis

Images were transferred to a workstation (Syngo Via; Siemens Medical Solutions) for analysis. Initially, the calcium score was calculated. Extracardiac findings were evaluated. Detailed images were obtained from thin-section axial data using two-dimensional maximum intensity projection (MIP), multi-planar reconstruction (MPR), and three-dimensional volumetric imaging techniques. MPR and MIP images were utilized for a detailed examination of arterial lumens, arterial walls, and heart chambers.. Three-dimensional volumetric imaging was preferred for a clearer depiction of the complex anatomical structure of the coronary arteries and stenoses. The best systolic and diastolic reconstructions were automatically identified at the workstation, facilitating a detailed examination of the coronary arteries with minimal artifacts. All images were evaluated by consensus by two radiologists, each with five years of experience in cardiovascular radiology, who were blinded to the ICA images.

CAD-RADS scoring

The assessment of coronary arteries utilized the CAD-RADS scoring criteria, categorizing based on the highest degree of stenosis observed in each patient. Stenosis classification was as follows: CAD-RADS 0 signifies no stenosis, CAD-RADS 1 indicates 0 to 24% stenosis, CAD-RADS 2 is 25 to 49% stenosis, CAD-RADS 3 ranges from 50 to 69% stenosis, CAD-RADS 4 entails 70% or more stenosis, and CAD-RADS 5 represents complete occlusion. Additional modifiers were applied to provide clarity; N for non-diagnostic, S for stent, G for graft, and HRP for high-risk plaque, defined by having at least two characteristics of HRP, including low attenuation plaque below 30 Hounsfield Units, positive remodeling, spotty calcification, or the “napkin ring” sign associated with low attenuation plaques [2].

LMCA, left anterior descending artery (LAD), left circumflex artery (LCX), and RCA were analyzed separately and similarly for vessel-based stenosis rates. Ca score was grouped as follows: 0, 1–100 = P1, 101–300 = P2, 301–999 = P3, > 999 = P4. Patients with a history of stent and bypass were not included in the Ca score evaluation. Treatments were classified as no treatment required, medical treatment, stent treatment, and surgery.

The diagnostic performances of CCTA and ICA were compared, and the effects of risk factors such as smoking, hypertension, dyslipidemia, and diabetes on Ca score and HRP diagnosis were examined. Additionally, non-contrast images covering the entire thorax with a wide FOV provided information about thoracic parenchyma, and triphasic scanning protocol provided insights into right heart chambers and pulmonary arteries. This allowed the evaluation of extracardiac and extracoronary cardiac findings.

Statistical analysis

Statistical analyzes were performed using SPSS Statistic software 24 (SPSS Inc., Chicago, III). Categorical variables were presented as numbers and percentages. The sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) of detecting coronary stenosis by CTA were compared with ICA by using a 2 × 2 cross-tabulation model. To evaluate the association between variables, the Chi-square test or Fisher exact test was applied. p < 0.05 was considered statistically significant.

Results

Patients' basic characteristics

In our study, 214 patients were evaluated, with 144 (67.3%) being male and 70 (32.7%) female. The median age was 56 years, ranging from 24 to 82. Detailed patient demographics are provided in Table 1. An analysis of Ca scoring revealed that 52 (24.3%) of the patients had a score of 0, 60 (28%) were scored between 1–100 (P1), 35 (16.4%) between 101–300 (P2), 30 (14%) between 301–1000 (P3), and 11 (5.1%) exceeded a score of 1000 (P4). Moreover, regarding the CAD-RADS evaluation, 18 (8.4%) patients were categorized as CAD-RADS 0, 42 (19.6%) as CAD-RADS 1, 46 (21.5%) as CAD-RADS 2, 35 (16.4%) as CAD-RADS 3, 57 (26.6%) as CAD-RADS 4 (Fig. 1), and 16 (7.5%) as CAD-RADS 5 (Fig. 2). Furthermore, 18 (8.4%) patients had a myocardial bridge (Fig. 3), and 32 (15%) had HRP. Of all the patients included in the study, 11 (5.1%) had one stent, 7 (3.3%) had two stents, 2 (0.9%) had three stents, and 1 (0.5%) had four stents. Additionally, 4 (1.9%) patients had two-vessel bypass grafts, and 4 (1.9%) had three-vessel bypass grafts.

