- Email: [email protected]
Measurement of right ventricular mass in normal and dilated cardiomyopathic ventricles using cine magnetic resonance imaging
MEnfoDs
Measurement of Right Ventricular Mass in Normal and Dilated Cardiomyopathic Ventricles Using Cine Magnetic Resonance Imaging Neil E. Doherty III, MD, Naoya Fujita, MD, Gary R. Caputo, MD, and Charles B. Higgins, MD
The accurate quantification of right ventricular (RV) mass has eluded conventional imaging modalities. Accordingly, tine magnetic resonance imaging was used for quantification of RV as well as left ventricular (W) mass in 10 normal subjects and in 10 patients with dilated cardiomyopathy with an LV ejection fraction <0.40. Hearts were imaged wlth 10 mm thick short-axis slices from apex to base with a short echo delay time of 5 ms. Each slice was partitioned into 3 sections: RV free wall, ventricular septum and LV free wall, for calculation of end-diastolic and end-systolic mass and UkRV free wall ratio. RV end-diastolic mass in normal subjects was 45 f 8 g, which was similar to the values determined in previously published postmortem studies, mean 46 g (range 23 to 66). The value determined in patients with dilated cardiomyopathy was higher (50 f 11 g), but this difference was not significant. LV:RV free wall ratio in cardiomyopathy (3.6 f 1.0) was greater than in normal subjects (2.4 f 0.3), because of the greater LV free wall mass in dilated cardiomyopathy, where LV free wall end-diastolic mass was 173 f 40 g vs 107.1 f 19.9 g in normal subjects (p <0.05). RV mass measurements had 6.4 f 3.6% interobserver and 7.3 f 6.1% intraobserver variability. There were no significant differences between end-diastolic and end-systolic mass measurements. Thus, tine magnetic resonance lmaglng can reproducibly calculate RV mass. The values in normal subjects correspond to previously reported postmortem values for a population without heart dlsease. (Am J Cardiol lSS2;69:1223-1226)
From the Department of Radiology, University of California San Francisco School of Medicine, San Francisco, California. Dr. Doherty was supported by National RAearch Award HL07570-07 from the National Institutes of Health, Bethesda, Maryland. Manuscript received August 14,199l; revised manuscript received and accepted December 23, 1991. Address for reprints: Charles B. Higgins, MD, University of California, 505 Pamassus Avenue, L-308, Department of Radiology, San Francisco, California 94143-0628.
ecently, it has beenshown that in vitro right ventricular (RV) masscan be accurately measured postmortem with spin-echo magnetic resonance imaging (MRI).’ Further interest in the measurement of RV masswas stimulated by the in vivo measurement of RV mass by ultrafast computed tomography.2 BecauseMRI also is basedon tomographic imaging, but is more available than ultrafast computed tomography, the current study sought to assessthe capability of tine MRI to measure RV mass in normal subjects and in patients with dilated cardiomyopathy. Values calculated from tine MRI were compared with normal values established by postmortem studies. RV mass measurements in dilated cardiomyopathy have not been previously reported.
R
METHODS Study subThe study group included 10 normal volunteers, aged 25 to 56 years, and 10 patients with dilated cardiomyopathy, aged 34 to 73 years. The diagnosis of dilated cardiomyopathy was based on clinical information and 2-dimensional echocardiographic or scintigraphic measuredejection fraction <0.40. All patients were consideredto be in New York Heart Association Class II (n = 3) or III (n = 7). No patients had significant valvular regurgitation. The tine MRI studies were approved by the Human Research Committee at the University of California, San Francisco,and all subjects gave informed consent. Imaging technique: Cine MRI was performed with a commercially available 1ST superconducting magnet (General Electric, Milwaukee, Wisconsin). Cine MRI was recordedin the short-axis plane3with gradient refocused echoeswith a short echo delay time of 5 ms, TR of 32 ms, and flip angle 30’ in 10 mm slicesof the heart from apex to base. No gap between slices was accomplished by interdigitating slices on alternating acquisitions. Two slices were acquired on each acquisition. To encompassthe entire right ventricle, 8 to 12 tomographic levels were acquired for each study. bnage analysis: Images were analyzed independently by 2 observersoff-line on a Signa console (General Electric, Milwaukee, Wisconsin). Standardized width and level settings were used for image display to trace epicardial and endocardial borders for each slice4 To accurately define edges,signal intensity region-of-interest measurementswere obtained on each image for RV free wall, chest wall, RV inferior wall, liver, RV cavity, interventricular septum, LV free wall, LV cavity and lung. Levels were set at myccardial signal intensity f 0.15 X signal intensity of the adjacent surface. Widths
MEASUREMENT OF RIGHT VENTRICULAR MASS USING CINE MRI
