Evaluation of adenosine-induced splenic switch-off in Tc-99m tetrofosmin (Myoview) myocardial perfusion studies as marker of stress adequacy by semiquantitative analysis of acquired single-photon emission computed tomography/computed tomography images

Abstract

Aim/Introduction

Adenosine pharmacological stress is commonly used for the detection of ischemia by myocardial perfusion imaging (MPI) using radiotracers and gated single-photon emission computed tomography/computed tomography (SPECT/CT) imaging. Recently, a phenomenon, splenic switch-off (SSO), has been observed where reduced perfusion to the spleen occurs after adenosine, suggesting an adequate effect of adenosine on the heart. The aim of the study is to evaluate the SSO effect in adenosine stress ^99mTc tetrofosmin MPI studies using semiquantitative analysis of SPECT/CT data.

Materials and methods

Seventeen adenosine MPI studies were selected. Each study included stress and rest, which were reoriented in short, horizontal, and vertical long axes. Also, data were reconstructed in the body transaxial, coronal, and sagittal planes. In the resulting images, the spleen was identified on the CT component, and regions of interest were drawn on the spleen, thoracic vertebra, and myocardium. The ratios spleen/vertebra (S/V_Ad and S/V_Rst) and myocardium/vertebra (M and M/_t) were calculated. Statistical analysis was done using a paired t test.

Results

Mean age was 60.35 years, with 9 males (53%). Mean S/V_Ad was 0.98 ± 0.38 and S/V_Rst 1.87 ± 0.87 (t = −3.707, P < 0.01). The mean M/V_Ad was 2.86 ± 1.65 and M/V_Rst 3.75 ± 2.41 (t = −1.323, P = 0.204).

Conclusion

Using the method outlined, especially the S/V ratio, evaluation of the SSO can be achieved from the acquired data, which shows promise as a marker for adequate effectiveness of adenosine stress in doubtful cases.

Introduction

Adenosine pharmacological stress myocardial perfusion imaging (MPI) is a validated method for the noninvasive evaluation of coronary artery disease in individuals who are unable to exercise adequately. Adenosine acts as a vasodilator, resulting in a hyperemic response, eliciting changes of regional myocardial perfusion using Tc-99 m tetrofosmin (Myoview) single-photon emission computed tomography/computed tomography (SPECT/CT), which are compared to the baseline perfusion at rest for detection of ischemia [1]. However, the diagnostic performance of such use is contingent on obtaining adequate pharmacological stress. This can be done by gauging hemodynamic responses such as heart rate and blood pressure and the clinical response during adenosine administration, including chest pain and other symptoms (flushing and headache), which may not accurately reflect the actual physiologic vasodilation, leading to a risk of underdiagnosis of ischemic heart disease. This matter is particularly important in patients who do not fully adhere to the preparation instructions, especially abstention from caffeine ingestion before the test.

Recently, adenosine-induced splenic switch-off (SSO), which is a brief reduction in splenic perfusion detectable by imaging, is becoming more widely acknowledged as an objective and reliable sign of appropriate adenosine responsiveness during the stress test. Although its exact mechanism is still not well defined, studies using both PET and cardiac MRI [2] have shown that SSO can be useful as a marker for adequate vasodilation, implying that it could play a role in quality assurance for MPI stress regimens. Recent studies indicate that semiquantitative and quantitative assessment of SSO from the acquired images can provide reliable correlation with the adequacy of pharmacologic stress [3,4]. So far, however, there is little evidence on the diagnostic utility of adenosine-induced SSO on Tc-99 m tetrofosmin (Myoview) MPI gated SPECT/CT using semiquantitative image analysis. Standardizing such an approach could help refine stress adequacy assessment, providing a potential quality control application in this procedure.

Accordingly, the aim of this study is to test a semiquantitative method for assessment of adenosine-induced SSO in Tc-99 m Myoview gated SPECT/CT MPI as a simple and reliable marker of stress adequacy, thus allowing for more accurate MPI interpretation and better clinical management of patients.

