Effect of the extent of short-duration exercise stress combined with adenosine triphosphate on the artifacts of two different regions on myocardial perfusion single-photon emission computed tomography

Background: We aimed to evaluate the differences in the reduction of undesired artifacts among different low-grade exercise stress levels combined with drug stress and their related factors by using myocardial perfusion single-photon emission computed tomography (SPECT) in patients with suspicious coronary artery disease. Methods: We divided patients into 4 groups as follows: group A, adenosine triphosphate (ATP) for 6 min; group A+25 W, ATP+25 W exercise for 6 min; group A+35 W, ATP+35 W exercise for 6 min; group A+45 W, ATP+45 W exercise for 6 min) and enrolled only those whose summed stress scores were <3. Undesired artifacts were evaluated on the basis of heart-to-liver activity (H/L) ratio and heart-to-10 pixels below the heart (H/below the H) ratio. Results: The logarithmic values of H/L and H/below the H ratios were signicantly higher in groups A+35 W and A+45 W than in group A. In all the patients, the logarithmic values of H/L and H/below the H ratios positively correlated with heart rate (p=0.004 and p=0.008, respectively), rate pressure product (RPP; p=0.005 and p=0.008, respectively), and the increment of RPP (p=0.002 and p=0.005, respectively) after stress in the univariate analysis. The left ventricular end-diastolic volume (LVEDV) after stress (p=0.002) negatively correlated with the logarithmic value of H/below the H ratio, but not H/L ratio. Although male sex was independently associated with the logarithmic values of both H/L (p=0.014) and H/below the H ratios (p=0.001) in multivariate regression analysis, LVEDV was also independently associated with the logarithmic value of H/below the H ratio (p<0.001). Conclusion: ATP plus ≥ 35 W exercise stress for 6 min was useful for reducing undesired artifacts after stress in myocardial perfusion SPECT.


Background
Streak artifacts are the most common source of error in myocardial single-photon emission computed tomography (SPECT) imaging. Stress with vasodilators has become increasingly used in myocardial perfusion SPECT, but the reduced activities in the inferior regions on SPECT images are explained by the liver-to heart artifact due to the intense tracer uptake in the liver [1,2]. High liver activity adjacent to the inferior wall results from oversubtraction of the activity from the inferior wall. Myocardial uptake was greatest with administration of vasodilators than with exercise, and this effect was not attenuated by combining vasodilator administration with exercise [3]. Exercise stress is well known to increase heart-tobackground count ratios as compared with drug stress alone and to reduce hepatic tracer uptake relative to the heart. Although adding low-level exercise stress to drug stress may help overcome some of the limitations of drug infusion alone and is useful for reducing the streak artifact [4][5][6], the appropriate extent of short-duration exercise stress remains controversial. However, no direct comparison has been made between the standard adenosine stress and the pharmacological stress combined with bicycle exercise at various low grades. In this study, we hypothesized that a protocol combining a 6-min adenosine triphosphate (ATP) infusion with several different low-grade bicycle exercise protocols would improve streak artifacts among patients unable to generate an adequate stress during exercise, as compared with the standard ATP-only protocol. We evaluated the differences in the reduction of undesired artifacts between two different background regions and the related factors on myocardial perfusion SPECT in patients with suspicious coronary artery disease. We only included patients whose summed stress scores (SSSs) were < 3 for a robust and accurate evaluation of artifacts.

