Cardiac MRI is a rapidly evolving modality in veterinary diagnostic imaging.1 Authors have used calculations of SV based on morphological MRI as a criterion-referenced standard to test new ultrasonography2–4 and CT5 techniques. Two-dimensional PCA is a validated and established technique for use in human cardiac MRI6,7; the technique allows measurement of cardiac output, shunt fraction, collateral flow around obstructions, and flow differences in the left and right pulmonary arteries.8 Use of 2-D PCA also provides information regarding stenotic flow, similar to information provided by Doppler examinations.6,7,9 Phase-contrast angiography involves the use of flow-induced phase effects after application of a bipolar pulse in gradient echo sequences. The sum of the velocities in a region of interest multiplied by the area in each phase of the cardiac cycle represents the flow volume per cardiac cycle.7,10
A basic challenge in cardiac MRI is the acquisition of high-quality images with minimal blood flow artifacts. In contrast to adult human cardiac MRI, but similar to pediatric cardiac MRI, anesthesia is required in veterinary patients for basic immobilization and also for the induction of breath holding. Recently, the effect of different anesthesia protocols on cardiac flow and quantification of mitral valve regurgitant fraction in veterinary medicine have been reported.11,12 The circulatory situation of cardiac patients and the effects of sedative drugs on cardiovascular parameters require close attention.11,13
In a closed cardiovascular system, results of blood flow quantification for the aortic, pulmonary, mitral, and tricuspid valves should theoretically be the same and should yield the same results as for morphological quantification of cardiac action. Flow-sensitive sequences measure ventricular inflow and outflow, respectively. Morphological techniques rely on changes in ventricular volume over time that result in SV when measured over an entire cardiac cycle.7
The purpose of the study reported here was to compare results of SV calculated on the basis of cardiac MRI morphology with SV calculated on the basis of 2-D flow-sensitive sequences in healthy dogs. The hypothesis was that quantification of SV calculated on the basis of blood flow through the aortic, pulmonary, mitral, and tricuspid valves would yield the same, reproducible results as quantification of SV calculated on the basis of morphology. Another hypothesis was that SV calculations would not differ between the left and right ventricles.
Materials and Methods
Animals
Ten healthy Beagles (4 males and 6 females) were included in the study. Mean ± SD age of the dogs was 3.8 ± 0.9 years, and mean body weight was 13.3 ± 2.3 kg. The dogs were considered to be healthy on the basis of results of a clinical examination, hematologic analysis, urinalysis, and echocardiography performed by a board-certified veterinary cardiologist (JNM) and a resident in a veterinary cardiology training program (MBT). All procedures were conducted in accordance with guidelines established by the Animal Welfare Act of Switzerland, and the Cantonal Veterinary Office approved the study design (license No. 144/2013).
Experimental procedures
A resident in a veterinary anesthesia training program (IUC) was responsible for the anesthetic procedures. Dogs were premedicated with butorphanol tartratea (0.2 mg/kg, IM). A catheter was placed in a cephalic vein of each dog. Anesthesia was induced with propofolb (4 to 6 mg/kg, IV) and maintained with isofluranec combined in oxygen and air (fraction of inspired oxygen, 0.6 to 0.7). From the time of placement of the catheter in a cephalic vein until recovery from anesthesia, the dogs received lactated Ringer solution.d Positive-pressure ventilatione with a volume-controlled mode (tidal volume, 10 to 15 mL/kg) was performed, with adaptation of the respiratory rate to achieve an end-tidal partial pressure of CO2 of 40 mm Hg. Several factors had potential impacts on ventricular dimensions; thus, flow velocities, SV, heart rate, and arterial blood pressure were constantly monitored and recorded every 5 minutes to enable investigators to recognize possible influences of these factors on the various measurements during anesthesia.
Cardiac MRI and PCA were performed twice on each dog. Cardiac MRI was performed with a 3-T systemf by a board-certified veterinary radiologist (MD). Cardiac action was monitored with a wireless vector cardiography unit,g with 4 MRI-safe electrodesh attached on the sides of the chest wall over the heart (2 electrodes were attached on each side).
Breath holding during image acquisition was used to prevent cardiac MRI motion artifacts. A calibrated monitori that included spirometry and capnography was used to guide breath holding. Dogs were first hyperventilated (by increasing the respiratory rate) to reduce the end-tidal partial pressure of CO2 to 35 mm Hg. Then, expiratory apnea was induced by discontinuing use of the ventilator for the duration of the acquisition of a single slice.