1223
TABLE I Right Ventricular Mass and Free Wall Thickness in Normal Subjects and in Patients with Dilated Cardiomyopathic Ventricles
RV free wall end-diastolrc mass RVfree wall end-systolrc mass Interventricular septum enddiastolic mass Interventricular septum endsystolic mass LVfree wall end-diastolic mass LV free wall end-systolic mass LV/RV end-diastolic mass LV/RV end-systolic mass RVfree wall end-diastolic thickness RV end-diastolic volume
Normal
Drlated Cardiomyopathy
44.6 & 7.8 g 44.3 ir 7.0 g 53.8 k 12.2 g
49.7 ? 11.2 g 49.3 k 11.8 g 76.4 + 12.8 g*
52.2 _f 11.1 g
73.6 + 10.7 g*
107.1 105.4 2.4 2.4 3.8
? -t 2 f k
19.9 g 19.0 g 0.3 0.2 0.3 mm
117 * 19ml
173.2 175.3 3.6 3.5 4.5
? k k !I +
40.0 gt 39.3 gt l.O* 0.9* 0.6 mm*
128 ? 17 ml
*p
were set as the difference in signal intensity between adjacent surfaces.Images were analyzed independently by 2 observersand repeated 14 weekslater by the first observer. Myocardial volumes were calculated using Simpson’srule. Mass was calculated by multiplying this value by the density of heart muscle, 1.055 g/ml. Each slice was partitioned into 3 sections: RV free wall, inter-ventricular septum, and LV free wall for calculations of end-diastolic and end-systolic mass. Figure 1 illustrates the partitioning of a typical midventricular slice. The method for partitioning the ventricular myocardium is shown in Figure 2. The RV free wall is parti-
TABLE II Interobserver Variability of Mass Measurements in Normal Subjects and in Patients with Dilated Cardiomyopathic Hearts
RV free wall end-drastolic mass RVfree wall end-systolic mass Ventricular septum end-diastolic mass Ventricular septum end-systolrc mass LVfree wall end-diastokc mass LV free wall end-systolic mass LV:RV free wall end-diastolic mass LV:RV free wall end-systokc mass
4.1 + 2.7 5.0 + 3.2 8.1 r 5.4
6.7 k 4.6 6.4 k 3.6 8.3 + 6.1
8.0 + 6.1
7.4 k 7.1
5.3 k 4.0 4.9 k 4.5 5.2 k 5.0
6.4 + 5.1 5.4 + 4.1 8.5 k 7.9
5.5 + 4.9
9.6 + 6.7
tioned by a line continuing from the RV cavity-interventricular septal junction, continuing through to the epicardial RV free wall-interventricular junction point. Straight lines are drawn from the superior epicardialseptaljunction and inferior epicardial-septal junction to the center of the LV cavity in order to partition the interventricular septum and LV free wall. Masseswere summed for each segmentfrom apex to base. Figure 1 is photographedat level, width and window settings appropriate for the anterior free wall, where the crosshair is placed. The lower free wall is not fully visualized in Figure 1 becauseit has a lower signal intensity,
FlGURE2.Mi~ofnonnal-with--
THE AMERICAN JOURNAL OF CARDIOLOGY VOLUME 69
Dilated Cardiomyopathy (%I
Values are mea” 2 standard dewatlon. LV = left ventricular: RV = right ventricular.
&at8ndepicardidedgestracedinsihoMtetodepktparti-
1224
Normal (%I
MAY 1, 1992
and was therefore measuredat a lower window level but without having to extrapolate from adjacent portions of the RV free wall. Statistiis: Values are expressedas mean f standard deviation. Differences between observers and observations in massand LV:RV free wall ratio were compared with a Student’s 2-tailed t test with differences considered significant at p <0.05. The interobserver and intraobserver percent variability for each mass measurement was determined as the quotient of the absolute value of the difference betweenthe 2 measurementsand the mean of the 2 measurements.5,6Linear regression analysis was performed to calculate interobserver and intraobserver reproducibility with standard error of the estimate (SEE) for each mass measurement and LV/ RV mass ratio. RESULTS Table I lists mean values for RV free wall, interventricular septum and LV free wall end-diastolic and end-
Interobserver Normals-RVESM 70 grams
TABLE III lntraobserver Variability of Mass Measurements in Normal Subjects and in Patients with Dilated Cardiomyopathic Hearts
RV free wall end-diastolic mass
3.8 k 1.9
RV free wall end-systolic mass
4.8 2 4.3
7.3 + 6.1 5.4 -t 6.5
Ventricular septum end-diastok mass
7.6 2 6.8
7.6 + 5.7
Ventricular septum end-systolic mass
5.1 + 4.3
5.6 2 3.6
LV free wall end-diastolic mass
3.3 + 2.2
LVfree wall end-systolic mass
3.7 + 2.7
5.1 2 3.9 4.9 + 3.6
LV:RV free wall end-dlastollc mass
4.6 + 3.8
7.3 k 6.9
LV:RV free wall end-systolic mass
4.7 2 5.4
7.6 + 4.7
Values are mea” 2 standard dewatton. LV = left ventricular; RV = right ventricular.
Interobserver Normals-RVEDM
Reproduclblllty 70 grams
1
Reproduclblllty
1
60
w1 St Obs
1st Ohs
w-
50
40
40-
.
30!/
A
Dllated Cardiomyopathy (%)
Normal (%)
1
30
40
50 2nd
30
60 grams
Observer
B
40
50 2nd
Observer
60 grams
lnterobserver Reproducibility Dilated Cardiomyopathy-RVESM 801
grams .1 70 60-
1st +jgObs . 5.60 + 0.89x
SEE = 3.9
20’
C
20
*
’
30
*
’
* ’
40
50
2nd
’
Observer
’ 60
*
’
*
70
’ 60
grams
FIGURE 3. Plots show -er~f8rtightventricubrd-systdk(RVEsM)art4l8Itd~m888 (RVEDM) far normal subjects (A, 6) md patids with dilated cmdiomyopathy ford-.
