diagnosis and assessment of mitral and aortic valve disease by cine-flow magnetic resonance imaging

MAGNETICRESONANCEINMEDICINE 12,181-197 (1989)

Diagnosis and Assessment of Mitral and Aortic Valve Disease by Cine-Flow Magnetic Resonance Imaging

LESLIE MITCHELL, * JEREMY P. R. JENKINS, * YVONNE WATSON, * DEREK J. ROWLANDS,? AND IAN ISHERWOOD*

Departments of* Diagnostic Radiology and TCardioIoa, University of Manchester, Oxford Road, Manchester, England

Received July 27, 1988; revised January 12, 1989

Seventy-six aortic and mitral valves, in 44 patients and 5 normal volunteers, were studied by Cine-Flow MRI (on a0.26-T superconducting magnet system), utilizingcom- pound oblique imaging planes and a Field Echo Even Rephasing sequence. All patients had had cardiac catheterization and echocardiography. All patients with valvular stenosis and aortic sclerosis ( n = 45) showed complete signal loss distal to the respective valve. Length of signal loss distal to the aortic valve in those in whom it was measured ( n = 15) allowed differentiation of aortic stenosis ( n = 9) from sclerosis ( n = 6). This also permitted grading of stenosis with highly significant correlation ( T = 0.86; P < 0.002) with pressure gradient measurement. In mitral stenosis ( n = 12) calculation of the area of signal loss distal to the mitral valve as a percentage of left ventricular cross-sectional area showed a highly significant correlation ( T = 0.77; P = 0.001) with pressure gradient measurement. Clinically significant valvular regurgitation was graded by size and duration of signal loss proximal to the valve with concordance with angiocardiography. It is concluded that Cine-Flow MRI has a clinical role in the diagnosis and assessment of valvular heart disease. 0 1989 Academic Press, Inc.

INTRODUCTION

The diagnosis of valvular heart disease is primarily clinical. Cardiac catheterization, which until recently has been the only reliable method of determining its severity, is not without risk requiring the use of iodinated contrast medium and ionizing radiation ( I ). Conventional noninvasive methods of investigation, viz. phonocardi- ography and echocardiography, can be used to confirm or exclude disease but are unreliable in the assessment of its severity or hemodynamic significance (2, 3 ) . The development of Doppler techniques has improved reliability in some patients ( 4, 5 ) but may be impossible to perform in some circumstances, e.g., pulmonary em- physema.

Magnetic Resonance Imaging (MRI) is a noninvasive technique unaffected by the factors which may prevent echocardiography. It has been used to demonstrate cardiac morphology and in addition can provide functional information ( 6 ) . Recent devel- opments have enabled the demonstration and quantification of flow at different stages of the cardiac cycle in any plane (Cine-Flow MRI) . It now offers an alternative noninvasive method for the investigation of valve disease ( 7, 8 ) .

181 0740-3 194/89 $3.00 Copyright 0 1989 by Academic Press, Inc. A11 rights of reproduction in any form reserved.

182 MITCHELL ET AL.

CINE-FLOW MRI IN VALVULAR HEART DISEASE 183

TABLE I

Diagnoses following Cardiac Catheterization and 2D-Echocardiography

Aortic valve

Stenosis 23 (gradient 20- I20 mm Hg) Sclerosis 10 No gradient and normal 2D-echo 11 Normal volunteers 5 Regurgitation 19 mild 8

moderate 7 severe 4

Mitral valve

Stenosis 12 No gradient and normal 2D-echo 9 Normal volunteers 5 Regulgitation I3 mild 7

moderate 4 severe 2

The purpose of this study was to evaluate the clinical role of Cine-Flow MRI, compared with cardiac catheterization and echocardiography, in the assessment of valvular heart disease.

MATERIALS AND METHODS

Seventy-six valves in 49 subjects (age range 28-78, mean 52 years) were studied by MRI. Forty-four were patients attending the Department of Cardiology at Man- Chester Royal Infirmary with suspected valvular or ischemic heart disease. All patients were selected on the basis that they were to undergo cardiac catheterization and two-dimensional ( 2D) echocardiography. Five were normal volunteers with no history or evidence of heart disease. Two further patients were unable to undergo MRI due to claustrophobia and were excluded from the study. Of the 49 subjects, 2 1 were female and 28 male. Forty-one were in sinus rhythm and 8 in atrial fibrillation. In the 44 patients MRI was performed between 4 days before and 4 weeks after (mean 14 days after) catheterization.