Methods

A cross-sectional study was performed on a total of seventeen adult patients who underwent adenosine stress Tc-99 m Myoview MPI gated SPECT/CT studies. The cases were selected consecutively from the departmental archives. All chosen studies included both stress and rest parts. After checking for adequate patient preparation for the test (fasting for 4–6 h, stopping caffeine for 24 h), each patient received an intravenous infusion of adenosine at a rate of 140 µg/kg/min over 6 min while blood pressure and ECG were monitored. At the third minute of adenosine infusion, an intravenous bolus of Tc-99 m Myoview was administered (740 MBq, 20 mCi). Afterwards, the patient was given a standardized meal and imaging was done about 30–45 min post-injection of the radiotracer using a dual-head camera equipped with low-energy high-resolution collimators according to the gated SPECT/CT institutional imaging protocol. Rest imaging followed a similar process on a different day.

The stress and rest images were processed for interpretation and reporting using the Myovation software (Xeleris workstation, GE Healthcare, USA) that generated aligned stress–rest short axis, vertical long axis, and horizontal long axis perfusion images of the left ventricle (Fig. 1). Functional images for wall motion and ejection fraction display were also produced. In addition to the standard gated SPECT/CT processing of the heart, the same data sets were reconstructed in the conventional anatomic planes of the body (i.e., transaxial, coronal, and sagittal) using the Volumetrix software, which is also available on the Xeleris platform. This allowed SPECT/CT dual display of cardiac and extracardiac structures in the acquired field of view used for cardiac imaging, including the heart, spleen, and vertebrae. Semiquantitative analysis of the resulting SPECT/CT images was performed by drawing circular regions of interest (ROIs) over three structures for both stress and rest images. The first structure was the spleen, identified using the CT slices as the soft-tissue structure in the left upper quadrant; the second structure was the myocardium, using visual correlation with cardiac CT and perfusion slices; and the third structure was an adjacent thoracic vertebral body as seen on the CT images, acting as a reference structure (Fig. 2). The average count per pixel within each ROI was recorded as follows: spleen adenosine (S_Ad), myocardium adenosine (M_Ad), vertebra adenosine (V_Ad), spleen rest (S_Rst), myocardium rest (M_Rst), and vertebra rest (V_Rst). Also, ratios of the recorded ROI uptake were calculated for the stress and rest study in each patient as follows: S_Ad/V_Ad, M_Ad/V_Ad, and S_Rst/V_Rst (see Table 1).

Fig. 1.

Display of MPI stress and rest images in the heart short, horizontal long, and vertical long axis that are used for diagnosis of ischemia. MPI, myocardial perfusion imaging.

Fig. 2.

Display of the imaging data for stress or rest images in the body axes with the ROls on spleen, myocardium, and vertebra. ROIs, regions of interest.

Table 1.

Results of the semiquantitative analysis for mean uptake in spleen, myocardium, and vertebra and the calculated ratios of spleen/vertebra and myocardium/vertebra for adenosine and rest studies

Region	Count/pixel	t-value P value	Ratio to vertebra	t-value P value
S Ad	308 ± 136	4.64 <0.001	0.98 ± 0.38	−3.707 <0.01
S Rst	144 ± 77		1.87 ± 0.87
M Ad	793 ± 282	7.612 <0.001	2.86 ± 1.65	1.323 0.204
M Rst	263 ± 149		3.75 ± 2.41
V Ad	349 ± 170	6.888 <0.001	1
V Rst	80 ± 26		1

M_ad, myocardium adenosine; M_Rst, myocardium rest; S_Ad, spleen adenosine; S_Rst, spleen rest; V_Ad, vertebra adenosine; V_Rst, vertebra rest.

The data were statistically analyzed using IBM SPSS Statistics software with descriptive statistics (mean and SD) for patients’ demographics. The paired t-test was applied for the semiquantitative data analysis to the 95% confidence interval.

In addition to its usefulness for applying the semiquantitative method, reconstruction of the SPECT/CT imaging data in the orthogonal planes allowed visual inspection of coronary calcifications and any incidental findings in the part of the lungs in the field of view (data not included).

Results

A total of 17 patients [mean age: 60.35 ± 12.97 years; nine males (53%) and eight females (47%)] were analyzed. The main indication for the performance of the study was to evaluate for ischemic heart disease. Clinically, 11 (65%) had ischemia, one (6%) had a scar, and five (29%) were normal based on the findings of the MPI studies. The clinical monitoring of the patients during adenosine infusion showed a mild increase in heart rate of about 10–15 beats/min in all of them and a slight decrease in systolic blood pressure of 10–20 mmHg in 14 patients (82%). Some patients (seven patients; 41%) exhibited changes in the ECG involving ST-segment depression with or without T-wave inversion that reverted to baseline at the end of the stress part. All cases complained of one or more symptoms during adenosine infusion, including chest tightness or chest pain, headache, and fatigue, all of which resolved shortly after the end of the infusion. The semiquantitative method was relatively straightforward to perform, with all the structures required for analysis (spleen, myocardium, and thoracic vertebra) readily in the field of view of SPECT/CT acquisition.