(C, LB).ComeWon
coefkhb
(R) were X.93
MEASUREMENT OF RIGHT VENTRICULAR MASS USING CINE MRI 1225
systolic massmeasurements.Table I also lists mean values for RV free wall thickness at end-diastole. There were no significant differencesin massmeasurementsat end-diastole compared with end-systole.Mean value for RV free wall measurementsat end-diastole in normal volunteers was 45 f 8 g. LV:RV free wall massratio at end-diastole was 2.4 f 0.3. RV free wall thickness was
opathy was 76 f 13 g compared with 54 f 12 g in normal hearts (p
lntraobaerver Reproduclbllty Normals-RVEDM
1
70 grams 1 60 --I 1st Oba
1st Obs
4040
- 3.96 + 1.11x R I 0.99 SEE E 0.7 g
Y=
3oI/ 30
A
30
40 2nd
50 Observation
60 grams
B
grams
30
9L
40 2nd
1 50
1 60 grams
Observation
lntraobaerver Reproduclbillty Dilated Cardiomyopathy-RVEDM
60WI-
grmll. so
70 -
70 i
1st Oba
/
60 -
- 2.3 + 1.07x A I 0.93 SEE I 4.6
40
-20
SO
40
C
2nd
FlGuRE4.Pbt8dnwinb-~forrighi (RVEDM) for nomd sdjects faran-.
1226
60
60
0b..rv6t10”
(A, B) and paiienb
70
60 own.
with dated
THE AMERICAN JOURNAL OF CARDIOLOGY VOLUME 69
3oK’ 30 D
’
40
. 2nd
’
60
.
Obaervatlon
’
60
.
’
70
*
grema
vmbidar emd-systdic (RVESM) and enddiastolic mass cmtiomyopathy (C, DJ, ComeWon coef&kb (R) were >0.93
MAY 1, 1992
’
60
tine MRI with short echo delay times can be used to were not usually seen,although their position could be measureRV free wall masswith a high degreeof repro- ascertainedby differencesin flow betweenright atrium, ducibility. Measurements of mass by tine MRI agree right ventricle and pulmonary artery. However, this inwith necropsy measurementsof RV free wall massand formation itself was not usually necessaryto determine LV:RV free wall ratios7 Mean value for RV free wall where the RV muscle stopped. The inferior border of measurementsin normal volunteers at end-diastole was the right ventricle could sometimes be clearly distin45 f 8 g. Mean value for LV:RV free wall mass ratio guished from the liver, but usually the difference in sigwas 2.4 f 0.3. These values are comparable to necrop- nal characteristics between inferior RV myocardium sy measurementsof RV free wall mass (mean 46 g, and liver was small and it was then necessaryto extrag range 23-68) and LV:RV free wall ratio 1.9 in subjects olate from RV free wall to the inferior wall adjacent to aged <76 years and without evidenceof heart disease.7 the septum,which was well delineated from other strucIn contrast, ultrafast computed tomography reported tures. The conus section of the right ventricle was readily defined and this portion was included in the measureslightly higher values for mean RV free wall mass,55 f ments. 3 g in 7 young healthy male volunteers.* The accuracy of nontomographic techniques for the Cine MRI showed slightly higher values for RV mass in dilated compared with normal ventricles but estimation of RV mass is limited. The electrocardiothese differences were not significant. However, RV gram is very insensitive for RV hypertrophy and is a free wall thickness was significantly greater in dilated poor predictor of absolute RV mass.lOJ1Surface echocardiomyopathic hearts than in hearts of normal sub- cardiography is limited by poor resolution due to inadejects. LV:RV free wall ratio in dilated cardiomyopathic quate acoustical windows and therefore gives poor estiRadionuclide ventriculography hearts at end-diastole was significantly higher than in mates of RV mass.12-15 Biplane normal hearts due to a higher LV mass.The increasein cannot assessRV wall thickness or mass.16-18 LV mass in dilated cardiomyopathy has been shown angiography cannot determine RV mass or wall thickpreviously with MRI.6 Previous studies have shown a nessdue to overlap with LV mass.19,20 RV massmeasurementsmay be useful for monitorwide variability of RV massand RV free wall thickness in patients with left-sided cardiac discase.7,8An echo- ing the effects of therapy in patients with right-sided cardiographic evaluation of the RV wall thickness with cardiac diseases*lor pulmonary hypertension.** It has proximate confirmation by postmortem measurements also been speculated that RV mass may be an imporindicated both increased and normal wall thickness in tant prognostic measurementin somediseases,*particupatients with LV cardiomyopathy.* The slight but not larly chronic obstructive pulmonary disease.A full assignificant increase in RV free wall mass in patients sessmentof RV function, including volumes, ejection with class II and III functional status in the current fraction and mass,is now possiblewith either ultrafast study is consistent with this variable occurrence of RV computed tomography23or MRI.24,25 The responseof the right ventricle to diseasesaffecthypertrophy in patients with dilated cardiomyopathy. The reproducibility data of the current study suggest ing the left ventricle can be studied with these tomothat tine MRI is a highly reproducible technique for graphic techniques. The current study showed slightly measurementof RV mass,with interobserver variabili- greater RV mass in dilated cardiomyopathy. The meaties of 4 f 3 to 7 f 5% and intraobserver variabilities of surement of RV mass in hypertensive LV hypertrophy 4 f 2 to 7 f 6%. Furthermore, the variability of RV and hypertrophic cardiomyopathy should be studied to massmeasurementswas not significantly different from see how these conditions influence the right ventricle. the variability of LV massmeasurements,which were in the range of interobserver variability reported by Semelka et a15*6for normal and dilated cardiomyopathy. The