A stack of 2-D gradient echo slices (time to echo, 2 milliseconds; time to repetition, 4 milliseconds; turbo field echo factor, 8; number of signal averages, 1; sensitivity encoding; echo train length, 6; slice thickness, 4 mm; space between slices, 5 mm; field of view, 200 × 181; flip angle, 45°; and matrix, 200 × 190) was obtained for the short-axis plane (perpendicular to the left ventricular long axis) and provided ventricular volume from the apex to the base of the heart. Additional 4-chamber and left ventricular 2-chamber views acquired with the same scan settings provided information for the long-axis plane. In addition, 2-D PCA scans (time to echo, 2.9 milliseconds; time to repetition, 4.7 milliseconds; turbo field echo factor, 2; number of signal averages, 2; sensitivity encoding; echo train length, 2; slice thickness, 8 mm; field of view, 180 × 180; flip angle, 10°; and matrix, 120 × 120) were acquired in the aortic, pulmonary, mitral valve, and tricuspid valve planes perpendicular to the direction of the blood flow. These scan settings allowed the cardiac cycle to be divided into 30 phases for the morphological sequences and 40 phases for the flow-sensitive sequences.
The DICOM images were exported to a workstation.j The cine echo gradient sequences were evaluated by use of cardiac MRI analysis softwarek and the PCA with suitable software.l Analysis of all images was performed by a board-certified veterinary radiologist (MD).
The SV was first calculated on the basis of morphological changes over 1 cardiac cycle for the short-axis (left and right ventricle) and long-axis (only left ventricle) planes. Endo- and epicardial contours of both ventricles were manually drawn in each short-axis slice and each phase (Figure 1). The stack of short-axis slices covered the heart from the apex to the atrioventricular valve plane (last slice at which the ventricular blood lumen was surrounded by at least 50% myocardium). The right ventricular outflow tract was included in the volume as long as it remained visibly connected with the right ventricular lumen over the entire heart cycle. The software did not allow a separate analysis of papillary muscle volume; therefore, papillary muscle volume was included in the ventricular cavity.7,14 Left ventricular volumes were calculated for the long-axis plane. Endo- and epicardial contours were manually drawn in the 4-chamber and left ventricular 2-chamber views at end diastole and end systole. The software automatically calculated ejection fraction, SV, and cardiac output.
For calculations made on the basis of blood flow, regions of interest were manually drawn in each phase of the corresponding plane for the aortic, pulmonary, mitral, and tricuspid valves (Figure 2). These regions of interest defined the area for blood flow quantification.
The second cardiac MRI was performed on each dog within 2 weeks after the first examination. Results for both examinations were used to assess repeatability.
Statistical analysis
One investigator (HR) performed the statistical analysis with commercially available statistical software.m Descriptive statistics included the mean, SD, median, and range. Normal distributions for all values were tested by use of histograms, Q-Q plots, and the Shapiro-Wilk test. Differences between SV calculated by use of the various methods were compared with a Wilcoxon signed rank test. Significance was set at values of P < 0.05. The CV was calculated to express the precision and repeatability of the measurements obtained for all dogs (intraindividual variability).
Results
The first examination could not be performed on 1 dog because of hardware problems unrelated to the examination technique or dog. For 2 dogs, blood flow measurements for the pulmonary valve obtained during the first examination had to be excluded because of inappropriate plane orientation attributable to operator error.
Mean ± SD heart rate of the dogs was 88.2 ± 14.1 beats/min. Heart rate of each dog changed only minimally throughout the study, with an SD of 5.2 beats/min (minimal SD, 1.8 beats/min; maximal SD, 11.2 beats/min). Heart rates of the dogs were highly variable (minimal SD, 12.5 beats/min; maximal SD, 16.4 beats/min). However, heart rates differed only minimally between the first and second examinations (CV = 0.13).
Mean ± SD arterial blood pressure was 81 ± 5 mm Hg, and it was always ≥ 60 mm Hg throughout the anesthetic period for all dogs. Mean systolic and diastolic arterial blood pressure was 106 ± 6 mm Hg and 57 ± 6 mm Hg, respectively. Variability of blood pressure measured during the first and second examinations was low for all systolic (CV = 0.08), mean (CV = 0.1), and diastolic (CV = 0.15) blood pressure measurements.