Cardiac Catheterization

The extent of catheter studies varied according to clinical indication. The majority of patients ( n = 30) underwent full right and left heart catheterization ( 12 via the

FIG. 1. Compound oblique images through cardiac long-axis plane utilizing (A) Spin-Echo and (B) FEER sequences. Flowing blood gives a signal void in (A) and high signal in (B). v, left ventricle; a, left atrium; r, right ventricle; curved arrow, aortic valve; long straight arrow, mitral valve; short straight arrow, aorta.

184 MITCHELL ET AL.

FIG. 2. Normal subject: Six images in the same cardiac long-axis plane at 100 ms intervals through the cardiac cycle using the FEER sequence from ( A ) 30 ms to (F) 530 ms after the ECG R-wave. Curved arrow in (A) , signal loss distal to mitral valve at end-diastole; straight arrow in (C), peripheral signal loss distal to aortic valve in mid-systole: small arrows in (F), signal loss adjacent to mitral valve leaflets in diastole.

trans-septa1 route) and had both mitral and aortic valve pressure gradients measured. Mitral regurgitation was assessed by left ventricular angiography and aortic regurgitation by aortic root angiography. Fourteen patients with ischemic heart diesease or mild aortic valve disease underwent left heart catheterization only. This procedure involved left ventriculography and aortic valve pressure gradient measurement. Five of these patients also underwent aortic root angiography.

Aortic valve stenosis was graded according to peak to peak systolic pressure measurements and mitral valve stenosis by mean diastolic pressure gradients. Valvular regurgitation was graded qualitatively into mild, moderate, or severe on the basis of the angiographic findings. Those patients ( n = 10) with thickened, abnormal aortic valve leaflets on 2D-echocardiography, but with no pressure gradient at catheterization, were designated as having aortic valve sclerosis.

Magnetic Resonance Imaging Scanning was performed on a Picker International 0.26-T superconducting magnet

system. Informed consent was obtained in all subjects. All images were gated to the R-wave of the electrocardiograph (ECG).


FIG. 3. ( A ) Mid- and ( B ) end-diastolic images in a patient with a normal mitral valve demonstrating signal loss in the left ventricle following atrial systole (arrowed in B).

Six to nine image multisets in the same anatomical plane were performed using a Field Even Echo Rephasing (FEER) sequence, at 50-100 ms intervals for 600-800 ms after the ECG R-wave. Each data set utilized 128 gradient increments with two signal excitations and was collected in 256 cardiac cycles (3-6 min of patient's scanning time). Three compound oblique imaging planes (see below) were used. Total patient scan time was 35-55 min. If magnetic resonance (MR) blood flow velocity measurements were performed, an additional 15-35 min scanning time was allo- cated.

FEER Sequence

The FEER sequence described by Nayler et al. ( 9 ) , utilizes a field echo ( TE22-28 ms) with even-echo rephasing of the applied gradients to return the signal from flowing protons into phase with that from static tissue. This provides a higher signal from coherently flowing blood compared with stationary tissues as the constant replace- ment of protons in the slice due to flow prevents saturation (Fig. 1 ). Random or incoherent motion of spins is responsible for signal loss and can be observed in the flow jets of stenotic or regurgitant valves. A reduced (60") flip angle was employed to reduce saturation still further and permit rapid repetition of images in the same

186 MITCHELL ET AL.

FIG. 4. Patient with aortic and mitral stenosis and mitral regurgitation-six images through cardiac cycle using the FEER sequence and cardiac long axis plane with timings at 100 ms intervals from ( A ) 30 ms to (F) 530 ms after the ECG R-wave. Long arrow, signal loss distal to aortic valve in systole; curved arrow, signal loss proximal to mitral valve in systole; short arrow, signal loss distal to mitral valve in diastole.

anatomical plane. Intervals as low as 50 ms between images were obtained without significant signal loss from flowing blood. A gating delay after the R-wave was imple- mented to provide an additional set of interleaved scans if required.

