| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Original Research |
1 Department of Medical Radiology, Institute of Diagnostic Radiology, University
Hospital Zurich, Zurich 8091, Switzerland.
2 Division of Cardiovascular Anaesthesiology, Institute of Anaesthesiology,
University Hospital Zurich, Zurich 8091, Switzerland.
3 Clinic for Cardiovascular Surgery, University Hospital Zurich, Zurich 8092,
Switzerland.
4 Department of Radiology, Spitaeler Chur AG, Loestrasse 170, Chur 7000,
Switzerland.
Received September 3, 2004;
accepted after revision October 25, 2004.
Address correspondence to T. Boehm
(thomas_boehm{at}gmx.net).
Abstract
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
SUBJECTS AND METHODS. Contrast-enhanced MDCT was performed in 37 patients who have a normal mitral valve, as shown on transesophageal echocardiography. Twenty CT data sets covering the valve apparatus were reconstructed every 5% step of the R-R interval. Multiplanar reconstructions were performed in the parallel short axis and perpendicular long axis of the left ventricle. Two independent blinded reviewers evaluated the image quality for dynamic cine-mode visualization of the valve components in systole and diastole and during the transitional phases in between.
RESULTS. Interobserver agreement for image quality ratings of valve components in all cardiac cycle phases ranged from good to excellent. Image quality for the visualization of valve leaflets, apposition zone, commissures, and mitral annulus (ranging from adequate to excellent) was significantly superior on perpendicular plane images than on parallel plane images for all cardiac phases (p < 0.05). Tendinous cords were visualized on both perpendicular and parallel planes with bad to adequate quality, whereas visualization of the papillary muscles was adequate to excellent on both imaging planes. Visualization of each valve component was superior in systole and diastole in both imaging planes as compared with the transitional phases (p <0.001).
CONCLUSION. Noninvasive cine-mode imaging of the mitral valve using retrospectively ECG-gated MDCT is feasible and allows accurate visualization of the moving valve. Perpendicular long-axis reconstructions yield images of superior quality when compared with the short-axis reconstructions and enable a determination of its functional morphology.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Transthoracic or transesophageal echocardiography is commonly the primary diagnostic tool for assessing anatomy and pathology of the mitral valve. Besides its capability of providing near real-time morphologic information, it delineates flow and derives hemodynamic data [4]. Echocardiography usually defines cardiac anatomy and function satisfactorily, often obviating further cardiac imaging.
The clinical applications of echocardiography include the detection and quantification of mitral stenosis, regurgitation, and mitral valve prolapse and the assessment of mitral annular calcification. Furthermore, it is able to depict systolic anterior motion of the mitral valve in patients with hypertrophic cardiomyopathy. However, echocardiography strongly relies on the morphologic characteristics of the patient; that is, it depends on a window that gives the interrogating beam adequate access to cardiac structures. For example, transthoracic echocardiography might be difficult in obese patients, patients with chest wall deformities, and those with chronic lung disease. On the other hand, transesophageal echocardiography is an invasive procedure and is contraindicated in patients with recent oral intake, prior esophageal surgery, unstable cervical spine injuries, or unevaluated gastrointestinal bleeding. Furthermore, echocardiography is strongly operator-dependent.
MRI is generally accepted as an accurate technique for monitoring adaptational changes in chamber dimensions associated with valve disease, assessing ventricular function, and quantitatively evaluating regurgitant valves [2]. However, current MRI techniques remain inferior for the depiction of valvular morphology, leaflet abnormalities, and motion [5, 6]. In addition, the inherent constraints of MRI, such as pacemaker implants, morbid obesity, and claustrophobia hinder MRI in becoming a clinically relevant technique for imaging cardiac valves.