The results of the average count rate per pixel for the myocardial ROIs (see Table 1) were highest for the stress (793 ± 282) and decreased significantly to 263 ± 149 at rest (t = 7.612, P < 0.001) with a 300% decrease. A similar pattern, although to a lesser magnitude, showed the splenic count rate average after adenosine was 308 ± 136 and fell significantly to 144 ± 77 at rest (t = 4.64, P < 0.001), representing a mean reduction of about 50% in splenic radioactivity. The vertebral count rates per pixel declined from 349 ± 170 during stress to 80 ± 26 at rest (t = 6.888, P < 0.001), a reduction of 440%.

When normalized to vertebral counts, the spleen-to-vertebra ratio was substantially lower during stress (S_Ad/V_Ad: 0.98 ± 0.38) compared with rest (S_Rst/V_Rst: 1.87 ± 0.87), and this reduction was statistically significant (t = –3.707, P < 0.01). On the other hand, the myocardium-to-vertebra ratio (M_Ad/V_Ad) was 2.86 ± 1.65 during stress and M_Rst/V_Rst was 3.75 ± 2.41 at rest, a difference that did not reach statistical significance (t = 1.323, P = 0.204). This would mean that the observed deviation of the pattern of splenic uptake from the expected pattern at rest in the vertebra, which is related to background bone marrow uptake, might reflect the phenomenon of SSO.

Discussion

Using imaging data acquired during adenosine stress and rest Tc-99 m tetrofosmin (Myoview) gated SPECT/CT MPI, this study demonstrates that semiquantitative analysis of S/V and M/V uptake ratios can provide a simple and objective method for evaluating adenosine-induced SSO. This phenomenon, characterized by a clear and measurable reduction in splenic uptake following adenosine infusion, has been increasingly recognized as a potential indicator of adequate pharmacological stress [2]. The underlying physiological mechanism is thought to involve adenosine-mediated splenic vasoconstriction by sympathetic activation, thus leading to transient hypoperfusion of the spleen, possibly as a compensating mechanism aiming at maintaining circulating blood volume, as occurs in hemorrhage. This is in contrast to vasodilation of the coronary and systemic vasculature. This splenic reaction to adenosine infusion, although not clinically relevant per se, could act as a readily available tool to ascertain whether adenosine caused the pharmacological effects expected from its use. Technically, it is amenable to study using available SPECT/CT technology that can directly assess splenic perfusion thanks to coverage of the area during acquisition and availability of simple tools for quantitative analysis. Also, in addition to evaluation for SSO, the semiquantitative method provides information on the perfusion to myocardium after adenosine stress and rest, thus giving a more comprehensive cover of the organs of interest in this study (spleen and heart).

From a technical point of view, the semiquantitative method used data already available from the acquired Tc-99 m tetrofosmin gated SPECT/CT MPI stress and rest studies for each patient case. The selection of the regions of interest for the semiquantitative analysis was based on known kinetics of the Tc-99 m tetrofosmin radiotracer, which distributes in the body according to regional perfusion except for the brain, liver, and kidneys. In the vertebral column, Tc-99 m tetrofosmin would be expected to be taken up by the bone marrow, thus providing a suitable background region that is only affected by the delivery of radiotracer to it, that is, its perfusion without spurious uptake as seen in the liver or bowel. More importantly, the vertebrae are clearly seen on CT with the SPECT/CT cardiac acquisition, which enables easy placement of ROIs on them.

In the cases studied, S/V ratios were consistently reduced in the adenosine stress phase compared to rest (data not shown). In addition, the M/V ratios provided uptake patterns that could be compared to those in the spleen, which exhibited the most marked relative change and suggests that while both myocardial and vertebral counts decreased between adenosine stress and rest, their proportional relationship remained relatively unchanged, contrary to the findings in the spleen. This observation agrees with prior reports that spleen uptake reduction is more pronounced and consistent than myocardial uptake variation when assessing stress adequacy [2].