most important technical refinement to enable REFERENCES measurementof RV massby tine MRI was shortening 1. Mackey E, Sandier MP, Campbell RM, Graham TP, Atkinson JB, Pricer R, Moreau GA. Right ventricular myocardial mass quantification with magnetic of the echo delay time to 5 ms. Initial attempts at resonance imaging. Am J Cardiol 1990;65:529-553. quantification of RV massusing an echo delay time of 2. Hadjuczok ZD. Weiss RM, Stanford W, Marcus ML. Determination of right 17 ms gave unsatisfactory edge discrimination and ex- ventricular mass in humans and dogs with ultrafast computed tomography. Circu1990,82:202-212. cess slow-flow artifact for accurate delineation of RV lation 3. Dinsmore RE, Wismer GL, Levine RA, Okada RD, Brady TJ. Magnetic borders. Earlier reports using spin-echo images with resonance imaging of the heart: positioning and gradient angle selection for echo delay time of 34 ms measuredRV free wall thick- optimal imaging planes. AJR 1984;143:1135-1142. J, Caputo GR, Kondo C, Higgins CB. Cine MR imaging of valvular nessbut did not report RV free wall massand were only 4.heartSuzuki disease: d&play and imaging parameters affect the size of the signal void able to give a clear image of 68% of the RV free wa11.9 caused by valvular regurgitation. AJR 1990;155:723-727. Semelka RC, Tomei R, Wagner S, Mayo J, Kondo C, Suzuki JI, Caputo G, There were a few areas of the heart where measure- 5.Higgins CB. Normal left ventricular dimensions and function: interstudy repro ment of RV mass by tine MRI was difficult. The RV ducibility of measurements with tine MR imaging. Radio&y 1990;174:763-768. apex usually, but not always, extended to the level of 6. Semelka RC, Tomci R, Wagner S, Mayo J, Caputo G, O’Sullivan M, Parmley WW, Chatterjee K, Wolfe C, Higgins CB. Interstudy reproducibility of dimenthe LV apex. When the 2 ventricles shared the apex and sional and functional measurements between tine magnetic resonance studies in no cavity was present, assignment of mass to right or the morphologically abnormal left ventricle. Am Heart J 1990;119:1367-1373. left ventricle was inferred by comparison to adjacent 7. Fulton RM, Hutchinson EC, Jones AM. Ventricular weight in cardiac hyperBr Heart J 1952;14:413-420. levels.Other areasof difficulty were identifying the bor- trophy. 8. Prokash R. Determination of right ventricular wall thickness in systole and ders of the pulmonic and tricuspid valves. The valves diastole. Echocardiographic and nccropsy correlation in 32 patients. Er Heart J MEASUREMENT OF RIGHT VENTRICULAR MASS USING CINE MRI
1227
1978;40:1257-1261. 9. Suzuki JI, Sakamoto T, Taken&
K, Kawakubo K, Amano K, Takahashi H, Hasegawa I, Shiota T, Hada Y, Sugimoto T, Nishikawa JI. Assessment of the thickness of the right ventricular free wall by magnetic resonance imaging with patients with hypertrophic cardiomyopathy. Br Heart J 1988;60:440-445. 10. Roman CT, Walsh TJ, Massie E. Right ventricular hypcrtrophy. Am J Cardiol 1961;7:481-487. Il. Milnor WR. Electrocardiogram and vectorcardiogram in right ventricular hypertrophy and right bundle branch block. Circulation 1957;16:348-367. 12. Watanabe T, Katsume H, Matsukobo HJ, Furukawa K, Ijichi H. Estimation of right ventricular volume with two-dimensional echocardiography. Am J Cardial 1982;49:1946-1953. 13. Starling MR. Crawford MH, Sorensen SC, O’Rourke RA. A new twodimensional echocardiographic technique for evaluating right ventricular size and performance in patients with obstructive lung disease. Circulafion 198266:
612-620. 14. Levine RA, Gibson TC, Aretz T, Gillam LD, Guyer DE, King ME, Weyman AE. Echocardiographic measurements of right ventricular volume. Circulation 1984;69:497-505. IS. Baker BJ, Scovil JA, Kane JJ, Murphy ML. Echocardiographic detection of right ventricular hypertrophy. Am Heart J 1983;105:611-614. 16. Cohen HA, Baird MG, Rouleau JR, Fuhrmann CF. Bailey IK, Summer WR, Strauss HW, Pitt B. Thallium-201 myocardial imaging in patients with pulmonary hypertension. Circulation 1976;54:790-795.
1228
THE AMERICAN
JOURNAL
OF CARDIOLOGY
VOLUME
69
17. Khaja F, Alam M, Goldstein S, Anahe DT, Marks DS. Diagnostic value of visualization of the right ventricle using thallium-201 myocardial imaging. Circulation 1979;59:182-188. 18. Ohusuzu F, Handa S, Kondo M, Yamazaki H, Tzugu T, Kubo A, Takagi Y, Nakamura Y. Thallium-201 myocardial imaging to evaluate right ventricular overloading. Circulation 1980;61:620-625. 19. Gentzler RD. Briselli MF, Gault JH. Angiographic estimation of right ventricular volume in man. Circulation 1974;50:324-330. 20. Shimazaki Y, Kamashima Y, Mori T, Beppi S. Angiographic volume estimation of right ventricle. Chest 1980;77:390-395. 21. Markiewicz W, Sectem U, Higgins CB. Evaluation of the right ventricle by magnetic resonance imaging. Am Hearf J 1987;113:8-15. 22. Bouchard A, Higgins CB, Byrd BF 111, Amparo EP, Osaki L, Axelrod R. Magnetic resonance imaging in pulmonary artery hypertension. Am J Cardioi 1985:56:938-942. 23. Reiter SJ, Rumberger JA, Feiring AJ, Stanford W, Marcus ML. Precision of right and left ventricular volume measurements by tine computed tomography, Circulation 1986;74:890-900. 24. Markiewicz W, Sechtem U, Kirby R, Derugin N, Caputo GC, Higgins CB. Measurement of ventricular volumes in the dog by nuclear magnetic resonance imaging. J Am CON Cardiol 1987;10:170-177. 25. Swhtem U, Pflugfelder PW, Gould RG, Cassidy MW, Higgins CB. Measurement of right and left ventricular volumes in healthy individuals with tine MR imaging. Radiology 1987;163:697-702.
