Cardiac MRI Assessment of Left and Right Ventricular Parameters in Healthy Australian Normal Volunteers

Cardiac magnetic resonance imaging (MRI) is being utilised increasingly for the purposes of cardiovascular imaging. Limited data suggest a high degree of reproducibility for parameters such as left ventricular (LV) ejection fraction (EF), mass, end-diastolic and end-systolic volumes (EDV and ESV). We sought to investigate reproducibility and establish means for these parameters in a selected normal non-Aboriginal Australian population, using cardiac MRI. Sixty normal volunteers underwent cardiac MRI investigation using a 1.5 T MRI system. Steady state free precession imaging was performed with short axis cine images through the left ventricle obtained. All images were acquired with cardiac gating. Two independent observers then analysed the data set. Data were collected for assessment of left ventricular EF, EDV, ESV, mass and right ventricular volumes. Data are presented as mean ± S.D. Total imaging time was approximately 15 min. All patients were able to complete the full protocol. Left ventricular parameters: EF 58.5 ± 8.0%, LV mass 114.2 ± 40.6 g, EDV 117.3 ± 33.4 mls and ESV 50.0 ± 22.2 mls. Right ventricular parameters: EF 45.6 ± 11.6%, EDV 163.5 ± 52.2 mls and ESV 89.5 ± 34.3 mls. Intraclass correlation coefficients for LV: EF 0.84, LV mass 0.84, EDV 0.85 and ESV 0.89. Cardiac MRI provides high quality information about cardiac function with a high level of reproducibility. Cardiac MRI parameters in a normal non-Aboriginal Australian population are provided. Keywords Cardiac MRI Ventricular function Normal volunteers Introduction Magnetic resonance imaging (MRI) has been shown to be an accurate and reproducible tool for the assessment of left and right ventricular (RV) measurements. 1–5 The information that can be obtained in a single examination includes ventricular volumes, mass and ejection fraction (EF) which provide both diagnostic and prognostic information. 3 MRI as with echocardiography, is a non-invasive investigation and does not utilise ionising radiation. However, the additional advantage is the higher inter-observer reproducibility compared to echocardiography, as MRI does not rely on geometric assumptions. This allows the serial assessment of ventricular parameters particularly the monitoring of ventricular remodelling with therapy. Earlier cardiac MRI studies for ventricular function and volumes utilised a technique with a pulse sequence using segmented k-space turbo gradient echo. 2 This has remained in current use, but a more recent technique is the steady-state free precession (SSFP) technique. This technique has been validated in animal studies. 6 Compared to the gradient echo technique, SSFP allows improved definition of the endocardial borders and faster acquisition time. 2,7 In this study, we sought to establish normal ranges for left ventricular (LV) volumes, mass and ejection fraction as well as right ventricular volumes in normal healthy non-Aboriginal Australian volunteers. We also sought to establish that the assessment of these parameters using cardiac MRI with SSFP is reproducible. Methods Study Population Sixty (41 males, 19 females, mean age 51 ± 13 years) consecutive normal healthy volunteers with no history of cardiovascular or respiratory disease and a normal resting blood pressure and a normal resting electrocardiogram were studied. All volunteers gave their informed consent for the study. Ethics approval was obtained from the institutional Research Ethics Committee (Royal Adelaide Hospital). Exclusion criteria were contraindications to MRI scanning and arrhythmia. Image Acquisition All cardiac MR imaging studies were performed with subjects in the supine position using a 1.5 T MRI scanner (Siemens Sonata, Germany) and a phased array surface coil. Long-axis reference views were used for positioning the 8–12 perpendicular LV short-axis slices from the level of the mitral valve to the left ventricular apex. Images were obtained during breath-hold (8–10 s) with prospectively ECG-gated True FISP (Fast Imaging with Steady-State Precision) sequences. The short axis section thickness were 6 mm with intersection gaps of 4 mm, based on previous published studies. 1,2,5 Acquisition time was 90% of the RR-interval, image matrix 256 × 150, field of view 380 mm, repetition time 52.05 ms, echo time 1.74 ms and flip angle 70°. 