A review of reported identification of SSO showed that some previous studies relied on qualitative visual identification of SSO, which may lack sensitivity, may be challenging in borderline cases, and is subject to inter-observer variability. The results presented in this study are in agreement with the work of Manisty et al. [2], who noted that SSO correlates with hemodynamic and symptomatic markers of adequate adenosine effect and suggested that a quantitative approach could improve diagnostic confidence, particularly in patients with absent or equivocal clinical responses. On the other hand, Kuijpers et al. [5] concluded that the SSO effect was not of great value in the prediction of adenosine nonresponders after certain circumstances, such as caffeine intake. Moreover, a study by Hosking [6], aiming at evaluating the importance of identifying the absence of SSO for detecting false-negative adenosine stress perfusion cardiac magnetic resonance (CMR) scans, found that the rate of lack of SSO was 20.7% in the false-negative group versus 13.1% in true positives (P = 0.37). Moreover, the lack of SSO is prevalent, easily measurable, and has the potential to improve on hemodynamic criteria as a marker of adenosine under stress in CMR perfusion scans. As such, the SSO effect can be utilized as an indicator of the effectiveness of adenosine stress MPI when it is used alongside certain additional stressors, such as exercise, or simply eliminating certain influences, as prior ingestion of caffeine before imaging. The notion that the absence of SSO was more likely to occur in patients with a false-negative CMR than in the group of patients with true-negative results was highlighted by Manisty et al. [2]. Their conclusion was that failed SSO with adenosine is an observation seen in under-stressed patients who are most likely to give a false-negative result on perfusion MR images. Thus, they suggested repeat examination of the under-stressed patients with negative SSO and negative results on perfusion MR images in order to increase the sensitivity of the study. Of note, Karamitsos et al. [7] increased the dose of adenosine injection in patients with negative results but failed to show changes during stress. Other reports stated that increasing the dose of adenosine injection more than 140 µg/kg/min, which is the rate used in this study, did not result in further coronary vasodilation on MPI but would only induce a more exaggerated peripheral hemodynamic response [8], possibly resulting in a higher rate of SSO.

The overall understanding currently is that SSO could have a role in indicating the effectiveness of adenosine stress MPI; however, its utilization in cases of negative results should prompt a re-evaluation of the patient with alternative stressors and proper, adequate preparation for MPI, for example, limiting caffeine intake. Thus, SSO’s role would be to support the finding of the observed true-negative MPI results. In summary, the potential clinical benefit of this approach lies in its ability to act as an additional marker of stress adequacy. This is particularly true in cases where patient cooperation is suboptimal, infusion is interrupted, or the anticipated side effects (e.g. flushing and chest discomfort) are absent.

Finally, this study focused on the technical aspects for evaluation of SSO as described above, using a relatively small sample size of cases in the analysis, which rather limited its statistical power to be recommended for routine clinical use for the purposes mentioned earlier as a test for adenosine stress adequacy. Hence, expansion of the presented analysis to include a higher number of patients would be warranted to establish its potential role in the confirmation of the effectiveness of the adenosine pharmacological intervention, leading to the elimination of false-negative cases and thus higher accuracy of the whole method. A formal ROC analysis would then enable the definition of cutoff values for the ratios generated in the semiquantitative analysis, especially the S/V and M/V ratios. This expansion would also include systematic correlation of SSO findings with observed hemodynamic parameters such as heart rate and blood pressure changes and ECG data, in addition to coronary angiographic findings. Other avenues for exploration would be comparison with exercise stress or other pharmacological agents to provide a broader determination of the S/V ratio changes in cardiac stress for MPI.

Conclusion

This study demonstrates that semiquantitative spleen-to-vertebra ratio analysis is a simple and objective method for detecting adenosine-induced splenic switch-off in Tc-99 m Myoview MPI. The consistent reduction in splenic uptake in terms of S/V ratio during stress supports its use as an objective physiological marker of adequate pharmacologic vasodilation, particularly when traditional clinical or hemodynamic indicators are inconclusive. Because this method can be derived from standard SPECT/CT acquisitions without additional imaging or radiation exposure, it is readily adaptable to routine clinical workflows. Larger studies and cross-stress modality comparisons are warranted to validate its diagnostic utility, with the ultimate goal of integrating quantitative SSO assessment into MPI interpretation guidelines.

Acknowledgements

Conflicts of interest

There are no conflicts of interest.