Methods
Subjects with suspected angina pectoris who were unable to generate adequate stress during exercise because of antianginal medications or confounding factors that might have hindered them from achieving the target heart rate underwent elective 99mTc-tetrofosmin stress myocardial perfusion imaging (MPI). The patients (n = 280, from December 2015 to September 2017) were randomly allocated to one of the four stress protocols (n = 70 each) as follows: ATP [7,8] at 120 µg/kg for 6 min (group A) and ATP at 120 µg/kg for 6 min concomitantly combined with various exercise intensities for 6 min at 25, 35, and 45 W (groups A + 25 W, A + 35 W, and A + 45 W, respectively). The subjects exercised on a bicycle ergometer and received 110 MBq of 99m Tc tetrofosmin injection at 3 min after initiation of ATP infusion.
Image acquisition was initiated 45 min after the 99m Tc injection. A 2-min planar image was obtained, followed immediately by the acquisition of a standard gated SPECT image. Perfusion imaging was performed using a 2-detector gamma camera (ADAC Forte), with a circular 180° orbit. The perfusion data were displayed in a 17-segment model in accordance with the American Society of Nuclear Cardiology guidelines [9]. The 17 segments for each image were scored from 0 to 4, with 0 indicating no perfusion defect and 4 indicating no tracer. In this study, we enrolled only patients whose SSSs were < 3 because of robust clari cation of artifacts. Therefore, we analyzed only the data of 109 patients (group A, n = 23; group A + 25 W, n = 34; group A + 35 W, n = 28; and group A + 45 W, n = 24). Whereas the stress and rest studies were interpreted together in the usual manner, the four-group images were analyzed in random order to avoid bias by the expert reader. The exclusion criteria were as follows: a contraindication to adenosine (moderate to severe chronic obstructive pulmonary disease or asthma, second-or third-degree atrioventricular block or sinus node disease, known as hypersensitivity to adenosine); hemodynamic instability; decompensated congestive heart failure; or the use of theophylline or dipyridamole within the preceding 48 h. This study complied with the tenets of the Declaration of Helsinki, and all patients provided written informed consent to participate.
Undesired background artifacts were evaluated using two different modes. To obtain the heart-to-liver activity (H/L) ratio, the regions of interest were drawn around the heart and liver, excluding the gall bladder. "Heart counts" and "liver counts" were obtained from the point of maximum counts in each region of interest (Fig. 1). Undesired artifacts were also evaluated on the basis of heart-to-10 pixels below the heart ratio (H/below the H ratio). The pro le lines were placed on the heart and below the heart from the bottom of the heart image to 10 pixels. "Heart counts" were obtained from the maximum count in the pro le line, and "below heart counts" were obtained from the minimum count in the pro le line. These ratios were determined from the anterior planar images.
Continuous variables are expressed as means ± standard deviations, whereas categorical variables are presented as percentages. Differences in categorical variables among the groups were assessed using chi-square tests (4 × 2), while those in continuous variables were assessed using one-way analysis of variance, and those between two groups were made using a post hoc Bonferroni test. The signi cance of the correlation between two variables was assessed with regression analysis. A p value of < 0.05 was considered statistically signi cant.

Results
The clinical characteristics of the patients are shown in Table 1. No signi cant differences in age, sex, incidence of comorbidities, medications, and laboratory data except for low-density lipoprotein cholesterol levels were found among the four groups. The values of the hemodynamic parameters were signi cantly higher in the groups with concomitant exercise than in those without exercise ( Table 2). Although left ventricular end-diastolic volume (LVEDV), end-systolic volume, and ejection fraction were not signi cantly different among the four groups before and after the protocol in terms of SPECT data, the logarithmic value of the heart-to-background count ratios such as the H/L ratio and H/below the H ratio was signi cantly higher in groups A + 35 W and A + 45 W than in group A (Fig. 2). BNP, brain natriuretic peptide; ADA, antidiabetic agents; ATA, antithrombotic agents; CCB, calcium-channel blocker; RAS-I, renin-angiotensin system inhibitor. In all the patients, the logarithmic values of H/L and H/below the H ratios positively correlated with heart rate (p = 0.004 and p = 0.008, respectively), rate pressure product (RPP; p = 0.005 and p = 0.008, respectively), and the increment of RPP (delta RPP; p = 0.002 and p = 0.005, respectively) after stress in the univariate analysis ( Table 3). The LVEDV after stress (p = 0.002) negatively correlated with the logarithmic value of H/below the H ratio, but not H/L ratio. Although male sex was independently associated with the logarithmic values of both H/L ratio (p = 0.014) and H/below the H ratio (p = 0.001) in the multivariate regression analysis, LVEDV was also independently associated with the logarithmic value of H/below the H ratio (p < 0.001; Table 3). H/L, heart-to-liver activity ratio; H/below the H, heart-to-below the heart; HR, heart rate; RPP, rate pressure product; EDVex, end-diastolic volume after exercise No signi cant side effects except for mild nausea and hypotension attributable to the ATP infusion in combination with exercise were observed, and our protocol was well tolerated in this study.