Morphological quantification for the short-axis plane resulted in a mean ± SD SV of 13.4 ± 2.7 mL for the left ventricle, which differed significantly (P = 0.008) from the mean SV of the right ventricle (8.6 ± 2.4 mL). Morphological quantification of left ventricular SV for the long-axis plane yielded mean values of 15.2 ± 3.3 mL for the 4-chamber view, which differed significantly (P = 0.008) from the mean value for the 2-chamber view (20.7 ± 3.8 mL). Furthermore, the calculated left ventricular SV for the short-axis plane was significantly less than the value for the 2-chamber (P = 0.008) and 4-chamber (P = 0.028) views, and the right ventricular SV for the short-axis plane was significantly (P = 0.008) less than the SV for both the 2-chamber and 4-chamber views.
Values for the ventricular dimensions, independent of slice orientation, changed little between the 2 examinations (CV = 0.08). Consequently, SV changed only minimally (CV = 0.14), and no significant differences for results between the first and second examination were identified (Table 1).
Values for cardiac function assessed on the basis of morphological information from MRI scans of 10 healthy Beagles* examined twice (interval of ≤ 14 days between examinations).
Structure | Plane or view | Examination | Variable | End-diastolic volume (mL) | End-systolic volume (mL) | Ejection fraction (%) | Heart rate (beats/min) | SV (mL) | Cardiac output (L/min) |
---|---|---|---|---|---|---|---|---|---|
Left ventricle | SAX | First | Mean ± SD | 28.2 ± 5.9 | 14.6 ± 4.1 | 49.0 ± 5.5 | 91.4 ± 12.7 | 13.7 ± 2.5 | 1.3 ± 0.3 |
Median | 26.4 | 14.6 | 47.8 | 91.0 | 13.3 | 1.3 | |||
Minimum | 20.2 | 9.1 | 39.4 | 75.0 | 9.7 | 0.7 | |||
Maximum | 36.6 | 20.2 | 55.0 | 114.0 | 17.4 | 2.0 | |||
Second | Mean ± SD | 30.2 ± 7.7 | 17.0 ± 6.0 | 44.9 ± 7.2 | 92.8 ± 14.6 | 13.2 ± 2.8 | 1.2 ± 0.3 | ||
Median | 28.8 | 15.5 | 46.3 | 92.5 | 12.6 | 1.2 | |||
Minimum | 20.2 | 9.9 | 29.4 | 68.0 | 10.0 | 0.9 | |||
Maximum | 42.4 | 27.4 | 55.3 | 129.0 | 18.5 | 1.9 | |||
Left ventricle | 4-chamber | First | Mean ± SD | 28.7 ± 9.6 | 15.4 ± 5.6 | 51.6 ± 7.8 | 90.9 ± 13.7 | 15.8 ± 2.5 | 1.4 ± 0.2 |
Median | 28.0 | 14.5 | 55.8 | 89 | 15.6 | 1.4 | |||
Minimum | 10.3 | 8.5 | 34.6 | 72 | 12.2 | 1.1 | |||
Maximum | 42.2 | 24.3 | 62.8 | 118 | 20.7 | 1.7 | |||
Second | Mean ± SD | 32.3 ± 7.3 | 17.4 ± 6.5 | 47.5 ± 12.1 | 90.2 ± 14.4 | 15.0 ± 3.9 | 1.3 ± 0.3 | ||
Median | 23.3 | 16.9 | 47.4 | 89.5 | 14.6 | 1.3 | |||
Minimum | 22.2 | 7.2 | 30.2 | 66.0 | 9.1 | 1.0 | |||
Maximum | 44.8 | 31.2 | 69.5 | 117.0 | 21.5 | 1.8 | |||
Left ventricle | 2-chamber | First | Mean ± SD | 37.1 ± 9.8 | 15.8 ± 7.4 | 59.3 ± 8.5 | 90.7 ± 14.7 | 21.3 ± 3.3 | 2.0 ± 0.5 |
Median | 34.0 | 12.9 | 62.1 | 88.0 | 20.4 | 2.0 | |||
