Introduction

Coronary artery disease (CAD) remains one of the most important causes of morbidity and mortality among women [1]. Although overall cardiac mortality has declined on average 3 % per year, with rates in men have decreasing substantially, little change has been reported in women [1]. As a result, interest and emphasis on research concerning women and heart disease have grown considerably, along with an increasing recognition of the importance of heart disease in relation to the female gender. However, a worrying gap persists in knowledge, understanding and general awareness of ischaemic heart disease in women. Therefore specific diagnostic testing algorithms are necessary for early identification of at-risk women [2, 3].

Research studies in large cohorts of women undergoing single photon emission computed tomography (SPECT) myocardial perfusion imaging have shown an independent and incremental value for both diagnostic and prognostic applications [4, 5]. The possibility of identifying women at low and high risk with myocardial SPECT has been described in several reports, including >1,000 women [6–8]. Building on this result, a recent consensus paper from the American Society of Nuclear Cardiology suggested a strategy based on this technology, albeit providing little supportive clinical trial evidence [9]. In fact, published studies have demonstrated the impact of myocardial perfusion imaging by standard SPECT for the evaluation of women with known or suspected ischaemic heart disease on target intervention and clinical treatment, but they do not include current studies with the new CZT camera [10–12]. The use of a lower radiation dose while maintaining high sensitivity and specificity could not only increase patient safety but also help identify ischaemia in this complex population.

The purpose of the present study was therefore to evaluate the diagnostic accuracy of CZT myocardial perfusion imaging in relation to gender in a prospective database of patients with known or suspected CAD, using coronary angiography as the reference standard.

Material and methods

Patients

Between September 2009 and November 2012 a total of 2,673 patients, of whom 1,997 (75 %) were men, were studied with CZT imaging. From these, we enrolled a group of 309 consecutive patients (248 men, 80 %) admitted for known or suspected CAD, referred to our Institution for stress–rest SPECT and who underwent coronary angiography up to 1 month after the evaluation of ischaemia with CZT imaging. Patients with acute or recent (<3 months) ST segment elevation myocardial infarction and unstable angina were excluded from the study. The study was approved by the local ethics committee.

Stress protocols

Patients were instructed to discontinue beta-blockers, calcium antagonists and nitrates 48 h before testing. A bicycle exercise stress test (stepwise increments of 25 W every 2 min) or dipyridamole (intravenous administration of 0.56 mg/kg over 4 min) was chosen based on the patient’s ability to exercise and to reach at least 85 % of the maximal age-predicted heart rate. Exercise stress testing was performed in 197 patients (64 %, 165 men) and dipyridamole stress testing was performed in 112 patients (38 %, 83 men). Of the 197 patients who underwent exercise stress testing, 148 men (90 %) and 27 women (84 %) reached 85 % of the expected maximal heart rate. Adverse events were defined according to international guidelines [13].

Acquisition protocol

Each patient underwent stress–rest CZT imaging according to a single-day protocol. The protocol for injection of 99mTc-tetrofosmin required administration of 185 – 222 MBq during stress and 370 – 444 MBq at rest. These doses were chosen as 50 % of those used previously [11, 12], as recently reported. All patients with known prior myocardial infarction were injected at rest after sublingual administration of nitrates. In all patients, CZT imaging was performed 15 min after injection of the radiopharmaceutical with an acquisition time of 7 min. According to the single-day protocol, patients were injected at rest 30 min after the end of the first acquisition and then, after an interval of 30 – 45 min, a second acquisition was carried out for 6 min. Image acquisition times were calculated so as to obtain a number of counts generally acquired by the same CZT camera after a standard tracer injection, as previously documented [11, 12].

CZT technology and acquisition protocol

The CZT camera (Discovery NM 530c; GE Healthcare, Haifa, Israel) was equipped with a multiple pinhole collimator and 19 stationary CZT detectors, simultaneously imaging 19 cardiac views. Each detector comprised 32 × 32 pixelated 5 mm thick (2.46 × 2.46 mm) elements. The system design enabled high-quality imaging of the three-dimensional volume, imaged by all detectors (quality field-of-view), where the patient’s heart was to be positioned. Patients were imaged in the supine position with arms placed above the head. Once acquisition was started there was no detector or collimator motion of any kind. List-mode files were acquired and stored. Images were reconstructed on a standard workstation (Xeleris II; GE Healthcare, Haifa, Israel) using a dedicated iterative algorithm; 50 iterations were performed. A Butterworth postprocessing filter (frequency 0.37, order, 7) was applied to the reconstructed slices. The tomographic studies were also reprojected into 60 planar projections to emulate a standard SPECT layout. Images were reconstructed without scatter or attenuation correction.