MAY 1. 1992
Measurement of Right Ventricular Mass in Normal and Dilated Cardiomyopathic Ventricles Using Cine Magnetic Resonance Imaging Neil E. Doherty III, MD, Naoya Fujita, MD, Gary R. Caputo, MD, and Charles B. Higgins, MD
The accurate quantification of right ventricular (RV) mass has eluded conventional imaging modalities. Accordingly, tine magnetic resonance imaging was used for quantification of RV as well as left ventricular (W) mass in 10 normal subjects and in 10 patients with dilated cardiomyopathy with an LV ejection fraction <0.40. Hearts were imaged wlth 10 mm thick short-axis slices from apex to base with a short echo delay time of 5 ms. Each slice was partitioned into 3 sections: RV free wall, ventricular septum and LV free wall, for calculation of end-diastolic and end-systolic mass and UkRV free wall ratio. RV end-diastolic mass in normal subjects was 45 f 8 g, which was similar to the values determined in previously published postmortem studies, mean 46 g (range 23 to 66). The value determined in patients with dilated cardiomyopathy was higher (50 f 11 g), but this difference was not significant. LV:RV free wall ratio in cardiomyopathy (3.6 f 1.0) was greater than in normal subjects (2.4 f 0.3), because of the greater LV free wall mass in dilated cardiomyopathy, where LV free wall end-diastolic mass was 173 f 40 g vs 107.1 f 19.9 g in normal subjects (p <0.05). RV mass measurements had 6.4 f 3.6% interobserver and 7.3 f 6.1% intraobserver variability. There were no significant differences between end-diastolic and end-systolic mass measurements. Thus, tine magnetic resonance lmaglng can reproducibly calculate RV mass. The values in normal subjects correspond to previously reported postmortem values for a population without heart dlsease. (Am J Cardiol lSS2;69:1223-1226)
From the Department of Radiology, University of California San Francisco School of Medicine, San Francisco, California. Dr. Doherty was supported by National RAearch Award HL07570-07 from the National Institutes of Health, Bethesda, Maryland. Manuscript received August 14,199l; revised manuscript received and accepted December 23, 1991. Address for reprints: Charles B. Higgins, MD, University of California, 505 Pamassus Avenue, L-308, Department of Radiology, San Francisco, California 94143-0628.
ecently, it has beenshown that in vitro right ventricular (RV) masscan be accurately measured postmortem with spin-echo magnetic resonance imaging (MRI).’ Further interest in the measurement of RV masswas stimulated by the in vivo measurement of RV mass by ultrafast computed tomography.2 BecauseMRI also is basedon tomographic imaging, but is more available than ultrafast computed tomography, the current study sought to assessthe capability of tine MRI to measure RV mass in normal subjects and in patients with dilated cardiomyopathy. Values calculated from tine MRI were compared with normal values established by postmortem studies. RV mass measurements in dilated cardiomyopathy have not been previously reported.
R
METHODS Study subThe study group included 10 normal volunteers, aged 25 to 56 years, and 10 patients with dilated cardiomyopathy, aged 34 to 73 years. The diagnosis of dilated cardiomyopathy was based on clinical information and 2-dimensional echocardiographic or scintigraphic measuredejection fraction <0.40. All patients were consideredto be in New York Heart Association Class II (n = 3) or III (n = 7). No patients had significant valvular regurgitation. The tine MRI studies were approved by the Human Research Committee at the University of California, San Francisco,and all subjects gave informed consent. Imaging technique: Cine MRI was performed with a commercially available 1ST superconducting magnet (General Electric, Milwaukee, Wisconsin). Cine MRI was recordedin the short-axis plane3with gradient refocused echoeswith a short echo delay time of 5 ms, TR of 32 ms, and flip angle 30’ in 10 mm slicesof the heart from apex to base. No gap between slices was accomplished by interdigitating slices on alternating acquisitions. Two slices were acquired on each acquisition. To encompassthe entire right ventricle, 8 to 12 tomographic levels were acquired for each study. bnage analysis: Images were analyzed independently by 2 observersoff-line on a Signa console (General Electric, Milwaukee, Wisconsin). Standardized width and level settings were used for image display to trace epicardial and endocardial borders for each slice4 To accurately define edges,signal intensity region-of-interest measurementswere obtained on each image for RV free wall, chest wall, RV inferior wall, liver, RV cavity, interventricular septum, LV free wall, LV cavity and lung. Levels were set at myccardial signal intensity f 0.15 X signal intensity of the adjacent surface. Widths
MEASUREMENT OF RIGHT VENTRICULAR MASS USING CINE MRI
1223
TABLE I Right Ventricular Mass and Free Wall Thickness in Normal Subjects and in Patients with Dilated Cardiomyopathic Ventricles
RV free wall end-diastolrc mass RVfree wall end-systolrc mass Interventricular septum enddiastolic mass Interventricular septum endsystolic mass LVfree wall end-diastolic mass LV free wall end-systolic mass LV/RV end-diastolic mass LV/RV end-systolic mass RVfree wall end-diastolic thickness RV end-diastolic volume
Normal
Drlated Cardiomyopathy
44.6 & 7.8 g 44.3 ir 7.0 g 53.8 k 12.2 g
49.7 ? 11.2 g 49.3 k 11.8 g 76.4 + 12.8 g*