Minimum | 27.7 | 9.2 | 40.7 | 70.0 | 17.6 | 1.3 | |||
Maximum | 57.9 | 29.6 | 67.5 | 115.0 | 28.3 | 3.3 | |||
Second | Mean ± SD | 36.2 ± 7.0 | 16.0 ± 6.2 | 56.9 ± 11.4 | 88.8 ± 12.8 | 20.2 ± 4.1 | 1.8 ±0.4 | ||
Median | 35.4 | 14.7 | 63.1 | 88.0 | 19.7 | 1.8 | |||
Minimum | 26.7 | 7.1 | 37.0 | 66.0 | 13.7 | 1.3 | |||
Maximum | 48.4 | 25.3 | 73.5 | 117.0 | 28.4 | 2.5 | |||
Right ventricle | SAX | First | Mean ± SD | 33.1 ± 7.7 | 24.2 ± 6.3 | 27.1 ± 4.2 | 91.4 ± 12.7 | 8.9 ± 2.3 | 0.8 ± 0.3 |
Median | 31.4 | 22.8 | 27.3 | 91.0 | 5.6 | 0.7 | |||
Minimum | 22.2 | 16.2 | 20.6 | 75.0 | 6.0 | 0.5 | |||
Maximum | 43.7 | 34.7 | 33.9 | 114.0 | 13.1 | 1.4 | |||
Second | Mean ± SD | 35.5 ± 8.7 | 26.9 ± 7.5 | 24.5 ± 5.9 | 92.8 ± 14.6 | 8.5 ± 2.4 | 0.8 ± 0.2 | ||
Median | 36.5 | 27.0 | 25.2 | 92.5 | 7.3 | 0.7 | |||
Minimum | 23.6 | 16.0 | 15.8 | 68.0 | 6.5 | 0.5 | |||
Maximum | 46.7 | 37.5 | 32.1 | 129.0 | 13.1 | 1.3 |
The first examination could not be performed on 1 dog because of hardware problems.
SAX = Short-axis plane.
Mean ± SD values for SV calculated on the basis of blood flow were 17.8 ± 4.1 mL for the aortic valve, 17.2 ± 5.4 mL for the pulmonary valve, 17.2 ± 3.9 mL for the mitral valve, and 16.9 ± 5.1 mL for the tricuspid valve. Although the degree of individual variation for blood flow measurements was larger than that for the morphological examinations, results of flow quantification at the various locations were more evenly distributed, and no significant differences were detected in SV calculated on the basis of blood flow measurements. There was little variation (CV = 0.17) in the measurement of blood flow velocities for the vessels and cross-sectional areas of the atrioventricular valves between the first and second examinations. No significant difference was found in the results of any of the variables between the first and second examinations (Table 2).
Values for cardiac function assessed on the basis of PCA in 10 healthy Beagles* examined twice (interval of ≤ 14 days between examinations).
Structure | Examination | Variable | SV (mL) | Mean flux (mL/s) | Mean velocity (cm/s) | Area at diastole (cm2) | Area at systole (cm2) | Heart rate (beats/min) | Cardiac output (L/min) |
---|---|---|---|---|---|---|---|---|---|
Aortic valve | First | Mean ± SD | 18.9 ± 3.2 | 28.5 ± 5.4 | 16.2 ± 3.3 | 2.7 ± 0.7 | 1.8 ± 0.4 | 91.1 ± 13.4 | 1.7 ± 0.3 |
Median | 18.4 | 30.3 | 14.2 | 2.4 | 1.7 | 98.0 | 1.9 | ||
Minimum | 15.6 | 19.7 | 12.5 | 2.0 | 1.3 | 68.0 | 1.2 | ||
Maximum | 25.8 | 36.7 | 21.7 | 4.0 | 2.4 | 108.0 | 2.2 | ||
Second | Mean ± SD | 16.9 ± 4.6 | 23.8 ± 5.8 | 13.3 ± 3.8 | 1.95 ± 0.8 | 1.8 ± 0.6 | 85.9 ± 12.4 | 1.5 ± 0.3 | |
Median | 15 | 23.4 | 14.2 | 1.7 | 1.6 | 85.5 | 1.5 | ||