Image quality

The quality of both stress and rest images was graded visually on a four-point scale as 1 (poor), 2 (fair), 3 (good) or 4 (excellent). The following parameters were considered: myocardial count density and uniformity; endocardial and epicardial edge definition; visualization; and background noise, especially from the subdiaphragmatic area. Two experienced nuclear cardiologists performed the qualitative analysis independently and consensus was reached on all analyses. The readers were blinded to the coronary anatomy but not to the clinical information.

Semiquantitative analysis

Stress and rest perfusion images from the CZT camera were semiquantitatively scored according to the 17-segment model [14] of the left ventricle and a five-point scale (0 normal, 1 equivocal, 2 moderate and 3 severe reduction in radioisotope uptake, and 4 absence of detectable tracer uptake). The summed stress score (SSS) and summed rest score (SRS) were calculated by adding the scores of the 17 segments in the stress and rest images, respectively. To match the results with coronary angiograms, the 17 segments were grouped into territories of the three main coronary arteries, as previously outlined [14]: left anterior descending artery (LAD), circumflex artery (LCx) and right coronary artery (RCA). Two experienced nuclear cardiologists performed the semiquantitative analysis independently and consensus was reached on all analyses.

Analysis of gated images

Left ventricular volumes and ejection fractions were measured after stress and at baseline using previously validated software [15].

Coronary angiography

Selective conventional coronary angiography was performed using standard techniques (Innova 2000 GE; General Electric). Standard multiple projections were recorded for the left and right coronary arteries. Coronary angiograms were quantitatively analysed using an off-line computer-based software program (MEDIS CMS version 6.0; MEDIS Imaging Systems) with an automatic edge-contour detection algorithm using standard and previously validated qualitative and quantitative parameters and definitions. The variables explored were the presence of significant stenosis (≥70 % luminal diameter reduction) in the epicardial coronary arteries or ≥50 % in the left main trunk.

Statistical analysis

Continuous variables are presented as means ± standard deviation. Where indicated, differences were assessed using Student’s t test for paired data. Intraobserver and interobserver variability were measured using percent agreement and kappa values. Accuracy in coronary stenosis detection was assessed by the area under the receiver operating characteristic (ROC) curves. A p value <0.05 was used to define statistical significance. The analyses were carried out with Stata version 11 (Statacorp, College Station, TX).

Results

Patient characteristics

Table 1 summarizes the clinical characteristics of the general population. Prior myocardial infarction was present in 50 of the 309 patients (16 %), while 71 (23 %) and 18 (6 %) had had revascularization by means of coronary angioplasty and coronary artery by-pass grafting, respectively. No adverse events were observed during stress.

Table 1 Clinical characteristics of 248 male and 61 female patients
Full size table

Image quality and evaluation of semiquantitative scores

Stress images were graded “good” or better in 285 of 309 patients (92 %), and rest images in 278 of 309 patients (90 %). No differences were observed in image quality between men and women. In particular, in the evaluation of the stress images, the images in 228 of 248 men and in 57 of 61 women were graded “good” or better, and in the evaluation of the rest images, the images in 223 men and in 55 women were graded “good” or better. The intraobserver and interobserver variability were 93 % (kappa 0.86, 95 % CI 0.72 – 0.94) and 87 % (kappa 0.79, 95 % CI 0.70 – 0.92), respectively.

For the semiquantitative scores, the intraobserver and interobserver variability were 89 % (kappa 0.81, 95 % CI 0.69 – 0.92) and 84 % (kappa 0.75, 95 % CI 0.65 – 0.90), respectively.

Coronary anatomy

Left main trunk stenosis, LAD stenosis, LCx stenosis and RCA stenosis were seen in 3, 155, 142 and 131 patients, respectively. Absence of significant coronary stenosis was detected in 22/309 patients.

Correlation with coronary angiography

Women

The mean SSS and SRS were 8 ± 5 and 3 ± 1, respectively. ROC curves in relation to global and regional SSS values are shown in Fig. 1. The overall ROC area was 0.822 (95 % CI 0.685 – 0.959), and the regional ROC areas for each vascular territory were 0.961 (95 % CI 0.914 – 1.000) for the LAD, 0.952 (95 % CI 0.896 – 1.000) for the LCx, and 0.894 (95 % CI 0.812 – 0.977) for the RCA. Using an SSS cut off of >4, overall sensitivity and specificity were of 88 % and 70 %, respectively, the positive and negative predictive values were 87 % and 75 %, respectively, and overall diagnostic accuracy was 83 %.In the 49 patients (80 %) without a previous history of CAD, the sensitivity and specificity were 85 % and 72 %, respectively.