12–17 heart phases were acquired per repetition time interval. Image Analysis Ventricular analysis was performed off-line with a proprietary software program (Argus software, Siemens Medical Solutions, Germany). For the left ventricular data set, short-axis endocardial and epicardial contours were manually traced in end-diastole (start of R-wave) and in end-systole (smallest cavity area). Papillary muscles and trabeculations were excluded from the ventricular volume and were included if contiguous with the myocardial mass. The basal slice was selected as the slice where the blood volume was surrounded by >50% of ventricular myocardium. 1,2 For the right ventricular data set, one observer manually traced the endocardial contours at end-diastole and at end-systole. The selection of the RV basal slice was based on published methods where, if the pulmonary valve was seen, only the portion of the volume below the level of the pulmonary valve was included. 2 Both LV and RV end-diastolic and end-systolic cavity surface areas were summed up and volumes: end-diastolic (EDV) and end-systolic (ESV) estimated by multiplying with interslice intervals as per Simpson's rule. Ejection fraction was calculated as EF = [(EDV − ESV)/EDV] × 100 (%). LV mass = 1.05 × (epicardial volume − endocardial volume). RV mass was not measured. Reproducibility To assess the interobserver variability, a second independent observer re-measured the data sets for both left and right ventricular parameters. Statistical Analysis Data was collected for assessment of left ventricular EF, mass, EDV and ESV, and presented as mean ± standard deviation (S.D.). The normal range was calculated as 2S.D. above the mean. Interobserver variability was assessed using the Bland and Altman method 8 as well as intraclass correlation coefficients. The correlation calculations is a statistical method to quantify the relationship between two variables measured independently and where r = 1 is taken as a perfect correlation. Unpaired sample t -tests were performed for the difference in measurements between males and females with the level of significance taken as p < 0.05%. Statistical analyses were performed using SPSS software. Results Total imaging time was approximately 15 min. All patients were able to complete the full protocol and diagnostic imaging was obtained in all subjects. The left ventricular parameters were: EF 58.5 ± 8.0%, LV mass 114.2 ± 40.6 g, EDV 117.3 ± 33.4 mls and ESV 50.0 ± 22.2 mls. The right ventricular parameters were: EF 45.6 ± 11.6%, EDV 163.5 ± 52.2 mls and ESV 89.5 ± 34.3 mls. Fig. 1 shows a mid-ventricular short axis SSFP image in end-diastole and end-systole with left ventricular endocardial and epicardial contours traced and right ventricular endocardial contour traced. Tables 1 and 2 show the means ± standard deviation for all LV and RV volumes for both males and females. Intraclass correlation coefficients between two independent observers for LV parameters were: EF 0.82, LV mass 0.94, EDV 0.95 and ESV 0.91. Results from the Bland and Altman analyses for interobserver and intraobserver variability for LV and RV are presented as mean, the difference between the means ± 2S.D. Interobserver mean of LV EF was 56.6% with difference between the means of 0.7 ± 8.6%. Interobserver mean of RVEF was 51.1%, difference between the means −1.7 ± 22.7%. This is presented graphically for ejection fraction in Fig. 2 A and B for LV inter- and intraobserver data and in Fig. 3 A and B for RV inter- and intraobserver data. Interobserver LV EDV mean 124, difference 10.8 ± 33 mls, LV ESV mean 54.1, difference 3.3 ± 18.0 mls and LV mass mean 107, difference 5.24 ± 37 g. Interobserver RV EDV mean 151.8, difference 31.5 ± 48.4 mls and RV ESV mean 76.0, difference 18.21 ± 49 mls. There was a statistically significant difference in volumes and mass between males and females. This difference was present for LV EDV ( p = 0.01), RV EDV ( p < 0.001), RV ESV ( p < 0.01) as well as LV mass ( p < 0.0001). The difference between male and female LV mass measurements remained significant after indexation to body surface area ( p = 0.008) but not the difference between the measurements for LV EDV and RV EDV. There was no significant difference in LV EF between males and females. Discussion Cardiac MRI is a non-invasive and accurate method for assessing ventricular volumes, mass and function. This study has established the normal ranges for left and right ventricular parameters in a group of normal healthy Australian volunteers using cardiac MRI. We have used the more recent technique with steady-state free precession imaging, which allows for improved endocardial and epicardial border definition. The slice thickness and inter-slice gaps used (left ventricular slice thickness of 6 mm with a 4 mm inter-slice gap) were chosen on the basis of published studies of cardiac MRI assessment of LV function, volume and mass. 2,5 Although thinner slices without slice