Minimum | 12.1 | 16.2 | 5.3 | 1.2 | 1.3 | 69.0 | 1.1 | ||
Maximum | 25.4 | 34.4 | 18.6 | 3.7 | 3.3 | 113.0 | 2.1 | ||
Mitral valve | First | Mean ± SD | 17.6 ± 2.7 | 25.6 ± 5.6 | −6.4 ± 1.2 | 4.9 ± 1.2 | 4.7 ± 0.8 | 87.2 ± 14.7 | 1.6 ± 0.4 |
Median | 18.3 | 27.3 | −5.9 | 4.6 | 4.6 | 92.0 | 1.7 | ||
Minimum | 12.5 | 13.6 | −8.1 | 3.4 | 3.8 | 65.0 | 0.9 | ||
Maximum | 21.7 | 32.8 | −4.9 | 7.5 | 6.1 | 107.0 | 2.1 | ||
Second | Mean ± SD | 16.9 ± 4.7 | 23.1 ± 5.5 | −5.5 ± 1.0 | 5.1 ± 1.0 | 4.4 ± 0.7 | 83.7 ± 11.6 | 1.5 ± 0.4 | |
Median | 14.9 | 21.8 | −5.2 | 5.1 | 4.3 | 84.0 | 1.5 | ||
Minimum | 11.8 | 16.8 | −6.9 | 3.9 | 3.4 | 67.0 | 1.1 | ||
Maximum | 23.7 | 33.8 | −3.9 | 6.3 | 5.5 | 107.0 | 2.2 | ||
Pulmonary valve | First | Mean ± SD | 15.9 ± 3.7 | 21.7 ± 7.2 | 8.6 ± 2.9 | 2.3 ± 0.5 | 2.3 ± 0.5 | 83.4 ± 12.9 | 1.3 ± 0.4 |
Median | 16.3 | 20.7 | 8.3 | 2.2 | 2.0 | 87.0 | 1.3 | ||
Minimum | 9.9 | 9.7 | 4.6 | 1.8 | 1.7 | 65.0 | 0.6 | ||
Maximum | 20.1 | 31.8 | 12.5 | 3.3 | 3.3 | 101.0 | 2.0 | ||
Second | Mean ± SD | 18.4 ± 5.9 | 25.0 ± 6.7 | 9.7 ± 2.0 | 2.5 ± 0.4 | 2.4 ± 0.4 | 83.0 ± 11.5 | 1.5 ± 0.4 | |
Median | 16.0 | 23.5 | 9.0 | 2.5 | 2.4 | 82.5 | 1.4 | ||
Minimum | 11.8 | 17.8 | 7.1 | 1.9 | 1.6 | 68.0 | 1.1 | ||
Maximum | 31.3 | 40.2 | 13.5 | 3.2 | 3.0 | 102.0 | 2.4 | ||
Tricuspid valve | First | Mean ± SD | 16.1 ± 3.3 | 23.3 ± 6.2 | −5.8 ± 1.3 | 4.4 ± 0.9 | 4.6 ± 1.2 | 86.2 ± 13.6 | 1.4 ± 0.4 |
Median | 16.4 | 23.4 | −5.9 | 4.1 | 4.5 | 90.0 | 1.4 | ||
Minimum | 10.4 | 14.1 | −8.0 | 3.5 | 2.8 | 65.0 | 0.9 | ||
Maximum | 21.8 | 33.4 | −3.7 | 6.6 | 6.4 | 105.0 | 2.1 | ||
Second | Mean ± SD | 17.6 ± 6.2 | 23.5 ± 6.0 | −5.9 ± 1.4 | 4.8 ± 0.7 | 4.7 ± 0.8 | 82.8 ± 12.4 | 1.4 ± 0.3 | |
Median | 16.2 | 22.5 | −5.7 | 4.9 | 4.4 | 85.0 | 1.4 | ||
Minimum | 11.0 | 15.3 | −8.8 | 3.4 | 3.6 | 67.0 | 1.0 | ||
Maximum | 29.1 | 34.8 | −3.9 | 5.9 | 6.0 | 105.0 | 2.1 |
The first examination could not be performed on 1 dog because of hardware problems, and blood flow measurements for the pulmonary valve obtained during the first examination for 2 dogs had to be excluded because of inappropriate plane orientation attributable to operator error.
Left ventricular SV for the short-axis plane was significantly less than the SV for the aortic valve (P = 0.008), pulmonary valve (P = 0.011), mitral valve (P = 0.008), and tricuspid valve (P = 0.008). Similarly, right ventricular SV for the short-axis plane was significantly less than the blood flow–based SV (P = 0.008 for all 4 views). Comparison of left ventricular SV for the 2-chamber view differed significantly from blood flow–based volumes for the aorta (P = 0.011), pulmonary valve (P = 0.011), mitral valve (P = 0.008), and tricuspid valve (P = 0.021).
