Background

Coronary artery disease (CAD) is a primary cause of mortality in the United States, and ranks as the third most common cause of death globally, responsible for 17.8 million deaths annually [1]. X-ray coronary angiography is utilized to diagnose CAD; however, it is an invasive procedure, and its complications cannot be overlooked. Presently, coronary computed tomography (CT) is widely employed in clinical practice as a non-invasive examination method. Coronary artery CT boasts a high negative predictive value and is efficacious in ruling out CAD [2]. Furthermore, there is evidence that evaluating coronary plaque [3] and implementing CT-based strategies can enhance prognosis [4]. Additionally, cost-effectiveness is also favorable in low to moderate prevalence rates [5]. Despite the utility of coronary CT being extremely high, it does have several drawbacks including radiation exposure, the administration of contrast agent, and difficulties in utilizing the method for highly calcified coronary arteries.

Whole heart coronary magnetic resonance angiography (WHCA) is considered as an alternative to coronary CT, possessing advantages over CT such as no radiation exposure, less susceptibility to coronary calcification [6]. Prior meta-analyses have been conducted on the diagnostic capabilities of WHCA [7, 8]. However, it should be noted that non-contrast imaging is recommended for 1.5 T WHCA, whereas contrast imaging is recommended for 3.0 T WHCA. The rationale behind this recommendation is that in 1.5 T WHCA, it is difficult to achieve increased arterial contrast with the administration of contrast agents [9]. In contrast, in 3.0 T WHCA, the use of steady state free precession (SSFP) is challenging due to banding artifacts caused by specific absorption ratio (SAR) limitations and B1 inhomogeneity. Therefore, it is generally advised to use gadolinium based contrast agents in gradient echo (GRE) sequence for 3.0 T WHCA [10]. To date, no meta-analysis has compared the diagnostic accuracy of non-contrast 1.5 T WHCA and contrast-enhanced 3.0 T WHCA. There is also debate about the need for premedication (vasodilators and beta-blockers) prior to imaging and differences in utilization by region (Western vs. Asian countries). These issues have not been evaluated in prior meta-analyses. Therefore, the purpose of this study was to perform a comprehensive meta-analysis on the diagnostic accuracy of WHCA for detecting significant CAD on X-ray coronary angiography and to evaluate the differences in magnetic field strength and use of contrast agent, with and without premedication, and including differences in diagnostic accuracy by region.

Methods

A systematic literature search was conducted in accordance with the guidelines established by the Cochrane Collaboration and the Preferred Reporting Items for Systematic Review and Meta-analysis (PRISMA) on November 10, 2022, utilizing databases such as PubMed, Web of Science Core Collection, Cochrane advanced search, and EMBASE. Search terms utilized included “whole heart coronary magnetic resonance angiography”, “WHCA”, “MRI”, “coronary artery disease”, “diagnostic accuracy” (as outlined in Additional file 1: Material S1). Two evaluators (SK and MA) independently assessed the validity of all titles and abstracts, followed by a review of the relevant complete peer-reviewed studies; any discrepancies were resolved by a third reviewer. The protocol for this study was registered with the University Medical Information Network (registration number: UMIN000050172) and did not require institutional review board approval as it was a meta-analysis and did not involve clinical patient information. Both prospective and retrospective studies that included diagnostic performance of coronary WHCA at 1.5 T and 3.0 T for detecting significant CAD on X-ray coronary angiography were included for data extraction, while literature such as case reports, animal studies, and non-English language articles were excluded.