52.2 _f 11.1 g
73.6 + 10.7 g*
107.1 105.4 2.4 2.4 3.8
? -t 2 f k
19.9 g 19.0 g 0.3 0.2 0.3 mm
117 * 19ml
173.2 175.3 3.6 3.5 4.5
? k k !I +
40.0 gt 39.3 gt l.O* 0.9* 0.6 mm*
128 ? 17 ml
*p
were set as the difference in signal intensity between adjacent surfaces.Images were analyzed independently by 2 observersand repeated 14 weekslater by the first observer. Myocardial volumes were calculated using Simpson’srule. Mass was calculated by multiplying this value by the density of heart muscle, 1.055 g/ml. Each slice was partitioned into 3 sections: RV free wall, inter-ventricular septum, and LV free wall for calculations of end-diastolic and end-systolic mass. Figure 1 illustrates the partitioning of a typical midventricular slice. The method for partitioning the ventricular myocardium is shown in Figure 2. The RV free wall is parti-
TABLE II Interobserver Variability of Mass Measurements in Normal Subjects and in Patients with Dilated Cardiomyopathic Hearts
RV free wall end-drastolic mass RVfree wall end-systolic mass Ventricular septum end-diastolic mass Ventricular septum end-systolrc mass LVfree wall end-diastokc mass LV free wall end-systolic mass LV:RV free wall end-diastolic mass LV:RV free wall end-systokc mass
4.1 + 2.7 5.0 + 3.2 8.1 r 5.4
6.7 k 4.6 6.4 k 3.6 8.3 + 6.1
8.0 + 6.1
7.4 k 7.1
5.3 k 4.0 4.9 k 4.5 5.2 k 5.0
6.4 + 5.1 5.4 + 4.1 8.5 k 7.9
5.5 + 4.9
9.6 + 6.7
tioned by a line continuing from the RV cavity-interventricular septal junction, continuing through to the epicardial RV free wall-interventricular junction point. Straight lines are drawn from the superior epicardialseptaljunction and inferior epicardial-septal junction to the center of the LV cavity in order to partition the interventricular septum and LV free wall. Masseswere summed for each segmentfrom apex to base. Figure 1 is photographedat level, width and window settings appropriate for the anterior free wall, where the crosshair is placed. The lower free wall is not fully visualized in Figure 1 becauseit has a lower signal intensity,
FlGURE2.Mi~ofnonnal-with--
THE AMERICAN JOURNAL OF CARDIOLOGY VOLUME 69
Dilated Cardiomyopathy (%I
Values are mea” 2 standard dewatlon. LV = left ventricular: RV = right ventricular.
&at8ndepicardidedgestracedinsihoMtetodepktparti-
1224
Normal (%I
MAY 1, 1992
and was therefore measuredat a lower window level but without having to extrapolate from adjacent portions of the RV free wall. Statistiis: Values are expressedas mean f standard deviation. Differences between observers and observations in massand LV:RV free wall ratio were compared with a Student’s 2-tailed t test with differences considered significant at p <0.05. The interobserver and intraobserver percent variability for each mass measurement was determined as the quotient of the absolute value of the difference betweenthe 2 measurementsand the mean of the 2 measurements.5,6Linear regression analysis was performed to calculate interobserver and intraobserver reproducibility with standard error of the estimate (SEE) for each mass measurement and LV/ RV mass ratio. RESULTS Table I lists mean values for RV free wall, interventricular septum and LV free wall end-diastolic and end-
Interobserver Normals-RVESM 70 grams
TABLE III lntraobserver Variability of Mass Measurements in Normal Subjects and in Patients with Dilated Cardiomyopathic Hearts
RV free wall end-diastolic mass
3.8 k 1.9
RV free wall end-systolic mass
4.8 2 4.3
7.3 + 6.1 5.4 -t 6.5
Ventricular septum end-diastok mass
7.6 2 6.8
7.6 + 5.7
Ventricular septum end-systolic mass
5.1 + 4.3
5.6 2 3.6
LV free wall end-diastolic mass
3.3 + 2.2
LVfree wall end-systolic mass
3.7 + 2.7
5.1 2 3.9 4.9 + 3.6
LV:RV free wall end-dlastollc mass
4.6 + 3.8
7.3 k 6.9
LV:RV free wall end-systolic mass
4.7 2 5.4
7.6 + 4.7
Values are mea” 2 standard dewatton. LV = left ventricular; RV = right ventricular.
Interobserver Normals-RVEDM
Reproduclblllty 70 grams
1
Reproduclblllty
1
60
w1 St Obs
1st Ohs
w-
50
40
40-
.
30!/
A
Dllated Cardiomyopathy (%)
Normal (%)
1
30
40
50 2nd
30
60 grams
Observer
B
40
50 2nd
Observer
60 grams
lnterobserver Reproducibility Dilated Cardiomyopathy-RVESM 801
grams .1 70 60-
1st +jgObs . 5.60 + 0.89x
SEE = 3.9
20’
C
20
*
’
30
*
’
* ’
40
50
2nd
’
Observer
’ 60
*
’
*
70
’ 60
grams
FIGURE 3. Plots show -er~f8rtightventricubrd-systdk(RVEsM)art4l8Itd~m888 (RVEDM) far normal subjects (A, 6) md patids with dilated cmdiomyopathy ford-.
(C, LB).ComeWon
coefkhb
(R) were X.93
MEASUREMENT OF RIGHT VENTRICULAR MASS USING CINE MRI 1225
systolic massmeasurements.Table I also lists mean values for RV free wall thickness at end-diastole. There were no significant differencesin massmeasurementsat end-diastole compared with end-systole.Mean value for RV free wall measurementsat end-diastole in normal volunteers was 45 f 8 g. LV:RV free wall massratio at end-diastole was 2.4 f 0.3. RV free wall thickness was
opathy was 76 f 13 g compared with 54 f 12 g in normal hearts (p
lntraobaerver Reproduclbllty Normals-RVEDM
1
70 grams 1 60 --I 1st Oba
1st Obs
4040
- 3.96 + 1.11x R I 0.99 SEE E 0.7 g
Y=
3oI/ 30
A
30
40 2nd
50 Observation
60 grams
B
grams
30
9L
40 2nd
1 50
1 60 grams
Observation
lntraobaerver Reproduclbillty Dilated Cardiomyopathy-RVEDM
60WI-
grmll. so
70 -
70 i
1st Oba
/
60 -
- 2.3 + 1.07x A I 0.93 SEE I 4.6
40
-20
SO
40
C
2nd
FlGuRE4.Pbt8dnwinb-~forrighi (RVEDM) for nomd sdjects faran-.
