Impact of an abdominal belt on breathing patterns and scan efficiency in whole-heart coronary magnetic resonance angiography: comparison between the UK and Japan

We studied 30 patients (15 patients each in the UK and Japan) randomly selected from the patients who were referred for a routine clinical cardiovascular MR scan. Exclusion criteria included patients with general contraindications to MRI (e.g., pacemakers, claustrophobia), abdominal aortic aneurysm, pregnancy, severe pulmonary disease, heart failure (NYHA class III-IV) and any abdominal or thoracic pain. Patients' characteristics are shown in table 1. The local ethics committee of both institutes approved the study, and all patients gave written informed consent to participate.

The MR examination was performed on a commercial 1.5 Tesla MRI unit (Achieva, Philips Healthcare, Best, The Netherlands) equipped with cardiac software (release 2.53) and a commercial gradient system (33 mTm peak on axis, 80 mT/m/ms slew rate). At the end of the clinical routine cardiac MR scan, the navigator acquisitions were performed [15]. Each navigator beam consisted of a cylindrical 2D spiral excitation with four gradient cycles (diameter of 30 mm) with a flip angle of 60°. Temporal resolution of the navigator acquisition in each position was 82 ms. The patients were examined in the common supine position and no breathing commands were given. We used a gradient echo sequence scout to determine the position of the diaphragm, the thorax and the abdominal wall before placing the navigators. One navigator was placed through the dome of the right hemi-diaphragm to detect the cranial-caudal (CC) position of the diaphragm (Figure 1a). A second navigator was placed through the left hemi-diaphragm to also detect the CC position of diaphragm (Figure 1a). A third navigator was placed in anterior-posterior (AP) direction through the right chest wall at the height of the 3rd intercostal space to measure the AP position of the thorax (Figure 1b). A fourth navigator was positioned in left-right (LR) direction through the right chest wall at the height of the 3rd intercostal space to measure the LR position of the thorax (Figure 1c). A fifth navigator was positioned in AP direction through the abdominal wall at the height of the 5 cm below the xiphoid process to measure the AP position of the abdominal wall (Figure 1d). Since the available software only allowed the acquisition of three navigator echoes in one sequence two series with three navigators each were performed. These consisted of CC _{right diaphragm}, CC _{left diaphragm} and AP _thorax and CC _{right diaphragm}, LR _thorax and AP _{abdominal wall}. In each series data was acquired for one-minute. After the two navigator series acquisition, the phased array coil was removed from the patient. The 20-cm-wide tight-fitting BELT was wrapped around the patient's abdomen in expiration to suppress breathing related motion of the diaphragm (Figure 2) then the phased array coil was replaced on the patient's chest. A gradient echo sequence scout was then performed to position the navigators. Two further series of the navigator acquisition were performed using the same navigator positions as before the application of the BELT i.e. CC _{right diaphragm}, CC _{left diaphragm} and AP _thorax and CC _{right diaphragm}, LR _thorax and AP _{abdominal wall}.

After the scan, each patient completed the following brief questionnaire; how long would you tolerate the BELT if a future scan was needed? (A. 0 min, B. 0-10 min, B. 10-20 min, C. 20-30 min, D. > 30 min). We recorded the time of belt administration individually in each patient.

Non-contrast enhanced whole-heart coronary MRA was obtained in 6 normal volunteers (3 subjects in the UK and in Japan, respectively) with and without application of the BELT. Imaging was performed at 1.5-T (Achieva, Philips Medical Systems, Best, the Netherlands) using a 32-channel cardiac surface coil [16]. Drift correction was employed.

The navigator data were exported from the console computer and were converted into PASW Statistics, version 18.0.2, software (SPSS, Chicago, Ill) for further processing. The positioning of diaphragm, thorax and abdominal wall was determined on the basis of the original navigator data. The diaphragmatic end expiratory position (EEP), end inspiratory position (EIP) and the duration of end expiration (EED) were calculated with histograms for steps of 1 mm on the basis of the diaphragmatic CC navigator positioned through the right diaphragm [15](Figure 3). The diaphragmatic position which occurred most frequently for each breathing cycle was defined as the EEP. The most distant position from EEP with at least 2 counts on the histogram was defined as the EIP. The EED was determined from the EEP values of the histogram for each breathing cycle. The EED was defined as the number of counts within the EEP interval. Each count corresponds to 82 ms. EEP-EIP was defined as the distance from EEP to EIP. Breathing cycle and respiration rate were calculated from the navigator data. Five-millimeter gating window was defined as the range between the lines of ± 2.5 mm up and down from the averaged EEP. The duration within the gating window in each respiration and that in one minute was calculated by counting the navigator position within the gating window. Scan efficiency (SE) was calculated as follows; SE = [the duration within 5 mm gating window per minutes]/[cardiac cycle]/[heart rate]. The different navigator echo signals were correlated with each other. The relationship between the different navigator signals was assessed by regression analysis. The slope and regression coefficient of this graph was obtained as a correction factor and a determinant of the relative movement respectively [15]. The correction factors for the thoracic AP and LR motion and abdominal AP motion in relation to right diaphragmatic CC motion after the application of the BELT were corrected by considering the decrease of EEP-EIP as follows; cCF = (100+%increase) × CF/100, where cCF is corrected correction factor, CF is corrected factor and %increase was defined as; (parameter _belt - parameter _pre)/parameter _pre × 100.

Whole-heart coronary MRA images were analyzed using an image analysis workstation (Virtual Place, Aze, Tokyo, Japan). Curved multi-planar reconstruction (MPR) images were generated on the workstation.

Statistical analysis was performed using PASW Statistics, version 18.0.2, software (SPSS, Chicago, Ill). For all continuous parameters, results are given as the mean ± standard deviation. For continuous variables, Shapiro-Wilk test was performed to see whether or not the variables were skewed. For continuous variables with normal distribution, parametric tests were applied. For discrete variables and skewed continuous variables, non-parametric statistics were applied. To detect statistically significant differences within a group, the paired Student t test or the Wilcoxon signed-rank test was used. Differences between two groups were tested by applying the unpaired Student t test or the Mann-Whitney U test. The correlation between age, sex, height, weight, BSA or abdominal circumference and differences in EEP-EIP, RR or Scan efficiency between before and after the BELT was evaluated. Pearson correlation or Spearman's rank correlation was used to test the relationship between two variables. All tests were two-tailed, and P < 0.05 was considered to indicate a statistically significant difference.

Several study limitations should be acknowledged. Firstly, in this study, we only studied the navigator position rather than acquiring the actual whole-heart coronary MRA images. However we have demonstrated the feasibility of the BELT to improve image quality and scan duration in 3 respective MRA volunteer scans in the UK and in Japan. Secondly, navigator acquisition was done only for one minute. Due to these limitations, scan efficiency obtained in this study cannot be simply applied to a real WHCMRA scan. In future studies, whole-heart coronary MRA should be performed using the BELT technique employed in the present study. Scan duration, image quality, scan efficiency have to be compared before and after the administration of the BELT. Thirdly, direct influences of the BELT on cardiac motion were not evaluated in this study. Wide variation in the actual correction factor among individuals was reported indicating that usage of a fixed correlation factor will result in subject-dependent residual errors in heart position estimates [18]. Further investigation is required to evaluate if the BELT can reduce the variation of the correction factor among individuals.