A high agreement of results was demonstrated between perfusion images obtained with the CZT-SPECT and A-SPECT cameras. In a multicenter study involving 168 patients, Esteves et al.
[1] (
Table 1) compared images obtained with the Discovery NM 530c and standard gamma cameras. The images made it possible to detect or exclude ischemia with high, 91.9% agreement (compared to 92.5% agreement, obtained when the same studies acquired with a standard camera were assessed twice,
p = 0.99). Sharir et al.
[2] demonstrated, by means of a multicenter study involving 238 patients, high concordance of quantitative assessments of total perfusion deficits (for D-SPECT and A-SPECT in the stress and rest studies linear correlation coefficients equaled 0.95 and 0.97, respectively) with good agreement in three vascular territories (kappa coefficients for LAD, LCX and RCA were 0.73, 0.73 and 0.70, respectively, the concordance in detecting defects in each of the three vascular areas exceeded 90%).
Table 1. The most important information along with a list of literature on myocardial perfusion studies performed with CZT cameras.
Esteves et al. [1] |
|
A-SPECT 1 vs. CZT-SPECT 2 |
92% Agreement in Detection of Ischemia |
Sharir et al. [2] |
High concordance of TPD 3, concordance in vascular territories >90% |
Verger et al. [3] |
High >85% concordance for MIBI 4 and Thallium |
Mannarino et al. [4] |
Women with low to intermediate prob. of CAD 5—no correlation between SSS 6 values |
Cantoni et al. [5] |
CZT-SPECT vs. coronary angiography |
Diagnostic efficacy of CZT-SPECT slightly higher than of A-SPECT (AUROC 7 89% vs. 83%) |
Nudi et al. [6] |
Detection of CAD: sensitivity. 0.84, specificity 0.69 |
Zhang et al. [7] |
Detection of CAD: sensitivity 0.84, specificity 0.72 |
Gimelli et al. [8] |
Effective detection of patients with mutivessel CAD |
Van Dijk et al. [9] |
AC 8 |
Increase in specificity (0.45 to 0.67) |
Caobelli et al. [10] |
increase in specificity (0.40 to 0.100) |
Esteves et al. [11] |
AC does not affect visual interpretation |
Mirshavalad et al. [12] |
Two patient positions |
Higher specificity in prone than in supine position (0.86 vs. 0.67) |
Nishiyama et al. [13] |
Supine + prone—increase in specificity vs. supine only (0.85 vs. 0.50) |
Nakazato et al. [14] |
Upright + supine vs. upright (sensitivity 0.94 vs. 0.86, specificity 0.91 vs. 0.59) |
In addition, Verger et al.
[3] verified, on a group of 276 patients, the agreement of study results obtained with a semiconductor and standard gamma cameras in studies applying three different protocols: using a standard and low activity of [99mTc] Tc-MIBI, as well as [201Tl] thallium. In all cases, a high agreement was obtained between results on both types of cameras; with standard activity [99mTc] Tc-MIBI—98%; low—86% and using [201Tl] thallium—92%. There was also a high agreement (over 85%) between the A-SPECT and CZT-SPECT cameras in terms of the percentages of abnormal study results, ischemia and permanent perfusion defects. Other authors of studies on smaller groups of patients also demonstrated high agreement between results of studies on both discussed types of cameras
[3][4][9][15][16].
However, the consistency of the results was not so high in all patient groups. Mannarino et al.
[4] examined a group of women with low and intermediate probability of coronary artery disease using both types of cameras. There was no significant correlation between summed stress scores (SSS)—r = 0.18,
p = 0.06, and a significant but weak (48%) correlation between total perfusion deficit (TPD) obtained as a result of semi-automatic comparison of individual images with normal databases. The discrepancy concerned the perfusion of anterior wall. In A-SPECT, weaker uptake of the radiotracer was detected more often in this area, which was not observed in CZT-SPECT. Based on the normal wall thickening in the gated study, all results suspected of ischemia were classified as normal. CZT-SPECT was considered a more useful technique in this group of patients due to its greater effectiveness in excluding perfusion abnormalities.