Discussion
In this study, we enrolled only patients whose SSSs were < 3 to perform a robust and accurate evaluation of artifacts. The main nding of our study was that concomitant 6-min low-grade exercise supplementation with ≥ 35 W, through the use of the bicycle protocol, increased heart-to-background count ratios such as H/L and H/below the H ratios in the patients who underwent ATP stress MPI. The other ndings of this study were that in the case of the H/below the H ratio, the artifact was independently associated with the LVEDV after exercise. The differences in the hemodynamic parameters such as RPP were not important factors for the evaluation of artefactual grade.
Side effects are common during standard adenosine stress testing. However, few signi cant side effects attributable to the ATP infusion in combination with exercise were observed, and this procedure was well tolerated in this study. Exercise increases the sympathetic nerve activity, which improves atrioventricular conduction and suppresses the vasodilation of adenosine in diaphragmatic lesions [10]. Pannell et al. [3] reported signi cant reductions in noncardiac side effects and major arrhythmias in combination with semi-spine bicycle exercise and adenosine stress. Furthermore, they found that the heart-to-background count ratio was higher in the exercise groups and correlated with the exercise level achieved. Samady et al. found that a 6-min adenosine infusion with concomitant low-level treadmill exercise reduced unfavorable side effects, enhanced image quality, and may have resulted in greater detection of ischemia [11]. Enhancement of image quality was observed when the adenosine infusion study was performed as compared with the bicycle-exercise protocol in patients treated with beta-blockade [12]. However, the opposite results were reported by Jamil et al., who showed no change in defect size or severity compared with the standard 6-min adenosine infusion [13].
A clear increase in H/L ratio is an important nding because an increased H/L ratio has been shown to result in fewer artifacts in the inferior and inferoseptal regions of the heart [1,2]. In our study, we included only patients whose SSSs were < 3 to evaluate precisely the extent of background level. When combined with ATP plus short-duration exercise stress, mild stress was adequate to obtain a signi cant quality image. Some differences were found between the results in the two regions regarding the quanti cation of the target-to-background ratio. The H/below the H ratio represents a comparison of the image quality of the heart with those of organs adjacent to the heart, such as the diaphragm, intestine, and other digestive tissues, and was inversely dependent on LVEDV in association with male sex. Enlargement of the left ventricle may reduce the sharpness of the SPECT image because of the increase in the dispersion of 99m Tc tetrofosmin in case of H/below the H ratio. By contrast, only male sex, irrespective of left ventricular size, was an independent factor for H/L ratio. Sex-related differences in artifacts may result from the grade of fat tissue and complicated body structures, including the lung, spine, and chest wall. A pertinent question that should be answered in the future is how a higher image quality can be obtained in patients with enlarged left ventricles when the use of a new type of stress method is becoming widespread in myocardial perfusion SPECT.
A considerable interpatient variability exists in the heart-to-background count ratios, but some extent of the exercise level achieved greater heart-to-background count ratio, which did not depend on the extent of the RPP. A short-duration exercise stress of ≥ 35 W may enhance sympathetic activity abruptly and adequately, thus leading to the desirable results. Improvement of the heart-to-background count ratio was present at the lower exercise levels, which most patients could be expected to achieve. Six-min ATP infusion with a moderately low-grade exercise protocol during the same period is preferable because it can be incorporated into the standard drug stress regimen with no time loss.

Limitations
Depending on the protocol design, the type of tracer, moment of tracer injection, and type and duration of exercise may vary [14][15][16][17][18] and elicit different results. Whether these issues have any effects on the accuracy of the perfusion study are currently unclear. We could not use SPECT-computed tomography scanner to reduce artifacts in our study. As we randomly assigned patients into four groups but enrolled only those with SSSs of < 3, the number of patients differed among the groups. Under this condition, the defect severity and sensitivity to coronary artery disease in each group could not be evaluated.

Conclusion
ATP administration concomitant with mild exercise stress of ≥ 35 W was useful for reducing undesired artifacts evaluated on the basis of heart-to-background ratio, irrespective of the extent of the RPP. H/L ratio was associated with male sex, while H/below the H ratio was associated with the LVEDV after stress in addition to male sex. The factors that cause artifacts might be different between target organs.

Declarations
Ethics approval and consent to participate This study complied with the tenets of the Declaration of Helsinki, and the protocol was approved by the ethics committee of our hospital (Ethics Committee of Yao Municipal Hospital). All patients provided written informed consent to participate.

Consent for publication
All authors have read and approved the submission of the manuscript. Figure 1 (a) Heart-to-liver (H/L) ratio: The regions of interest are drawn around the heart and liver, excluding the gall bladder. "Heart counts" and "liver counts" were obtained from the point of the maximum counts in each region of interest. (b) Heart-to-below the heart (H/below the H) ratio: The pro le lines are placed on the heart and below the heart from the bottom of heart image to 10 pixels. "Heart counts" were obtained from the maximum count in the pro le line, and "below the heart counts" were obtained from the minimum count in the pro le line.