Morphology-based volume calculations for the 4-chamber view did not differ significantly from values calculated on the basis of blood flow through the pulmonary valve (P = 0.11) and tricuspid valve (P = 0.066; Figure 3). However, they differed significantly from values calculated on the basis of blood flow through the aortic valve (P = 0.008) and mitral valve (P = 0.015).
Discussion
The present study was conducted to compare SV calculated on the basis of morphology in various planes with SV calculated on the basis of PCA results for 4 cardiac valves of the canine heart. The hypothesis that quantification of SV did not depend on technique was rejected because of significant differences between the results of morphology-based versus flow-based SV calculations. However, the hypothesis that SV did not differ on the basis of location was valid for PCA, whereby SV values differed little among the aortic, pulmonary, mitral, and tricuspid valves. Furthermore, there was minimal day-to-day variability.
Ventricular function is a complex interaction of narrowing, shortening, and twisting.15 Calculation of SV on the basis of morphology in short- and long-axis planes relies on geometric models that do not include the contribution of motion perpendicular to the imaging plane (through-plane motion) to total SV.16 The investigated models made the most substantial geometric assumptions for calculation of SV for morphology of long-axis planes, which may explain the higher SV than that for the short-axis planes. Long-axis planes are best for revealing motion of the heart base and apex throughout the cardiac cycle, but the risk for partial volume effects is greater than for short-axis planes.7 Calculations of SV on the basis of blood flow measurements for the cardiac valves provided reproducible and similar results. Flow quantification can be used to measure through-plane motion and allows for adjustments of location and cross-sectional diameter of the region of interest for compensation of in-plane motion and vessel or valve dynamics, respectively. This may explain the robustness of the calculated values, compared with the values for the morphological calculations. Local independence of blood flow quantification is important for the measurement of pressure gradients, quantification of shunts, and evaluation of valvular regurgitation and stenosis.8 Indeed, the marked difference between flow-based and morphology-based SV calculations was a striking finding in the present study. It was expected that MRI would allow exact 3-D measurements of the left and right ventricles. Of particular interest were measurements of right ventricular volumes because the right ventricle essentially cannot be reliably and reproducibly assessed echocardiographically.5 Although there was no criterion-referenced standard for the present study that would have allowed a definitive indication of the technique that was most exact, the fact that the largest variability was between the 3 left ventricular morphology-based measurements, especially between left and right ventricular volumes, implied that morphology-based measurements were much less exact than expected. In view of all other results, it appeared that measurements of right ventricular volumes in particular were insufficiently exact (ie, falsely low values) for the study reported here. The twisted orientation of the right ventricular outflow tract with its great range of motion made it difficult to include it into calculations. In a study5 conducted to compare 3-D ultrasonography and cardiac MRI, calculations of right ventricular SV on the basis of short-axis morphology (including the right ventricular outflow tract) were in the range of flow measurements obtained for the present study. Therefore, a likely explanation for small right ventricular volumes in the present study would have been partial exclusion of the right ventricular outflow tract.
Results of the present study can be compared with results of morphological studies conducted to investigate left ventricular SV,2–4 right ventricular SV,5,17 and anesthetic protocols.11 Some results are similar, but there also are some important differences worthy of mention. With regard to similarities, basic cardiovascular variables, including heart rate and systemic blood pressure during anesthesia, were comparable with values reported in other studies.2–5,11 Thus, differences in anesthetic premedication protocols were not expected to be responsible for differences in results among studies. In addition, morphology-based values for left ventricular volumes and calculations of left ventricular SV for the present study were extremely similar to results for other studies.2–4,11 Finally, differences between left and right ventricular volumes for the present study were comparable to those of 2 other studies,11,17 which indicated that differences were mostly dependent on the technology, rather than on the operator.