Outcome measures

The primary objective of this meta-analysis was to estimate the diagnostic performance of coronary WHCA for significant coronary artery stenosis in known or suspected CAD using X-ray coronary angiography as the gold standard and to compare its value at 1.5 T and 3.0 T. Two reviewers (SK and MA) were invited to review the results of the studies, extracting the following study characteristics: author name, year of publication, country, patient disease, age, gender, magnetic resonance imaging (MRI) parameters such as magnetic field strength, sequence used, producer of MRI equipment, MRI coil information, use of gadolinium contrast, and examination time. Definition of significant CAD on X-ray coronary angiogram was also investigated. A meta-analysis of the diagnostic accuracy of coronary WHCA for significant coronary artery stenosis was performed using summary receiver operating characteristics (ROC) analysis. The analysis included the following. (1) All studies including the diagnostic performance of 1.5 T and 3.0 T were used to compare their diagnostic performance. The following sub-analyses were performed: 1. non-contrast enhanced 1.5 T WHCA vs. contrast enhanced 3.0 T WHCA, 2. drug administration (vasodilators and beta-blockers), 3. Comparison between Asian and Western countries. The Quality Assessment of Diagnostic Accuracy Studies-2 (QUADAS-2) were utilized to assess risk of bias [11].

Data integration and statistical analysis

Meta-analysis was conducted utilizing RevMan 5.41 (Cochrane Collaboration, London, UK) and R Statistical Software (v3.5.1, Boston, MA, USA). The diagnostic accuracy of WHCA was evaluated through summary ROC analysis. Three levels of diagnostic accuracy were analyzed: patient-based, vessel-based, and segment-based. Sensitivity and specificity were derived from ROC curves and the diagnostic performance was compared at magnetic field strengths of 1.5 T and 3.0 T. A random-effects model was employed to estimate imaging time by coil type. The inverse variance method was utilized to weight each study in the meta-analysis. Heterogeneity was indicated by I2, with 0% indicating no heterogeneity and 100% indicating strong heterogeneity [12]. P < 0.05 was considered statistically significant.

Results

Ultimately, 34 eligible papers were selected from a pool of 140 candidate papers, and data from 1899 patients were consolidated (Fig. 1) [13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45]. The characteristics of the included studies are summarized in Table 1. The publication years of these studies ranged from 2005 to 2022; 23 of these studies utilized 1.5 T MRI technology [13,14,15,16,17,18,19,20,21,22,23,24,25, 28, 30, 31, 33,34,35,36,37, 44,45,46], while 11 utilized 3.0 T technology [26, 27, 29, 32, 38,39,40,41,42,43, 45]. The countries of publication were diverse, with China having 11 reports [18, 26, 32, 34, 38, 39, 41,42,43,44,45], Germany having 8 reports [13, 15, 19, 25, 27, 29, 31, 35], Japan having 7 reports [14, 17, 24, 28, 30, 37, 40], United Kingdom [21, 46] having two reports and various other countries such as the United States [20], Korea [16], Switzerland [36], Portugal [33], Belgium [23], and Turkey [22] having one report each. In terms of study design, one study employed a prospective multicenter design [28], while 19 employed prospective single-center designs [13,14,15, 17, 21, 23, 25,26,27, 29,30,31,32, 37,38,39, 43,44,45,46]. The remaining studies were retrospective in nature. The MRI sequences utilized in these studies were steady state free precession [13,14,15,16,17,18,19,20,21,22,23,24,25, 28, 30, 31, 33,34,35,36,37, 40, 44, 46] or gradient-echo [26, 27, 29, 32, 38, 39, 41,42,43, 45]. Twenty-three studies demonstrated the diagnostic capability of 1.5 T WHCA [13,14,15,16,17,18,19,20,21,22,23,24,25, 28, 30, 31, 33,34,35,36,37, 44, 46], three of which used contrast [31, 33, 36] (Table 2). That is, the majority of studies (87%; 20/23) performed 1.5 T WHCA imaging without contrast. Eleven studies showed diagnostic performance of 3.0 T WHCA [26, 27, 29, 32, 38,39,40,41, 43, 45], of which two studies used non-contrast imaging [29, 45]. In other words, the majority of studies (82%, 9/11) evaluated the diagnostic performance of 3.0 T with gadolinium contrast administration. In terms of sequence used, all studies at 1.5 T used SSFP (100%, 23/23). In contrast, only one study at 3.0 T used SSFP [40], while the others used GRE-based sequencing (91%, 10/11). The information of pre-pulse and fat suppression was summarized in Table 2. The definition of significant CAD was luminal narrowing ≥ 50% in almost all studies (97%, 33/34). Only one study defined significant CAD as “luminal narrowing ≥ 90%, ≥ 50% in LMT, or FFR ≤ 0.80” [33]. In terms of cardiac coils, 32-channel coils were employed in 11 reports [29,30,31,32, 37, 39, 40, 42, 44,45,46]. Vasodilators such as sublingual nitroglycerin were utilized in 18 studies [14, 15, 17, 18, 24, 26, 27, 29,30,31,32,33,34,35, 37, 40, 41, 46], and beta-blockers were utilized in 12 studies [18, 19, 22, 26, 29, 32,33,34,35, 41, 43, 46]. The results of QUADAS-2 are summarized in Additional file 1: Material S2.