gaps may increase the accuracy of the ventricular imaging, problems with cross-excitation between slices and signal-to-noise become significant. Papillary muscles not contiguous with the myocardial mass were excluded in keeping with published studies, due to the variability created in attempting to include this. 1 However, new software options may allow inclusion of the papillary muscles in the future. The excellent reproducibility between observers using this technique for assessment of left ventricular parameters allows it to be a reliable tool for both baseline ventricular assessments as well as for serial measurements. This will be particularly useful for assessing ventricular remodelling with therapy in both clinical and research settings. In this study, we found that there was greater interobserver and intraobserver variability in the right ventricular parameters. This could be due to factors related to the shape of the RV, selection of the basal slice as well as the irregular surface of the RV endocardial surface due to trabeculations, which results in some difficulty in tracing the surfaces. These factors may be improved with further software development to automate boundary detection and also assist in the selection of the basal slice during analyses. Our results were comparable to the results obtained with SSFP sequences in normal populations except for slightly smaller left ventricular end-diastolic volumes in our group. This could be due to the known difficulty in selecting the basal slice for inclusion in ventricular volume analysis. Previous MRI studies on normal ventricular parameters have predominantly used the older gradient echo pulse sequence, although there are studies comparing it to the newer SSFP sequences. 2,7 The differences are small, but consistent between the techniques, and may relate to variations in imaging characteristics of the blood–myocardial interface with different sequences. In one study, the LV EDV and LV ESV were larger and LV mass was smaller with SSFP imaging than with the older sequences. 7 However, this does highlight the need for robust population based “normal” values for the different techniques, and furthermore highlights that echocardiography based indices should not be used for cardiac MRI normal indices. Our results demonstrate a statistically significant difference between male and female LV and RV volumes as well as in the LV mass, but no difference in the LV EF. This has also been previously described. When indexed to body surface area, the gender difference in LV mass persisted, but not the ventricular volumes. Conclusion Cardiac magnetic resonance imaging provides high quality information about cardiac structure and function, with a high level of reproducibility. Normal values for the non-Aboriginal Australian population are provided. Further work to define such values in Aboriginal Australians is required. References 1 C.H. Lorenz E.S. Walker V.L. Morgan S.S. Klein T.P. Graham Jr. Normal human right and left ventricular mass, systolic function, and gender differences by cine magnetic resonance imaging J Cardiovasc Magn Reson 1 1 1999 7 21 2 K. Alfakih S. Plein H. Thiele T. Jones J.P. Ridgway M.U. Sivananthan Normal human left and right ventricular dimensions for MRI as assessed by turbo gradient echo and steady-state free precession imaging sequences J Magn Reson Imaging 17 3 2003 323 329 3 F. Grothues G.C. Smith J.C. Moon N.G. Bellenger P. Collins H.U. Klein D.J. Pennell Comparison of interstudy reproducibility of cardiovascular magnetic resonance with two-dimensional echocardiography in normal subjects and in patients with heart failure or left ventricular hypertrophy Am J Cardiol 90 1 2002 29 34 4 F. Grothues J.C. Moon N.G. Bellenger G.S. Smith H.U. Klein D.J. Pennell Interstudy reproducibility of right ventricular volumes, function, and mass with cardiovascular magnetic resonance Am Heart J 147 2 2004 218 223 5 L.E. Hudsmith S.E. Petersen J.M. Francis M.D. Robson S. Neubauer Normal human left and right ventricular and left atrial dimensions using steady state free precession magnetic resonance imaging J Cardiovasc Magn Reson 7 5 2005 775 782 6 D.S. Fieno W.C. Jaffe O.P. Simonetti R.M. Judd J.P. Finn TrueFISP: assessment of accuracy for measurement of left ventricular mass in an animal model J Magn Reson Imaging 15 5 2002 526 531 7 J.C.C. Moon C.H. Lorenz J.M. Francis G.C. Smith D.J. Pennell Breath-hold FLASH and FISP cardiovascular MR imaging: left ventricular volume differences and reproducibility Radiology 223 3 2002 789 797 8 J.M. Bland D.G. Altman Statistical methods for assessing agreement between two methods of clinical measurement Lancet 1 8476 1986 307 310