Quantitative flow data (velocity measurements) were obtained by performing two sets of scans simultaneously. The first, a reference set, was performed as described above. In the second modifications were made to the gradient profile to encode for flow in a selected direction. Phase reconstructions were then performed and velocity flow maps prepared from subtraction of the two scans thus providing a direct measurement of flow rate in the selected direction.

Compound Oblique Imaging Planes

Plane 1 : Cardiac long axis as used in 2D-echocardiography demonstrates the proximal aorta, aortic valve, left ventricle, mitral valve, and left atrium, together with part of the right ventricle (Fig. 1 ). This plane was used in all patients in whom overall valvular assessment was required.


TABLE 2

Comparison of Pressure Gradient with Length of Aortic Signal Loss in Patients with Aortic Stenosis and Sclerosis

Pressure gradient (mm Hg) Length of signal loss (cm)

0 (n = 6) 5-9 (mean = 6) 20 12 30 13 30 13 40 14 40 15 60 16 80 16 80 17

120 16

Plane 2: Short-axis aorta and left ventricular outflow tract (LVOT) ensured the detection of small areas of signal loss. Such a plane is also perpendicular to the most probable line of flow and therefore provides the best opportunity for velocity measurements. Scans were performed 2 cm proximal and distal to the aortic valve and were used only when attempting to quantify aortic valve stenosis by velocity measurement.

Plane 3: A long-axis plane of the aortic arch was obtained in order to detect the full extent of signal loss distal to the aortic valve. If the aorta was tortuous a second parallel scan was performed to allow complete visualization of the whole arch. This was performed when scans in plane 1 showed signal loss distal to the aortic valve or if only aortic valve gradient was being assessed.

Due to length of scan time, especially if velocity maps were required, all three scans were not performed in all patients: 27 had scans in plane 1,20 in plane 2 ( 18 proximal and distal to the aortic valve and 2 distal only), and 23 in plane 3. Velocity maps were prepared from all the scans in plane 2. They were sensitized to measure flow in the slice select direction (i.e., along the long axis of the aortic valve) in order that a pressure gradient could be calculated by the modified Bernoulli equation as used in Doppler echocardiography ( 4 ) .

MR scans, including static and multiset cine-loop images, were assessed independently of catheterization and echocardiography results. The presence or absence of signal loss proximal and distal to the aortic and mitral valves was noted and its extent and duration recorded. If there was signal loss distal to the aortic valve and a scan along the aortic arch had been performed ( n = 15), the maximum length of signal loss in the aorta was measured. There was often a transitional zone in the area where signal returned and the linear measurement was therefore performed from the valve to the point where signal intensity value was halfway between that in the area of signal loss and that after full recovery. This measurement was correlated with catheter pressure gradient results using the Kendall’s rank correlation coefficient ( T ) . Multi- ple (4-7) measurements were made, and the mean value calculated, by one observer

188

18-

16 - 14-

12-

t lo- AORTIC SIGNAL LOSS 8- (cm)

6-

4-

2-

0-

MITCHELL ET AL.

+ + +

+ + *

+

+

+

+

0 20 40 60 m 100 120

PRESSURE GRADIENT (mmHg) - FIG. 5. Graph of peak-peak aortic valve pressure gradient against maximum length of signal loss distal

to valve on Cine-Flow MRI.

(L.M.) who had no knowledge of echocardiographic or cardiac catheter findings. Lin- ear measurements were in millimeters but were rounded to the nearest centimeter. These measurements were then independently checked by a second observer (JRRJ). The measurements never vaned by more than 0.6 cm. After an interval of several weeks the initial observer repeated the measurements which were again within 0.6 cm of the original value.

In patients with signal loss distal to the mitral valve in plane 1, the area of signal loss within the left ventricle was calculated using a full-width half-height method for boundary definition ( 10). This was expressed as a percentage of the left ventricular cross-sectional area (% LVA) and correlated with the mitral valve pressure gradient, measured at catheterization, using Kendall’s rank correlation coefficient ( T ) . In one patient with mitral stenosis the pressure recordings were lost and the data were therefore excluded.


FIG. 6. Aortic sclerosis-systolic images along aortic arch showing signal loss confined to the ascending aorta (long arrow). Note the coincidental mitral regurgitation (short arrow). Maximum length of signal loss = 5 cm.