1226
60
60
0b..rv6t10”
(A, B) and paiienb
70
60 own.
with dated
THE AMERICAN JOURNAL OF CARDIOLOGY VOLUME 69
3oK’ 30 D
’
40
. 2nd
’
60
.
Obaervatlon
’
60
.
’
70
*
grema
vmbidar emd-systdic (RVESM) and enddiastolic mass cmtiomyopathy (C, DJ, ComeWon coef&kb (R) were >0.93
MAY 1, 1992
’
60
tine MRI with short echo delay times can be used to were not usually seen,although their position could be measureRV free wall masswith a high degreeof repro- ascertainedby differencesin flow betweenright atrium, ducibility. Measurements of mass by tine MRI agree right ventricle and pulmonary artery. However, this inwith necropsy measurementsof RV free wall massand formation itself was not usually necessaryto determine LV:RV free wall ratios7 Mean value for RV free wall where the RV muscle stopped. The inferior border of measurementsin normal volunteers at end-diastole was the right ventricle could sometimes be clearly distin45 f 8 g. Mean value for LV:RV free wall mass ratio guished from the liver, but usually the difference in sigwas 2.4 f 0.3. These values are comparable to necrop- nal characteristics between inferior RV myocardium sy measurementsof RV free wall mass (mean 46 g, and liver was small and it was then necessaryto extrag range 23-68) and LV:RV free wall ratio 1.9 in subjects olate from RV free wall to the inferior wall adjacent to aged <76 years and without evidenceof heart disease.7 the septum,which was well delineated from other strucIn contrast, ultrafast computed tomography reported tures. The conus section of the right ventricle was readily defined and this portion was included in the measureslightly higher values for mean RV free wall mass,55 f ments. 3 g in 7 young healthy male volunteers.* The accuracy of nontomographic techniques for the Cine MRI showed slightly higher values for RV mass in dilated compared with normal ventricles but estimation of RV mass is limited. The electrocardiothese differences were not significant. However, RV gram is very insensitive for RV hypertrophy and is a free wall thickness was significantly greater in dilated poor predictor of absolute RV mass.lOJ1Surface echocardiomyopathic hearts than in hearts of normal sub- cardiography is limited by poor resolution due to inadejects. LV:RV free wall ratio in dilated cardiomyopathic quate acoustical windows and therefore gives poor estiRadionuclide ventriculography hearts at end-diastole was significantly higher than in mates of RV mass.12-15 Biplane normal hearts due to a higher LV mass.The increasein cannot assessRV wall thickness or mass.16-18 LV mass in dilated cardiomyopathy has been shown angiography cannot determine RV mass or wall thickpreviously with MRI.6 Previous studies have shown a nessdue to overlap with LV mass.19,20 RV massmeasurementsmay be useful for monitorwide variability of RV massand RV free wall thickness in patients with left-sided cardiac discase.7,8An echo- ing the effects of therapy in patients with right-sided cardiographic evaluation of the RV wall thickness with cardiac diseases*lor pulmonary hypertension.** It has proximate confirmation by postmortem measurements also been speculated that RV mass may be an imporindicated both increased and normal wall thickness in tant prognostic measurementin somediseases,*particupatients with LV cardiomyopathy.* The slight but not larly chronic obstructive pulmonary disease.A full assignificant increase in RV free wall mass in patients sessmentof RV function, including volumes, ejection with class II and III functional status in the current fraction and mass,is now possiblewith either ultrafast study is consistent with this variable occurrence of RV computed tomography23or MRI.24,25 The responseof the right ventricle to diseasesaffecthypertrophy in patients with dilated cardiomyopathy. The reproducibility data of the current study suggest ing the left ventricle can be studied with these tomothat tine MRI is a highly reproducible technique for graphic techniques. The current study showed slightly measurementof RV mass,with interobserver variabili- greater RV mass in dilated cardiomyopathy. The meaties of 4 f 3 to 7 f 5% and intraobserver variabilities of surement of RV mass in hypertensive LV hypertrophy 4 f 2 to 7 f 6%. Furthermore, the variability of RV and hypertrophic cardiomyopathy should be studied to massmeasurementswas not significantly different from see how these conditions influence the right ventricle. the variability of LV massmeasurements,which were in the range of interobserver variability reported by Semelka et a15*6for normal and dilated cardiomyopathy. The most important technical refinement to enable REFERENCES measurementof RV massby tine MRI was shortening 1. Mackey E, Sandier MP, Campbell RM, Graham TP, Atkinson JB, Pricer R, Moreau GA. Right ventricular myocardial mass quantification with magnetic of the echo delay time to 5 ms. Initial attempts at resonance imaging. Am J Cardiol 1990;65:529-553. quantification of RV massusing an echo delay time of 2. Hadjuczok ZD. Weiss RM, Stanford W, Marcus ML. Determination of right 17 ms