2. Diagnostic Efficacy of CZT Cameras
In a meta-analysis of 40 studies including a total of 7734 patient studies on the diagnostic efficacy of CZT-SPECT and A-SPECT, Cantoni et al.
[5] showed, in relation to coronary angiography, that CZT-SPECT fulfills its role properly in the diagnosis of coronary artery disease. Its diagnostic efficacy was slightly higher than that of A-SPECT (sensitivity and specificity compared to coronary angiography were 89% and 69% versus 85% and 66%, respectively, while the areas under the ROC curve amounted to 0.89 and 0.83). The accuracy of studies using CZT-SPECT was slightly, although statistically significantly, higher than that of studies with A-SPECT (
p < 0.05). This is explained by better spatial resolution and the resulting higher contrast of images obtained with semiconductor gamma cameras. Authors of previous meta-analyzes involving studies on semiconductor cameras combined with coronary angiography (Nudi et al.
[6]—2092 patients and Zhang et al.
[7]—2350 patients) obtained similar results regarding the sensitivity and specificity of studies in the detection of coronary artery disease, without showing significant differences between the two CZT cameras used in the research. Moreover, Gimelli et al.
[8] showed, on the material of 695 patients, that a CZT camera allows for the effective detection of CAD in patients with a multivessel disease. Areas under the ROC curves obtained on the basis of semi-quantitative measures—summed stress scores (SSS) in patients with two- and three-vessel disease, 0.83 and 0.79, respectively, did not differ significantly from the area under the ROC curve for single-vessel patients—0.80. This method allowed for the proper detection of the extent of coronary artery disease in 64% of patients. In addition, regional perfusion defects corresponding to the location of the three major coronary arteries were detected highly efficiently (areas under the ROC curves for LAD, LCX, and RCA amounted to 0.90, 0.88, and 0.87, respectively). The article of Gimelli et al. indicates an advantage of semiconductor over Anger cameras in terms of the effective detection of patients with a multivessel disease. There are many previous reports informing about difficulties in diagnosing such patients due to the relative nature of a radiopharmaceutical distribution in images, and also in the proper determination of the extent of the coronary artery disease using conventional cameras
[17][18].
As shown by the two previously mentioned meta-analyzes of the results of CZT-SPECT perfusion studies
[6][7], the specificity of those studies in detecting coronary artery disease is lower than their sensitivity (69% and 72% vs. 84% and 84%, respectively). The relatively high rate of false-positive results is attributed, as is the case with conventional cameras, to the attenuation of radiation in the patient body, which is largely related to obesity
[9]. In order to improve the quality of studies, attempts are made to correct the attenuation. Another approach to solve this problem is placing a patient in different positions during study acquisition, assuming that the filling of perfusion defects due to a change in the patient position indicates their false nature.
Attenuation correction (AC) aims to eliminate artifacts resulting from radiation absorption in a patient body. On the basis of studies using a hybrid semiconductor camera integrated with a CT (Discovery NM/CT 570c), van Dijk et al.
[16] showed that AC increases diagnostic confidence (the percentage of correct stress images increased from 45% of uncorrected—NC, up to 72% of attenuation corrected—AC, and up to 67% while considering both sets—corrected and uncorrected). A subsequent follow-up did not reveal a higher incidence of major cardiac events in patients whose NC studies were considered more normal than those with AC. The authors conclude that the use of AC allows for more frequent giving up of the rest study, and thus limiting the exposure to ionizing radiation to the dose of about 1.2mSv.
However, only a few sites in the world have a hybrid cardiac CZT camera. Attempts are being made to obtain the same information using a CT not integrated with a gamma camera. The question of whether AC is necessary for a reliable assessment of myocardial perfusion imaged by the CZT camera has not yet been resolved.
Caobelli et al.