Fig. 1
figure 1

Overall and per vessel ROC curve analysis in women. The curves represent for each SSS cut-off point pairs of sensitivity–specificity values. The areas under the ROC curves indicate very good accuracy in all vessels in the identification of CAD with the highest accuracy in the left anterior descending coronary artery territory

Full size image

Men

The mean SSS and SRS were 9 ± 5 and 2 ± 2, respectively. ROC curves in relation to global and regional SSS values are shown in Fig. 2. The overall ROC area was 0.884 (95 % CI 0.836 – 0.933), and the regional ROC areas for each vascular territory were 0.929 (95 % CI 0.898 – 0.959) for the LAD, 0.889 (95 % CI 0.848 – 0.931) for the LCx and 0.872 (95 % CI 0.827 – 0.917) for the RCA. Using an SSS cut off of >4, overall sensitivity and specificity were of 89 % and 78 %, respectively, the positive and negative predictive values were 87 % and 70 %, respectively, and overall diagnostic accuracy was 88 %.In the 179 patients (72 %) without a previous history of CAD, the sensitivity and specificity were 86 % and 75 %, respectively.

Fig. 2
figure 2

Overall and per vessel ROC curve analysis in men. The curves represent for each SSS cut-off point pairs of sensitivity–specificity values. Similar to the findings in women, the areas under the ROC curves indicate very good accuracy in all vessels in the identification of CAD with the highest accuracy in the left anterior descending coronary artery territory

Full size image

Gated images

In men, the mean baseline ejection fraction was 52 ± 18 %, the end-diastolic volume was 134 ± 31 ml, and the end-systolic volume was 64 ± 25 ml. Corresponding values after stress were 49 ± 21 %, 138 ± 35 ml and 68 ± 27 ml (p = not significant for all values). In women, the mean ejection fraction at rest was 52 ± 18 %, the end-diastolic volume was 122 ± 24 ml, and the end-systolic volume was 59 ± 18 ml. Corresponding values after stress were 49 ± 19 %, 125 ± 29 ml and 61 ± 23 ml, respectively.

Discussion

In this study in a cohort of patients with known or suspected CAD no differences in sensitivity and specificity using a CZT camera were detectable between men and women. These results are comparable with those previously reported [16], and most importantly, they were obtained with a low-dose radiation protocol, an extremely important feature for the safety of patients.

The purpose of the diagnostic work-up in women with symptoms suspicious for CAD is to identify significant lesions with optimal accuracy and to lay the basis for implementing preventive and therapeutic interventions [9]. More effective diagnostic strategies are critical in women at risk of CAD, because up to 40 % of initial cardiac events are fatal [17]. As suggested in the literature, compared to men, women are initially diagnosed with more advanced CAD because of a lack of early recognition and management [18]. Therefore, a better understanding of how the accuracy variation of different noninvasive tests for CAD varies according to gender could dramatically improve the outcome in many women.

Although most early studies of the utility of myocardial perfusion imaging in establishing the diagnosis and prognosis of CAD included predominantly men, studies have established that SPECT myocardial perfusion imaging using either 201Tl or 99mTc sestamibi/tetrofosmin provides excellent diagnostic and prognostic information in women [5, 19–20].

The data obtained in the present study, indicate that the use of a CZT camera can reproduce and improve the accuracy in detecting CAD in terms of sensitivity and specificity with a lower radiation exposure that is one of the main shortcomings of myocardial perfusion imaging by SPECT.

Previous limits overcome by CZT

Previous studies with standard myocardial SPECT in women have shown some limitations such as left ventricular size and soft-tissue attenuation that can potentially lead to reduced accuracy [21–24]. The smaller hearts in some women may potentially limit spatial resolution causing small areas of reduced perfusion to go undetected. Moreover, diagnostic specificity can also be influenced by breast tissue, which can cause false-positive results, especially in the anterior myocardial segments. These limitations have mostly been described with the use of 201Tl and are theoretically decreased with higher energy 99mTc radioisotopes, whose use is preferable in women [19, 21]. However, the use of gating improves the specificity of SPECT in distinguishing breast attenuation from scar based on regional wall motion [25, 26]. Supplementing gated SPECT with attenuation correction may also help to further reduce artefacts related to attenuation from breast tissue.

The use of a CZT camera could overcome these limitations due to the higher spatial resolution and consequently to the better identification of myocardial walls. Increased resolution contributes to a better identification of stress defects that may affect the myocardium in a nontransmural way. Beyond this, the real power of the CZT camera is in the design of the detector that through the multipinhole technique focuses the heart leading to a reduction in attenuation artefacts, thus increasing specificity. Finally, the increased resolution allows better evaluation of end-diastolic and end-systolic volumes with consequently better quantification of ejection fraction, without the underestimation of volumes obtained in women by standard SPECT.

Study limitations

The results of this study were based on a relatively small group of highly selected patients and this could have led to a selection bias. However, all the patients underwent coronary angiography after nuclear testing, and thus the evaluation of sensitivity and specificity was possible. Moreover, the patients showed a low occurrence of ventricular dysfunction, and thus volumes and ejection fractions were not examined in dilated hearts.

Conclusion

A low-dose protocol with a CZT camera can be routinely used in women with known or suspected CAD without loss of accuracy and with lower radiation exposure of the patients, and therefore with better safety.