Table 1 Demographic, Ca score, stent, and graft characteristics of patients
Fig. 1
figure 1

In the patient with a stent in the LAD, the stent is open but there is a CAD-RADS 4 lesion (arrow) (a: inverted image, b) in the distal part of the stent. In another patient, there is a CAD-RADS 4 lesion (arrow) in the LCX (d) and 3D image (e)

Fig. 2
figure 2

CCTA image (a, b: 3D image) and ICA image (c) of the patient with CAD-RADS 5 lesion (arrow) in the LCX show occlusion, and after the treatment the LCX is observed to be open (dashed arrow) (d)

Fig. 3
figure 3

The image of the patient with a myocardial bridge during diastole (a) and systole (b) shows narrowing in the ICA (arrow), the CCTA image (c) of the patient shows a deep myocardial bridge (arrow), and another patient has a malignant interarterial course in which the RCA originates from the left coronary sinus (d) (dashed arrow)

In the analysis, hypertension and smoking were found to be associated with CAD-RADS scoring (p = 0.006 and p = 0.002, respectively). Diabetes mellitus and dyslipidemia were not associated with CAD-RADS. Hypertension, smoking, diabetes mellitus, and dyslipidemia were all statistically significantly associated with Ca scores (p = 0.009, 0.019, 0.010, and 0.004, respectively).

Extracardiac and extracoronary findings detected in the wide FOV images obtained for the Ca score and during CCTA are presented in Table 2.

Table 2 Extracardiac and extracoronary findings

Vessel-based evaluation results

In our study, sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were calculated for vessel-based and patient-based analyses according to the 70% and 50% stenosis threshold values based on CAD-RADS 2.0. The results of the coronary vessel analysis for 856 vessels and the patient-based analysis are presented in Table 3.

Table 3 Coronary, vessel-based, and patient-based analysis

The results demonstrate that CCTA, especially in patients with high Ca scores, can detect stenoses with high sensitivity and accuracy. However, there are diagnostic challenges in patients with Ca scores P2 and P3, which have led to overdiagnosis (Table 4).

Table 4 Ca score analysis

HRP and Ca score analysis

Our study found that 32 patients (15%) had HRP (Fig. 4), and a significant relationship was found between vulnerable plaque and Ca score (p = 0.021). Additionally, HRP was significantly associated with both CCTA CAD-RADS score and ICA CAD-RADS score (p < 0.05). The median age of patients with vulnerable plaques was 56 (range: 35–73).

Fig. 4
figure 4

In the patient with CAD-RADS 4 lesion in the LAD; In coronal, sagittal and axial images (a), low-density HRP giving a napkin ring appearance (arrow), HRP showing positive remodeling in the RCA (b), CCTA (c) and ICA image of low-density HRP causing CAD-RADS 3 lesions (d) (arrow), nodule with irregular borders detected in the parenchymal window (dashed arrow) (e)

Bypass surgery analysis

In our study, 8 patients (3.7%) had bypass surgery, with a total of 20 grafts. The agreement between CCTA and ICA was excellent in patients with grafts, with 14 patients (70%) having stenosis. Sensitivity, specificity, PPV, and NPV were all 100%.

Stent analysis

In our study, 21 patients (9.8%) had stents, with a total of 35 stents (Fig. 5). Seven patients (20%) had stenosis in stented vessels. Sensitivity was 85.7%, specificity was 89.3%, PPV was 66.7%, and NPV was 96.2%.

Fig. 5
figure 5

In patients with stents, the stents are seen to be open (arrow) (a, b), in another patient with stents, a mixed type plaque causing 40% stenosis is also seen (c, d) (CAD-RADS 2)

CAD-RADS 3 analysis

We had 35 patients (16.4%) with CAD-RADS 3, of whom 31 (88.6%) were treated medically (median age: 59, male ratio: 61.3%), while 4 (11.4%) received stent treatment (median age: 64, male ratio: 75%). These patients were older and predominantly male, but the differences were not statistically significant. There were no significant differences in smoking, diabetes, hypertension, or dyslipidemia. All CAD-RADS 3 patients were monitored for major cardiac events; no major cardiac events were observed in the 31 medically treated patients (minimum 3 months, maximum 22 months, average 10 months) or in the 4 stent-treated patients (minimum 4 months, maximum 8 months, average 6 months).