In patients with signal loss proximal to a valve, the degree was assessed qualitatively as mild, moderate, or severe on the basis of size and duration of signal loss. This was correlated with the findings at angiocardiography.

Only results assessed by both MRI and cardiac catheterization have been included. Thus if left ventriculography and aortic valve catheter withdrawal were performed, only mitral regurgitation and aortic stenosis were considered. If MRI was limited to the aortic valve no comment was made on the mitral valve. A total of 76 valves were included (49 aortic and 27 mitral) and the diagnoses reached on cardiac catheterization and 2D-echocardiography are given in Table 1. Of the 10 patients with aortic sclerosis, 7 had aortic regurgitation of whom 5 also had rheumatic mitral valve disease. In the other 3 patients aortic root angiography was not performed.

RESULTS

MR images of diagnostic quality were obtained in all patients and volunteers. Good quality images were obtained in all subjects ( n = 41 ) in sinus rhythm (rate 50- 100 beats per min) and in six of eight patients with atrial fibrilation. In the remaining two, image quality was poor but still diagnostic. Respiratory motion artifact was not a

190 MITCHELL ET AL.

FIG. 7. Aortic stenosis (pressure gradient 80 mm Hg)-systolic images along aortic arch showing signal loss extending to descending aorta (mowed). Maximum length of signal loss = 16 cm.

problem in any subject. No differences in image quality or measurement were noted between scans utilizing TE’s of 22 and 28 ms.

Normal Appearances

In 6 of 16 normal aortic valves, partial signal loss, affecting mainly the peripheral part of the aorta, was observed distal to the valve during ventricular systole (Fig. 2). In two patients who had scans in plane 3, signal reduction extended no more than 2 cm from the valve. The remainder showed no signal loss distal to the valve. No signal loss was seen proximal to the valve in any normal volunteer. Small areas of signal loss were seen adjacent to the mitral valve leaflets in early diastole (Fig. 2). In 3 of 14 subjects an area of signal loss distal to the mitral valve was seen only at end-diastole following atrial systole (Figs. 2 and 3) . None showed signal loss proximal to the valve.

Aortic Stenosis/Sclerosis

All patients with aortic stenosis and sclerosis showed complete signal loss distal to the valve in systole (Fig. 4). Fifteen of these had scans along the aortic arch with measurement of maximum length of signal loss (Table 2; Fig. 5 ) . This allowed clear


FIG, 8. Cardiac long-axis images ( A ) early and ( B ) in mid-systole in a patient with aortic stenosis. Arrow shows signal loss proximal to the aortic valve in mid-systole.

differentiation between stenosis and sclerosis (Figs. 6 and 7 ) . In those with a pressure gradient there was a highly significant correlation ( T = 0.86; P < 0.002) between the seventy of pressure drop and the length of signal loss on MR. Mild, moderate, and severe stenosis could also be differentiated but at pressure gradients above approximately 60 mm Hg, no further increase in length of signal loss was seen (Fig. 5 ) . In one patient with coincidental aortic coarctation, signal loss was also seen arising from the coarctation segment.

During mid-systole 2 1 of 29 patients (4 not scanned proximal to the valve) with signal loss distal to the aortic valve also showed proximal signal loss which extended for 2-3 cm (Fig. 8). This mid-systolic proximal signal loss was seen in 5 of 8 patients with sclerosis and 16 of 2 1 with stenosis but not in patients without distal signal loss ( n = 16).

Velocity measurements were attempted in 20 subjects ( 19 patients and 1 volunteer). Velocity measurements in the normal range ( maximum postaortic valve velocities of 70-140 cm/s were achieved in the 8 subjects without aortic valve disease (7 patients, 1 volunteer). In the remaining 12 patients ( 10 with aortic stenosis, 2 with sclerosis) total signal loss prevented velocity measurement.

192 MITCHELL ET AL.

TABLE 3

Comparison of Pressure Gradient with Area of Left Ventricular Signal Loss as a Percentage of Cross-Sectional

Area (% LVA) in Mitral Stenosis

Pressure gradient (mm Hg) % LVA Aortic regurgitation"

4 6

10 I I 12 12 12 13 16 26

12 12 46 44 45 46 49 46 52 58

111 I1

I 111

I 0

111 I I

I1

Note. Degree of coincidental aortic regurgitation indicated. " Grading of regurgitation: 0 = none: I = mild: I1 = moderate; 111

= severe.