Fig. 1
figure 1

Preferred reporting items for systematic reviews and meta-analyses flow diagram

Full size image
Table 1 Characteristics of included studies
Full size table
Table 2 Information of magnetic resonance imaging
Full size table

Diagnostic accuracy of WHCA—1.5 T vs. 3.0 T

Figure 2 illustrates the summary receiver operating characteristic analysis of the ability of 1.5 T WHCA to detect significant stenosis when the gold standard is significant stenosis on X-ray coronary angiography. The area under the curve was 0.88 for patient-based analysis (N = 979 patients from 16 studies), 0.90 for vessel-based analysis (N = 2905 vessels from 15 studies), and 0.92 for segment-based analysis (N = 7171 segments from 16 studies). Figure 3 illustrates the summary receiver operating characteristic analysis of the ability of 3.0 T WHCA to detect significant stenosis. The area under the curve was 0.94 for patient-based analysis (N = 604 patients from 9 studies) and 0.95 for vessel-based analysis (N = 2032 vessels from 9 studies), and 0.96 for segment-based analysis (N = 4795 segments from 8 studies). Table 3 summarizes the sensitivity and specificity calculated from the receiver operating characteristic curves. The 3.0 T WHCA technique had significantly higher sensitivity in the segment-based analysis compared to 1.5 T (0.88, 95% confidence interval (CI) 0.84–0.91 vs. 0.80, 95% CI 0.72–0.86, P = 0.04). The individual diagnostic performance of each study is summarized in Additional file 1: Materials S3–S5.

Fig. 2
figure 2

Summary ROC curve of 1.5 T whole-heart coronary MRA

Full size image
Fig. 3
figure 3

Summary ROC curve of 3.0 T whole-heart coronary MRA

Full size image
Table 3 Sensitivity and specificity of whole heart coronary MRA for the detection of significant coronary stenosis on X-ray coronary angiogram
Full size table

Sub-analysis of diagnostic performance of WHCA

As previously stated, 1.5 T WHCA is typically performed without the use of gadolinium contrast, while 3.0 T WHCA is typically performed with contrast administration. Keeping this in mind, we conducted a comparison of the diagnostic performance of non-contrast 1.5 T WHCA and contrast-enhanced 3.0 T WHCA. The 3.0 T WHCA technique demonstrated significantly higher specificity compared to the non-contrast enhanced 1.5 T WHCA on a patient-based analysis (Table 4). Additionally, we performed a subgroup analysis based on the presence or absence of drug use and geographical difference (Asian and Western countries). There were no differences in diagnostic performance on a patient-based analysis were observed in the use of vasodilators and beta-blockers (Tables 5, 6) or comparison between Asian and Western countries (Table 7). However, in the vessel-based analysis, the sensitivity of studies using the vasodilator was lower than those not using it (P = 0.03) (Table 5). Mean heart rate in the study with beta-blocker administration was 64.6 bpm (95% CI 63.9–65.3 bpm) and in the study without beta-blocker administration mean heart rate was 68.6 bpm (95% CI 67.9–69.2 bpm), a significant difference was found between the two groups (P < 0.001). There was a significant difference in heart rate between trials with and without beta-blockers, but no difference in diagnostic performance. In the regional analysis, the body mass index (BMI) reported in Western countries was significantly higher than that in Asian countries (27.0 kg/m2, 95% CI 26.6–27.3 vs. 23.7 kg/m2, 95% CI 23.3–24.1, P < 0.001).