gave unsatisfactory edge discrimination and ex- ventricular mass in humans and dogs with ultrafast computed tomography. Circu1990,82:202-212. cess slow-flow artifact for accurate delineation of RV lation 3. Dinsmore RE, Wismer GL, Levine RA, Okada RD, Brady TJ. Magnetic borders. Earlier reports using spin-echo images with resonance imaging of the heart: positioning and gradient angle selection for echo delay time of 34 ms measuredRV free wall thick- optimal imaging planes. AJR 1984;143:1135-1142. J, Caputo GR, Kondo C, Higgins CB. Cine MR imaging of valvular nessbut did not report RV free wall massand were only 4.heartSuzuki disease: d&play and imaging parameters affect the size of the signal void able to give a clear image of 68% of the RV free wa11.9 caused by valvular regurgitation. AJR 1990;155:723-727. Semelka RC, Tomei R, Wagner S, Mayo J, Kondo C, Suzuki JI, Caputo G, There were a few areas of the heart where measure- 5.Higgins CB. Normal left ventricular dimensions and function: interstudy repro ment of RV mass by tine MRI was difficult. The RV ducibility of measurements with tine MR imaging. Radio&y 1990;174:763-768. apex usually, but not always, extended to the level of 6. Semelka RC, Tomci R, Wagner S, Mayo J, Caputo G, O’Sullivan M, Parmley WW, Chatterjee K, Wolfe C, Higgins CB. Interstudy reproducibility of dimenthe LV apex. When the 2 ventricles shared the apex and sional and functional measurements between tine magnetic resonance studies in no cavity was present, assignment of mass to right or the morphologically abnormal left ventricle. Am Heart J 1990;119:1367-1373. left ventricle was inferred by comparison to adjacent 7. Fulton RM, Hutchinson EC, Jones AM. Ventricular weight in cardiac hyperBr Heart J 1952;14:413-420. levels.Other areasof difficulty were identifying the bor- trophy. 8. Prokash R. Determination of right ventricular wall thickness in systole and ders of the pulmonic and tricuspid valves. The valves diastole. Echocardiographic and nccropsy correlation in 32 patients. Er Heart J MEASUREMENT OF RIGHT VENTRICULAR MASS USING CINE MRI
1227
1978;40:1257-1261. 9. Suzuki JI, Sakamoto T, Taken&
K, Kawakubo K, Amano K, Takahashi H, Hasegawa I, Shiota T, Hada Y, Sugimoto T, Nishikawa JI. Assessment of the thickness of the right ventricular free wall by magnetic resonance imaging with patients with hypertrophic cardiomyopathy. Br Heart J 1988;60:440-445. 10. Roman CT, Walsh TJ, Massie E. Right ventricular hypcrtrophy. Am J Cardiol 1961;7:481-487. Il. Milnor WR. Electrocardiogram and vectorcardiogram in right ventricular hypertrophy and right bundle branch block. Circulation 1957;16:348-367. 12. Watanabe T, Katsume H, Matsukobo HJ, Furukawa K, Ijichi H. Estimation of right ventricular volume with two-dimensional echocardiography. Am J Cardial 1982;49:1946-1953. 13. Starling MR. Crawford MH, Sorensen SC, O’Rourke RA. A new twodimensional echocardiographic technique for evaluating right ventricular size and performance in patients with obstructive lung disease. Circulafion 198266:
612-620. 14. Levine RA, Gibson TC, Aretz T, Gillam LD, Guyer DE, King ME, Weyman AE. Echocardiographic measurements of right ventricular volume. Circulation 1984;69:497-505. IS. Baker BJ, Scovil JA, Kane JJ, Murphy ML. Echocardiographic detection of right ventricular hypertrophy. Am Heart J 1983;105:611-614. 16. Cohen HA, Baird MG, Rouleau JR, Fuhrmann CF. Bailey IK, Summer WR, Strauss HW, Pitt B. Thallium-201 myocardial imaging in patients with pulmonary hypertension. Circulation 1976;54:790-795.
1228
THE AMERICAN
JOURNAL
OF CARDIOLOGY
VOLUME
69
17. Khaja F, Alam M, Goldstein S, Anahe DT, Marks DS. Diagnostic value of visualization of the right ventricle using thallium-201 myocardial imaging. Circulation 1979;59:182-188. 18. Ohusuzu F, Handa S, Kondo M, Yamazaki H, Tzugu T, Kubo A, Takagi Y, Nakamura Y. Thallium-201 myocardial imaging to evaluate right ventricular overloading. Circulation 1980;61:620-625. 19. Gentzler RD. Briselli MF, Gault JH. Angiographic estimation of right ventricular volume in man. Circulation 1974;50:324-330. 20. Shimazaki Y, Kamashima Y, Mori T, Beppi S. Angiographic volume estimation of right ventricle. Chest 1980;77:390-395. 21. Markiewicz W, Sectem U, Higgins CB. Evaluation of the right ventricle by magnetic resonance imaging. Am Hearf J 1987;113:8-15. 22. Bouchard A, Higgins CB, Byrd BF 111, Amparo EP, Osaki L, Axelrod R. Magnetic resonance imaging in pulmonary artery hypertension. Am J Cardioi 1985:56:938-942. 23. Reiter SJ, Rumberger JA, Feiring AJ, Stanford W, Marcus ML. Precision of right and left ventricular volume measurements by tine computed tomography, Circulation 1986;74:890-900. 24. Markiewicz W, Sechtem U, Kirby R, Derugin N, Caputo GC, Higgins CB. Measurement of ventricular volumes in the dog by nuclear magnetic resonance imaging. J Am CON Cardiol 1987;10:170-177. 25. Swhtem U, Pflugfelder PW, Gould RG, Cassidy MW, Higgins CB. Measurement of right and left ventricular volumes in healthy individuals with tine MR imaging. Radiology 1987;163:697-702.
MAY 1. 1992