An important discrepancy with results of another study11 is the difference in PCA-based SV across the valves. In the study reported here, calculated SVs were extremely similar across all 4 valves; however, SVs calculated for the atrioventricular valves were approximately 50% lower than SVs calculated for the semilunar valves in that other study.11 We cannot explain this difference. Regardless, inflow and outflow volumes should be essentially identical in a healthy heart, and noninvasive imaging is essential for evaluations of diseased hearts.6,7
Phase-contrast angiography is a noncontrast angiographic technique.7 The technique was introduced in veterinary diagnostic imaging to compare effects of anesthetic protocols on cardiac flow.11 Investigators likely use contrast media in an attempt to increase the signal-to-noise ratio, which is a concern in PCA.18 The present study involved the use of PCA without contrast medium, with a field strength of 3-T and a high temporal resolution. We compensated for loss of signal with an increase in slice thickness (8 mm, compared with 5 to 6 mm in the aforementioned study11). Calculations based on morphology in the present study used a 5-mm slice thickness (compared with a 6-mm slice thickness) and a higher temporal resolution (30 phases/cardiac cycle, compared with 20 phases/cardiac cycle) than was used in other recent studies.11,19 A reduction of slice thickness to 4 mm appears to be possible without a reduction of temporal resolution and probably would facilitate inclusion of the right ventricular outflow tract.3,4
We recognize the technical challenges encountered in the study reported here, and we share optimism concerning feasibility regarding use of cardiac MRI for canine patients. The learning curve gained during the present study appeared to be similar to that for authors of another study.12 In particular, the twisted course of the pulmonary artery caused problems during the first examination, with failure to obtain adequate data for planes in 2 dogs. This anatomic challenge did not play a role during the second examinations. Phase-contrast angiography depends on perpendicular alignment of scan plane and flow direction.7 Consequently, correct plane alignment is of fundamental importance in cardiac MRI for both morphological and flow-dependent evaluations.12
Lack of a criterion-referenced standard was an important limitation of the present study. Cardiac MRI has been used as a reference standard for comparing routine echocardiography with contrast-enhanced or 3-D echocardiography,2,3 echocardiography of the left ventricle with CT, and 3-D echocardiography of the right ventricle with CT. However, because of minor variations in results attributable to the anesthetic protocol,11 mild day-to-day variability, marked variability of individual calculations (eg, lower and upper level of agreement11), and marked variation between methods in the present study, cardiac MRI currently might not be considered a criterion-referenced standard for veterinary medicine. In addition, the study reported here included a small population of dogs of only 1 particular breed.
For the study reported here, PCA represented a simple method for rapid measurement of blood flow through the cardiac valves for quantification of SV in healthy dogs. Results for PCA appeared to be superior to those obtained for assessment of morphological sequences, independent of location and plane orientation. Calculation of SV on the basis of morphological measurements yielded more variable results with significant differences. The complex anatomy and distinct systolic shortening made calculations (especially those of right ventricular SV based on morphology) difficult and prone to errors. Results for the present study can be extrapolated to patients with cardiac disease, whereby quantification of effective SV, regurgitant fractions, and shunt volumes can be expected to be more reliable than results obtained on the basis of morphology. Cardiac MRI has primarily been used to document and evaluate complex cardiovascular anomalies. In such cases, PCA may be more useful for providing additional information regarding peripheral blood flow.20
Acknowledgments
Supported in part by the Albert Heim Foundation.
Presented in abstract form at the Conference of European Veterinary Diagnostic Imaging, Utrecht, The Netherlands, August 2014.
ABBREVIATIONS
CV | Coefficient of variation |
PCA | Phase-contrast angiography |
SV | Stroke volume |
Footnotes
Morphasol, Dr. E. Graeub AG, Bern, Switzerland.
Propofol 1%, Fresenius Kabi AG, Oberdorf, Switzerland.
IsoFlo, Abott AG, Altishofen, Switzerland.
Fresenius Kabi AG, Oberdorf, Switzerland.
Mallard, Mallard Medical, Redding, Calif.
Philips Ingenia 3T with dStream body coil solution, Philips AG, Zurich, Switzerland.
Philips AG, Zurich, Switzerland.
750 clear tape electrodes, Kendall, Anandic Medical System SA, Feuerthalen, Switzerland.
MRI monitor, GE Healthcare, Anandic Medical System SA, Feurthalen, Switzerland.
Extended MR WorkSpace, Philips AG, Zurich, Switzerland.
Cardiac Explorer, Philips AG, Zurich, Switzerland.
Q-FlowAnalysis, Philips AG, Zurich, Switzerland.
SPSS IBM, Schweiz, Zurich, Switzerland.