Discussion

This study focuses on evaluating the compatibility between CCTA and ICA using CAD-RADS 2.0 [2] in the diagnosis of CAD. This study, conducted with a large patient group, concentrates on the differences in diagnostic performance between CCTA and ICA, the impact of the Ca score, and the significance of the presence of HRP, offering an innovative perspective on CAD diagnosis and treatment approaches.

CAD continues to be a significant health issue in both developed and developing countries and maintains its position among the leading causes of death despite advancements in diagnostic methods and treatment options [3, 4]. In our study, with 144 (67.3%) being male and 70 (32.7%) female. The median age of the patients was 56 (range: 24–82), similar to studies in the literature indicating that CAD is more common in males and at advanced ages [5,6,7].

According to CAD-RADS 2.0, Ca scores are classified as P1, P2, P3, and P4 [2]. Our study found a statistically significant association between the Ca score and CAD-RADS score, consistent with other studies on Ca scoringn [8,9,10,11,12,13]. Specifically, CCTA provides detailed information about the characterization of the vessel wall and plaques, offering advantages over ICA. It suggests that CCTA may be a more useful guide for clinicians in follow-up and treatment.

When comparing our findings with a study conducted in 2024 [14], some differences are observed. In our study, the proportion of participants with a zero coronary Ca score is lower at 24.3%, and non-zero Ca scores appear approximately 15 years earlier in men than in women. Additionally, high-risk categorization (75th percentile) for Ca scores is reached at the age of 37 for men and 41 for women, indicating an earlier onset compared to the age limits set by the previous study. These differences can be attributed to our study population, which includes individuals with a higher suspicion of CAD, undergoing both CCTA and ICA; this suggests that our findings are derived from a high-risk group and should be evaluated in this context.

Our study includes patients with stents and grafts. Similar diagnostic parameters were identified between CCTA and ICA in evaluating these patients. Particularly, CCTA has been shown to offer comparable results to ICA in assessing stent and graft lumens. Additionally, CCTA provides extra advantages in evaluating extracardiac findings, extracoronary findings, and the assessment of intimal walls and lesion locations [15,16,17,18,19,20,21].

Our study includes the evaluation of extracardiac findings (Table 2) as recommended by the most latest version of CAD-RADS [2]. The capability of CCTA to diagnose these findings presents a significant advantage over ICA. Particularly when cardiac findings do not explain the patient's symptoms, exploring alternative causes is essential for appropriate treatment. Potential other causes of chest pain include smoking-related lung diseases, interstitial lung diseases, malignancies, anatomical anomalies, and pulmonary embolism. These conditions can be easily detected especially with the non-contrast CT examination for Ca scoring and the triphasic scanning protocol of CCTA [22,23,24,25]. Given the diverse patient population encountered in university hospitals, such examinations are deemed essential. Therefore, the examination for Ca scoring was conducted with a wide FOV covering the entire thorax and using a triphasic CCTA scanning protocol. In the literature, there are limited studies advocating the use of a wide FOV instead of a cardiac-focused FOV for diagnostic and cost benefits [24,25,26]. Therefore, especially for patients at medium and high risk recommended for CCTA, we suggest that the images taken for Ca scoring be scanned with a wide FOV considering factors such as age and smoking habits. According to the new guidelines published by the American Cancer Society in 2023, annual lung cancer screening is recommended for individuals aged 50 to 80 who currently smoke or have quit in the past and have a history of 20 pack-years or more. This recommendation can be combined with the recommended screenings for Ca scoring, making it possible to conduct both screenings simultaneously [27].

Studies have examined with similar rates in the literature (In patient-based analysis, patients with CAD-RADS 3 and higher scores) have shown that CCTA successfully detects [28,29,30,31,32,33] CAD with high sensitivity and specificity rates. Also, the study by Nikolou and colleagues [34] has shown high diagnostic performance for stenoses of 75% and greater. CCTA's diagnostic performance, when compared to the reference standard ICA, has been indicated to show high sensitivity and specificity in meta-analyses with exercise ECG and single photon emission computed tomography (SPECT) [35,36,37]. A study published in 2024 has demonstrated significant diagnostic performance of CCTA in detecting significant stenoses (more than 50%) with high sensitivity, specificity, NPV, and PPV [38]. All these studies and our research emphasize the strong diagnostic capacity of CCTA in the diagnosis of CAD and its advantages compared to alternative diagnostic methods.