Mitral Stenosis

All patients with mitral stenosis showed an area of signal loss distal to the mitral valve during mid-diastole which varied in extent and duration (Fig. 4). The % LVA showed highly significant correlation ( T = 0.77; P = 0.001 ) with pressure gradient results (Table 3; Fig. 9). Difficulty was found in defining the size of the area of signal loss in patients with coincident moderate or severe aortic regurgitation where areas of signal loss were seen in the left ventricle in mid-diastole proximal to the aortic and distal to the mitral valve (Fig. 10). Using the cine-loop display it was possible to separate these areas and measure that distal to the mitral valve.

Aortic and Mitral Regurgitation (Tables 4 and 5 )

All patients with moderate or severe mitral and aortic regurgitation ( n = 17) at angiography, showed signal loss proximal to the valve at the appropriate phase of the cardiac cycle (systole for the mitral valve (Fig. 4), diastole for the aortic (Fig. I 1 )) . The area of signal loss on MR was smaller than the regurgitant jet seen at angiography. In patients with more severe regurgitation the area of signal loss was larger and persisted for a longer proportion of the cardiac cycle. In patients with coincident mitral stenosis and aortic regurgitation, diastolic signal loss was seen in the left ventricle due to both lesions. In mild aortic regurgitation, signal loss proximal to the aortic valve was observed only in early systole before maximum signal loss from mitral stenosis. With moderate or severe aortic regurgitation the areas of signal loss were coincident (Fig. 10). Utilization of the cine-loop display permitted separation and measurement of these areas of signal loss.

In 16 patients qualitative grading of regurgitation by MR and catheter concurred. In one patient with aortic regurgitation, who had only scans in plane 2 (short axis),

CINE-FLOW MRI IN VALVULAR HEART DISEASE

50 -

t 40- g, LVA

+ +

++*+

193

+

0 4 8 12 16 20 24 28 MITRAL VALVE PRESSURE GRADIENT

(mmHg) - a percentage of left-ventricular cross-sectional area on Cine-Flow MRI.

FIG. 9. Graph of mean mitral valve pressure gradient against area of signal loss distal to mitral valve as

catheter grading was severe and MR moderate (Table 4). In 9 of 14 patients with mild angiographic regurgitation signal loss was seen proximal to the valve and grading by MR and catheter agreed. In 5 patients no signal loss was seen (2 aortic, 3 mitral).

DISCUSSION

Cine-Flow MRI gives a high signal from flowing blood and can differentiate throm- bus and slow-moving blood ( 1 I ). Abnormalities which produce disturbed flow can be identified as a signal void. Flow velocity can be measured with good precision allowing calculation of volume flow rates in a reasonable time scale (typically 6-8 min) ( 9 ) . Unstable flow occurs when velocity rises above a certain level which varies mainly according to vessel radius. It can develop into turbulent flow if this flow rate is maintained for sufficient time and there is associated irregularity of the luminal wall. In normal hearts flow rates rise to unstable levels for only short periods and it is uncertain whether this is sufficient to produce turbulent flow ( 12). It is not, therefore, surprising that at times of peak flow sufficient disturbance of flow coherence may occur to cause reduction of MR signal in normal subjects. During valve closure local- ized flow reversal, producing signal loss (Fig. 2), occurs around the valve cusps and is instrumental in causing closure ( 12).

Signal loss on MR in patients with valvular stenosis or sclerosis may be due to several factors. Valve cusp irregularity will trigger turbulence at unstable flow rates

194 MITCHELL ET AL.

FIG. 10. Patient with mitral stenosis and aortic regurgitation-cardiac long-axis images in ( A ) early and (B) mid-diastole showing signal loss proximal to aortic (long arrow) and distal to mitral (short arrow) valves.