Table 4 Comparison of diagnostic performance of non-contrast 1.5 T WHCA and contrast-enhanced 3.0 T WHCA for the detection of significant CAD
Full size table
Table 5 A comparative analysis of sensitivity, specificity, and AUC between WHCA with and without vasodilators
Full size table
Table 6 A comparative analysis of sensitivity, specificity, and AUC between WHCA with and without beta-blockers
Full size table
Table 7 A comparative analysis of sensitivity, specificity, and AUC between Asian and Western countries
Full size table

Discussion

The main findings of this study are as follows: Receiver operating characteristic analysis showed that the 3 T field strength was superior in detecting significant coronary arteries compared to the 1.5 T. In addition, a direct comparison of non-contrast 1.5 T WHCA and contrast-enhanced 3.0 T WHCA was performed as a practical comparison, with the latter showing significantly higher specificity on a patient-based analysis. Subgroup analyses also showed no significant difference in diagnostic performance of 1.5 T WHCA in terms of the use of vasodilators and beta-blockers on a patient-based analysis. Although BMI was higher in Western patients compared to Asian patients, there was no difference in the diagnostic performance of the 1.5 T WHCA. These results suggest that WHCA is useful for noninvasive detection of significant CAD.

WHCA is well-established as a non-invasive method for the screening of CAD and possesses a number of advantages, such as the absence of ionizing radiation exposure, decreased susceptibility to calcification, and the lack of a requirement for contrast agent administration at 1.5 T. However, there is limited evidence for its clinical utility. To date, numerous studies have utilized X-ray coronary angiography as the gold standard, yet the majority of these studies have been conducted on small patient populations at a single institution. The only prospective, multi-center study was conducted in Japan and reported a sensitivity of 88% and specificity of 72% for 1.5 T WHCA [28]. Unfortunately, there have been no further multi-center studies since. Additionally, 3 T MR is often performed utilizing the gradient echo method, which necessitates the administration of a contrast agent [26, 27, 32, 38,39,40,41,42,43]. After contrast agent administration, 3 T WHCA provides a higher signal-to-noise ratio compared to 1.5 T WHCA and has been reported to have high diagnostic performance for the detection of coronary artery stenosis. However, one of the major advantages of MRI, the lack of requirement for contrast agent administration, is lost with 3 T WHCA. Recently, attempts have been made to perform non-contrast 3 T WHCA imaging, with promising results, but the number of reports on this technique is limited [29, 45]. The administration of a gadolinium contrast agent is necessary for 3 T WHCA due to the difficulties in using SSFP caused by SAR limitations and banding artifacts resulting from B1 inhomogeneity. Therefore, GRE is generally the preferred imaging sequence, but in order to achieve sufficient vascular contrast, the administration of gadolinium contrast is required [10]. As the imaging methods and diagnostic accuracy of 1.5 T and 3 T coronary WHCA are fundamentally different, separate meta-analyses are required. However, meta-analyses reported to date have included a mixture of 1.5 T and 3 T WHCA systems [7, 8]. Therefore, the primary objective of the present meta-analysis was to compare the diagnostic performance of WHCA with two different magnetic field strengths. Our results demonstrated that 3 T provided superior diagnostic performance when compared to 1.5 T, however, the number of reports regarding 3 T WHCA was small and there was a large bias in the countries and facilities where the studies were conducted (8/11 reports from China), making it difficult to generalize the obtained data. Further evidence accumulation and large-scale, prospective, multi-center studies are needed in the future to further investigate the diagnostic performance of 3 T MR. The clinical significance of the difference in diagnostic performance between 1.5 and 3.0 T WHCA is debatable. While 3.0 T WHCA exhibits slightly superior diagnostic performance, its major disadvantage of requiring the administration of gadolinium-based contrast agents negates its advantages over coronary CTA. Therefore, 1.5 T WHCA, which has unique benefits such as no radiation exposure and no need for gadolinium-based contrast administration, may be more clinically practical.