In graft evaluation, while ICA is considered the gold standard, it increases the radiation dose exposed to both the patient and the doctor. Sometimes these evaluations can be inadequate [38, 39]. In a study conducted with 100 patients with bypass using ICA and CCTA, similar to our study, all diagnostic performance parameters were obtained as 100% [40]. In another study conducted on 84 patients with grafts, sensitivity was reported as 97% and specificity as 100% [41]. A meta-analysis examining 12 studies evaluating patients with grafts found sensitivity and specificity of 98% [42]. In our study, there were 20 grafts, and the evaluation in patients with grafts found perfect agreement between CCTA and ICA, with sensitivity, specificity, PPV, and NPV all found to be 100%.

In our study, there are a total of 35 stents in 21 patients, with CCTA finding sensitivity of 85.7%, specificity of 89.3%, PPV of 66.7%, and NPV of 96.2%. A meta-analysis evaluating 35 studies on patients with stents found sensitivity of 90% and specificity of 94% for stenoses of 50% and greater [43].

In the literature, the presence of HRP, even in low-risk patients such as those with CAD-RADS 1 and 2, is suggested to require more aggressive preventive treatments. In the literature, CT Fractional Flow Reserve (CT FFR), CT perfusion (CTP), stress imaging or, if necessary, ICA are recommended as additional examination in patients with CAD-RADS 3 and HRP [2]. The presence of HRP is thought to triple a patient's risk of death or nonfatal myocardial infarction (MI) [11]. A study found that HRP was less common in women and the risk of myocardial infarction (MI) was lower. Additionally, HRP was found to increase the risk of MI, independent of Ca score (Ca score), obstructive disease, gender and cardiovascular risk factors (Odds Ratio: 1,6) [44]. A 2024 article [45] emphasized that the optimal management of HRP has not yet been determined and that this issue needs further discussion. In our study, we found that HRP is more common in men and the average age is consistent with the literature. 13 patients with HRP (5 patients with CAD-RADS 1 and 2) received medical treatment, and 19 patients received stent and surgical intervention (11 stents, 8 surgeries). It is believed that the treatment approaches and follow-up of these patients could make significant contributions to the literature.

In the literature, it is recommended to use functional evaluation methods such as CT-FFR, CTP, or stress tests to document or rule out the presence of ischemia in patients with CAD-RADS 3 scores [2]. Our study emphasizes the need for personalized treatment for CAD-RADS 3 patients, as stated in the literature [2]. This approach helps determine the necessity of invasive treatment. In our study, CADRADS 3 patients were questioned for major cardiac events. A major cardiac event was not observed in 31 patients who received medical treatment (minimum 3 months, maximum 22 months, average 10 months) or four patients who received stent treatment (minimum 4 months, maximum 8 months, average 6 months). This finding is important, and similar follow-up studies with larger numbers of CAD-RADS 3 patients are needed.

CCTA has important advantages such as identifying complications related to stents and grafts, recognizing non-atherosclerotic stenosis, and determining plaque burden and characterization. These features make CCTA a valuable imaging method for both doctors and patients.

Our study is generally compatible with the treatment recommendations stated in the CAD-RADS score. The presence of HRP, which can be easily evaluated by CCTA, may increase the severity and possible risks of CAD. This supports the recommendation of more aggressive treatment methods, such as invasive evaluation and potential revascularization, especially for patients with CAD-RADS scores of 3, 4 and 5. The detection of HRP plays an important role in the early and accurate identification of CAD and is critical in determining appropriate treatment strategies.

Our study offers a significant advantage in evaluating patients with CAD-RADS 2.0 and presenting a high number of patients compared to similar studies in the literature. However, our study has some limitations. Particularly, the fact that the study is single-center and retrospective is an important limitation. Other limitations include the low number of patients with HRP and CAD-RADS 3 scores and the lack of advanced methods such as CT-FFR in our unit. Especially, we believe that diagnosing, treating, and following up patients with HRP and CAD-RADS 3 scores with more extensive participation studies is necessary for managing these patients..

In conclusion, our study supports the importance of HRP in CAD management and the efficacy of CCTA in detecting these plaques, while also demonstrating compliance with the treatment recommendations based on the CAD-RADS score. Additionally, our study emphasizes the advantages of the imaging protocol used for Ca scoring, taken with a wide FOV and a triphasic scanning protocol. These findings provide significant contributions to the development of diagnosis and treatment strategies for CAD.