(i.e., peak flow in normal subjects). Flow through a narrowed orifice produces accel- eration. In severe aortic stenosis postvalvar velocities may reach 600 cm/s and at such high velocities disturbed flow is inevitable. In the aortic valve, the area of the valve must be reduced by half (i.e., a 50% stenosis) before a measurable pressure gradient develops ( 1 3 ). Patients with aortic sclerosis, therefore, may have a degree of narrowing which causes turbulence but is insufficient to produce a measurable gradi-

TABLE 4

Comparison of Grading of Aortic Regurgitation (n = 19) by Cine-Flow MRI and Angiography

Angiographic grading"

0 I I1 111 0 2 1 6 I1 7 1

111 3

MRI grading

' Grading of regurgitation as in Table 3.


TABLE 5

Comparison of Grading of Mitral Regurgitation (n = 13) by Cine-Flow MRI and Angiography

Angiographic grading“ ~

0 1 I1 111 0 3 I 4

I1 4 111 2

MRI grading

Grading of regurgitation as in Table 3.

ent. In the area distal to a narrowed valve the kinetic energy gained is dissipated in the blood and surrounding tissues. The more severe the stenosis the faster the flow rate, and recovery of coherent flow takes longer and occurs further from the valve. Turbulence arising from a valve will have a destabilizing effect for a short distance “upstream” thus producing disturbance of flow proximal to the valve ( 12), In addition to severity of stenosis, the length of MR signal loss distal to the aortic valve will

Fk. 1 1 . Cardiac long-axisimagesin (A)early and (B) mid-diastole showing signal loss(arrowed) proximal to the aortic valve due to aortic regurgitation.

196 MITCHELL ET AL.

increase with prevalvar velocity. Signal loss may also vary with aortic diameter and elasticity together with angle and size of branch vessel origins. These will affect the rate of absorption of kinetic energy ( 12). In addition, a single two-dimensional plane does not pass through the center of the whole of an aortic arch and thus the line of measurement may deviate from the middle of the aorta. This would be more marked in a tortuous vessel and could lead to the measured length being several millimeters different from the “central” length.

Our results indicate that even without allowance for these factors, length of signal loss measured by this method within the aorta provided a highly significant correlation ( T = 0.86; P < 0.002) with the pressure gradient measurements in aortic valve stenosis. Using this sequence signal loss less than 10 cm occurred without measurable pressure drop. Signal loss above this length increased directly with gradient measurement up to approximately 60 mm Hg where no further increase was observed. This allowed clear separation of sclerosis and stenosis and differentiation into mild, moderate, or severe stenosis. Further grading in those with pressure gradients greater than 60 mm Hg was, however, not possible. Attempts on Cine-Flow MRI to grade stenosis by measuring postvalvar velocity failed due to the severity of signal loss.

Similarly in mitral valve stenosis, extent of signal loss may vary with flow rate and left ventricular volume and compliance. Our results have demonstrated a highly significant correlation ( T = 0.77; P = 0.001 ) between size of signal loss as a proportion to left ventricular cross-sectional area and catheter measured gradient with a leveling off at gradients greater than 16 mm Hg. Difficulty occurred with coincident moderate or severe aortic regurgitation when signal loss in the left ventricle due to regurgitation persisted into mid-diastole and summated with that due to mitral stenosis. Use of a cine-loop display aided separation in these patients.

Velocity measurements attempted in patients with signal loss (viz. all those with stenosis or sclerosis) failed as there was no signal information from which to calculate a velocity map. Attempted maps showed only random noise encompassing the whole range of the velocity window. The conclusion of this part of the study is that it is not possible to quantify valvular stenosis by measuring pre- and postvalvar velocities using this technique. The velocities obtained in normal subjects were within the range of normal.

In both aortic and mitral stenosis MR assessment can separate those without significant hemodynamic disease from those who may require surgical therapy. This latter group would require cardiac catheterization to assess the coronary arteries and, if necessary, to provide more exact grading of valve disease particularly at the severe end of the spectrum.

Regurgitation produced disturbance of flow proximal to a valve giving rise to a signal void which increased in size and duration with the degree of regurgitation. In all cases of clinically significant regurgitation (moderate or severe), such a signal void was seen on MR and concurred with catheter results in all but one patient. In that patient only a short axis view was available which probably accounted for the under- estimate on MRI. Nine of 14 valves with mild regurgitation at angiocardiography were identified at MRI and gradings concurred. In the other five valves the degree of regurgitation was insufficient to produce a signal void on MRI. In some of them it is possible that the angiographic reflux was spurious. The presence of a catheter in the left ventricle may interfere with the mitral valve apparatus and induce mitral regurgi-


tation. Rapid infusion of contrast (20-30 ml at 20 ml/s) can provoke mild regurgitation in a normal aortic valve especially if the catheter lies just above it. Despite these features all instances of regurgitation likely to affect patient management were cor- rectly diagnosed by MR and there were no false positive diagnoses.