References
1. Gilbert SH, McConnell FJ, Holden AV, et al. The potential role of MRI in veterinary clinical cardiology. Vet J 2010; 183: 124–134.
2. Kim JH, Lee MS, Lee SY, et al. Contrast echocardiography to assess left ventricular volume and function in Beagle dogs: comparison with 3-Tesla dual source parallel cardiac magnetic resonance imaging. Vet J 2013; 198: 450–456.
3. Meyer J, Wefstaedt P, Dziallas P, et al. Assessment of left ventricular volumes by use of one-, two-, and three-dimensional echocardiography versus magnetic resonance imaging in healthy dogs. Am J Vet Res 2013; 74: 1223–1230.
4. Sieslack AK, Dziallas P, Nolte I, et al. Comparative assessment of left ventricular function variables determined via cardiac computed tomography and cardiac magnetic resonance imaging in dogs. Am J Vet Res 2013; 74: 990–998.
5. Sieslack AK, Dziallas P, Nolte I, et al. Quantification of right ventricular volume in dogs: a comparative study between three-dimensional echocardiography and computed tomography with the reference method magnetic resonance imaging. BMC Vet Res 2014; 10: 242.
6. Gatehouse PD, Keegan J, Crowe LA, et al. Applications of phase-contrast flow and velocity imaging in cardiovascular MRI. Eur Radiol 2005; 15: 2172–2184.
7. Bogaert J. Cardiac function. In: Bogaert J, Dymarkowski S, Taylor AM, et al. Clinical cardiac MRI. Heidelberg, Germany: Springer-Verlag Berlin, 2012; 109–166.
8. Varaprasathan GA, Araoz PA, Higgins CB, et al. Quantification of flow dynamics in congential heart disease: applications of velocity-encoded cine MR imaging. Radiographics 2002; 22: 895–905.
9. Didier D, Ratib O, Lerch R, et al. Detection and quantification of valvular heart disease with dynamic cardiac MR imaging. Radiographics 2000; 20: 1279–1299.
10. Kondo C, Caputo GR, Semelka R, et al. Right and left ventricular stroke volume measurements with velocity-encoded cine MR imaging: in vitro and in vivo validation. AJR Am J Roentgenol 1991; 157: 9–16.
11. Drees R, Johnson RA, Stepien RL, et al. Effects of two different anesthetic protocols on cardiac flow measured by two-dimensional phase contrast magnetic resonance imaging. Vet Radiol Ultrasound 2015; 56: 168–175.
12. Sargent J, Connolly DJ, Watts V, et al. Assessment of mitral regurgitation in dogs: comparison of results of echocardiography with magnetic resonance imaging. J Small Anim Pract 2015; 56: 641–650.
13. Odegard KC, Dinardo JA, Tsai-Goodman B, et al. Anaesthesia considerations for cardiac MRI in infants and small children. Paediatr Anaesth 2004; 14: 471–476.
14. Alfakih K, Plein S, Thiele H, et al. 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 2003; 17: 323–329.
15. Buckberg G, Hoffman JIE, Mahajan A, et al. Cardiac mechanics revisited: the relationship of cardiac architecture to ventricular function. Circulation 2008; 118: 2571–2587.
16. Burbank F, Parish D, Wexler L. Echocardiographic-like angled views of the heart by MR imaging. J Comput Assist Tomogr 1988; 12: 181–195.
17. Baumwart RD, Meurs KM, Raman SV. Magnetic resonance imaging of right ventricular morphology and function in Boxer dogs with arrhythmogenic right ventricular cardiomyopathy. J Vet Intern Med 2009; 23: 271–274.
18. Saleh RS, Lohan DG, Nael K, et al. Cardiovascular MRI at 3T. Appl Radiol 2007; 36: 10–26.
19. Drees R, Johnson RA, Stepien RL, et al. Quantitative planar and volumetric cardiac measurements using 64 MDCT and 3T MRI vs. standard 2D and M-mode echocardiography: does anesthetic protocol matter? Vet Radiol Ultrasound 2015; 56: 638–657.
20. López-Alvarez J, Dukes-McEwan J, Martin MW, et al. Balloon dilation of an imperforate cor triatriatum dexter in a Golden Retriever with concurrrent double-chambered right ventricle and subsequent evaluation by cardiac magnetic resonance imaging. J Vet Cardiol 2011; 13: 211–218.