Another significant clinical query revolves around the necessity of nitroglycerin or beta-blockers in WHCA. Subgroup analyses of trials including and excluding both medications demonstrated comparable diagnostic performance in patient-based analyses, irrespective of drug usage (Tables 5, 6). Notably, there was no disparity in diagnostic performance, despite lower heart rates observed in studies employing beta-blockers. This could be attributed to the minimal absolute difference in heart rates (64.6 bpm vs. 68.6 bpm). Furthermore, although no distinctions were found in patient-based or segment-based analyses concerning vasodilator use, the sensitivity of studies employing vasodilators was lower than those that did not, as revealed by the vessel-based analysis (0.81 vs 0.91, P = 0.03, Table 5). Although the exact cause remains unclear, the vessel-based analysis exhibited higher AUC values for both groups, with an AUC of 0.90 for studies utilizing vasodilators and an AUC of 0.92 for studies without vasodilators. Given the trade-off relationship between sensitivity and specificity, the AUC does not appear to indicate a substantial disparity in diagnostic performance between studies with and without vasodilator use.

In addition, it is posited that coronary MRA is utilized by numerous institutions in Asian countries, with fewer employing it in Western countries. In light of this, we conducted a subgroup analysis of 1.5 T WHCA, taking into account the possibility of reduced diagnostic performance in larger patients due to their larger body size in Western countries (3.0 T studies could not be analyzed due to their small number). The findings indicated that BMI was significantly higher in patients from Western countries, but no significant differences in diagnostic performance were discerned between the two groups (Table 7). This suggests that WHCA can maintain its diagnostic efficacy even in patients with larger body mass.

The assessment of diagnostic efficiency between WHCA and coronary CT is a highly pertinent clinical inquiry. Nevertheless, there are few studies that have directly compared the two modalities. For instance, it has been demonstrated that WHCA is more diagnostically reliable than coronary CT in highly calcified segments of coronary arteries with calcification scores of 100 or above [18]. Conversely, other studies have found that 3 T WHCA has comparable diagnostic accuracy to CTA [43, 47]. Although such small-scale studies are dispersed, there is a lack of large-scale, coherent data, and it is challenging to statistically validate the comparison in this meta-analysis. Regardless, it is incontrovertible that coronary CT is the primary test for screening for CAD, owing to its spatial resolution, imaging duration, and reported high diagnostic accuracy. WHCA may serve as a viable alternative for patients for whom coronary CT cannot be performed, such as those with iodine allergies. Additionally, it should be utilized assertively in young patients who should not be exposed to radiation, female patients, and patients with coronary artery malformations [48] or coronary aneurysms in Kawasaki disease [49], which can be adequately evaluated with MRI resolution. Further accumulation of evidence on these points is also desirable.

Recent advancements in high-speed imaging techniques, such as compressed sensing, have the potential to shorten the imaging time for WHCA [50]. Furthermore, advancements in imaging techniques utilizing artificial intelligence are anticipated to enhance spatial resolution and decrease noise, thereby improving the image quality of WHCA. Deep learning reconstruction techniques have been used to improve the contrast-to-noise ratio and image quality of high-resolution WHCA [51]. A volunteer study has also demonstrated the potential of deep learning reconstruction for WHCA with sub-millimeter isotropic resolution at 3T [52]. These innovations in imaging technology are expected to further enhance the diagnostic accuracy of WHCA.

Limitations

First, many of the studies analysed were single centre studies with a limited number of cases, and the variability in study results cannot be ruled out. Prospective multicentre studies that include a larger number of patients are desirable. Second, we performed several subgroup analyses, but the number of included studies may be too small to produce statistically valid results.

Conclusions

The diagnostic performance of WHCA was deemed satisfactory, with contrast-enhanced 3.0 T WHCA exhibiting higher specificity compared to non-contrast-enhanced 1.5 T WHCA in a patient-based analysis. No significant differences in diagnostic performance were observed on a patient-based analysis based on the use of vasodilators, beta-blockers, or geographical regions (Asian and Western countries). Further large multicentre studies are imperative to facilitate the global adoption of WHCA.