The cardiac long-axis plane (i.e., plane 1 ) was the most useful, allowing proximal and distal assessment of both mitral and aortic valves. It provided good overall display of the left heart chambers and allowed the best overall appreciation of valve function. The aortic and LVOT short-axis plane (plane 2) was of limited value. Signal loss occurred across the whole diameter of the aorta or LVOT and, therefore, the theoretical risk of missing a small jet when using a long-axis view did not materialize. The aortic arch long-axis plane (plane 3 ) was useful in patients with signal loss distal to the aortic valve and provided the only method for complete assessment.

CONCLUSION

Our results indicate that Cine-Flow MRI can reliably detect valvular heart disease even in the presence of an irregular rhythm. In the aortic valve, sclerosis can be clearly differentiated from stenosis and mild and moderate stenosis is separated from severe. Mitral valve stenosis can be graded by measuring the area of signal loss distal to the valve as a proportion of left ventricular cross-sectional area. Clinically significant regurgitation can be graded by observation of the size and duration of signal loss proximal to the valve. It is concluded that Cine-Flow MRI has a clinical role in the diagnosis and assessment of valvular heart disease.

ACKNOWLEDGMENTS

We acknowledge generous financial support from the Medical Research Council, Cancer Research Cam- paign, British Heart Foundation, Department of Health and Social Security, and Picker lntemational Ltd.

REFERENCES

1. W. GROSSMAN (Ed.), in “Cardiac Catheterization and Angiography” p. 25, Lea & Febiger, Philadel-

2. G. D. COPE, J. A. KISSLO, M. L. JOHNSON, AND V. S. BEHAR, Circulation 52,664 ( 1975). 3. A. N. DEMARIA, W. BOMMER, J. JOYE, G. LEE, J. BOUTELLER, AND D. T. MASON, Circulation 62,

4. L. HATLE, B. A. ANGELSEN, AND A. TROMSDAL, Br. Heart J. 43,284 ( 1980). 5 . M. BERGER, R. L. BERDOFF, P. E. GALLERSTEIN, AND E. GOLDBERG, J. Amer. Coll. Cardiol. 3, 150

6. J. P. R. JENKINS AND I. ISHERWOOD, in “Recent Advances in Cardiology” (D. J. Rowlands, Ed.), Vol.

7. S. R. UNDERWOOD, D. N. FIRMIN, K. H. UIPSTEIN, R. S. 0. REES, AND D. B. LONGMORE, Br. Heart

8. C. B. HIGGINS, W. HOLT, P. PFLUGFELDER, ANDU. S E C H T E M , ~ ~ ~ ~ ~ . Reson. Med. 6, 121 (1988). 9. G. L. NAYLER, D. N. FIRMIN,ANDD. B. LONGMORE, J.C.A.T. 10,715 (1986).

10. X. P. ZHU,D. R. CHECKLEY, D. S. HICKEY,AND~. ISHERWOOD, J.C.A.T. 10,96( 1986). I I. L. MITCHELL, J. P. R. JENKINS, W. C. BROWNLEE, AND I. ISHERWOOD, Clin. Radiol. 39,458 ( 1988). 12. C. G. CARO, T. J. PEDLEY, R. C. SCHROTER, AND W. A. SEED, “The Mechanics of the Circulation,”

13. S. S . YANG, L. G. BENTIVOGLIO, V. MARANHAO, AND H. GOLDBERG, “From Cardiac Catheterization

phia, 1980.

304 ( 1980).

(1984).

10, p. 2 19, Churchill Livingstone, Edinburgh, 1987.

J. 57,404 ( 1987).

Oxford Medical Publications, Oxford, 1980.

Data to Hemodynamic Parameters,” p. 177, Davis, Philadelphia, 1978.

diagnosis and assessment of mitral and aortic valve disease by cine-flow magnetic resonance imaging

Documents