[10] report that, in comparison with results of coronary angiography, external CT attenuation correction increases the specificity of the CZT-SPECT study (specificity for the AC and NC studies were, respectively: in the assessment of a whole myocardium perfusion—100% vs. 40%; in the assessment of vascular perfusion—LAD: 63% vs. 36%, LCX: 70% vs. 33%; RCA: 81% vs. 19%). AC significantly increased the diagnostic accuracy of the global myocardial score (
p = 0.01) as well as in the RCA vascular territory (
p = 0.02), but not in the LAD (
p = 0.35) and LCX areas (
p = 0.08).
On the other hand, Esteves et al.
[11] claim that the attenuation correction of CZT-SPECT images does not significantly affect their visual interpretation. Moreover, CT exposes patients to an additional radiation dose and, due to a possible patient motion resulting in a SPECT and CT misalignment, may affect the credibility of the examination.
As already mentioned, another method of dealing with artifacts resulting from the absorption of radiation in the patient body is to perform the examination in two different positions. Mirshavalad et al.
[12], in a meta-analysis of studies using standard and semiconductor cameras, showed on two separate groups of patients that the prone position study—1490 patients—allows for similar sensitivity (83% vs. 86%) and higher specificity (79% vs. 67%) than the study in the supine position—1138 patients. The pooled sensitivity and specificity of the prone position in detecting the right coronary artery territory defects were 70% and 84%. Nishiyama et al.
[13] showed, on a group of 276 patients, that taking into account images from both positions (supine + prone) significantly improves the perfusion interpretation in terms of specificity (
p = 0.02) and accuracy (
p = 0.04) in relation to the supine position: sensitivity, specificity and accuracy: 85% vs. 78%, 85% vs. 50%, 85% vs. 76%, respectively.
Study acquisition in prone and supine positions is possible only with the Discovery NM 530c camera. In the case of D-SPECT, it is possible to acquire studies in the upright, supine and bikerlike positions. Combined assessment of images acquired in the sitting and supine position (upright + supine), in comparison with the results of coronary angiography, provided a sensitivity and specificity of 94% and 86%, respectively, compared to studies acquired only in the upright (91% and 59%) or only supine (88% and 73%) positions; areas under the ROC curves differed statistically significantly (0.94 vs. 0.88 and 0.89,
p < 0.05)
[14]. The forward leaning bikerlike position decreased a distance between the heart and the detector and eliminated attenuation artifacts in relation to the sitting and supine position
[19].
The Discovery NM 530c camera is prone to artifacts in very obese patients resulting from a small field of view and a pinhole collimation geometry
[20][21].
In D-SPECT cameras this problem does not occur and, in doubtful situations, caused by diaphragmatic artifacts, it is possible to perform a study in sitting and supine positions
[4][14].
It is worth emphasizing that the CZT camera, due to its higher sensitivity, provides greater flexibility in the selection of a study protocol. Compared to classic cameras, one can shorten the acquisition time of a study with preserved radiopharmaceutical activity, reduce activity or apply both, to a lesser extent. By administering a lower activity patient radiation dose can be reduced
[1]. As reported by Einstein et al.
[22], it is even possible to reduce the effective dose to 1mSv in the case of normal stress studies when the other parts are not necessary.
Among the limitations of CZT cardiac cameras, the difficulties in observing patient movement during study acquisition are stressed
[23]. This movement can have a significantly negative impact on reconstructed images. Software for Anger cameras makes it possible to observe patient movement, for example by displaying raw images in a cinematic mode or by observing the discontinuities of heart coordinates. Patient movement can be easily corrected before a reconstruction procedure.
However, in the case of CZT cardiac cameras, due to their different geometry, correction of the patient movement, although possible, is much more difficult and, above all, time-consuming. This limits its application to cases significantly suspected of patient movement during study acquisition.
A high cost of those devices is also stressed, especially because they are dedicated to a strictly defined type of studies
[23].