MDCT has emerged as an imaging technique that can fully evaluate both cardiac structure and function. When combined with retrospective ECG-gating, MDCT allows visualization of the coronary arteries [7], detection and quantification of coronary calcification [8, 9], imaging of coronary artery soft-tissue plaques [10], and assessment of left ventricular ejection fraction [11]. To further broaden the clinical applications of MDCT, it would be desirable to evaluate valvular morphology and function as well. Early experience with conventional CT for the assessment of valve morphology has been reported [12]; however, the limited temporal resolution with consecutive motion artifacts has rendered this technique of almost no clinical value. More recently, retrospectively ECG-gated 4-MDCT has yielded good visualization of morphologic details of the mitral valve [13]. However, this study used data from one slice at a fixed interval during mid-diastole, and therefore only single, 2D, and static images of the mitral valve apparatus were obtained.
Today, imaging of the heart should be able to incorporate 3D data and, in addition, include the elements of time and motion, thus yielding information about the function of the cardiac valves [14, 15]. The purpose of this study was to evaluate the feasibility and image quality of retrospectively ECG-gated 16-MDCT for the dynamic visualization of the normal mitral valve apparatus using transesophageal echocardiography as the standard examination. By using the cine mode with 20 reconstructions in 5% steps of the ECG phase and covering the volume of the whole valve apparatus in two planes, we aimed to assess the dynamic morphology of the normal mitral valve throughout the heart cycle.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
MDCT
All 37 patients underwent scanning on a 16-MDCT scanner (Sensation 16,
Siemens Medical Solutions) with a gantry rotation time of 0.375 sec. One
hundred milliliters of iodixanol (Visipaque 320, 320 mg I/mL, Amersham Health)
was administered via a 20- to 22-gauge needle that was placed in a superficial
vein in the antecubital fossa. The contrast medium was administered using a
power injector (CT Injector, Ulrich Medical) at a rate of 4 mL/sec. For
optimal intraluminal contrast enhancement, the delay time between the start of
contrast medium administration and the start of imaging was determined for
each patient using a bolus-tracking technique (CARE-Bolus, Siemens Medical
Solutions). The region of measurement was placed in the ascending aorta, and
the threshold was set at 150 H. The contrast medium bolus was followed by a
30-mL saline chaser bolus administered at the same rate.
Repetitive low-dose monitoring examinations (120 kV, 10 mA, 0.5-sec scanning time, 1-sec inter-scan delay) were performed 10 sec after contrast medium injection began. After reaching the preset contrast enhancement level of 150 H, the MDCT examination was initiated automatically. Data acquisition was performed in a craniocaudal direction with a collimation of 16 x 0.75 mm, a table feed of 3 mm per rotation, and a gantry rotation of 0.375 sec (pitch, 0.25). The X-ray tube potential was 120 kV, and the effective tube current was 550 mA.
MDCT Data Postprocessing
Axial CT images were reconstructed from the CT raw data using a slice
thickness of 1 mm and an increment of 0.5 mm. For image reconstruction, a
segmented adaptive cardiac reconstruction algorithm was used
[16]. This algorithm uses raw
data from one subsegment of consecutive helical MDCT data from the same heart
period at heart rates below 65 bpm. At higher heart rates, two subsegments
from adjacent heart cycles contribute to the partial scan data segment.
Depending on patient anatomy, the reconstructed field of view was individually
fitted to the actual size of the heart in each patient (mean field of view,
211 mm; SD, 19; range, 182-268 mm; image matrix, 512 x 512 pixels).
Twenty data sets of axial image reconstructions at 5% steps of the R-R
interval were performed using a Bf30 medium soft-tissue kernel (Siemens
Medical Solutions).
From these data sets, multiplanar reconstructions parallel and perpendicular to the mitral valve ring were reconstructed in all 20 phases of the cardiac cycle using the multiplanar reconstruction postprocessing module on a workstation (Leonardo 3D-Card, Siemens Medical Solutions) by a radiologist with 5 years of experience in cardiovascular radiology. These two reconstruction planes were chosen to resemble the midesophageal transesophageal echocardiography views.
All reconstructions were planned using the axial CT series at 5% of the cardiac cycle (closed mitral valve). The parallel imaging plane was placed parallel to the mitral valve in the axial, coronal, and sagittal imaging planes using the multiplanar reconstruction tool. The reconstruction template was saved and was used for the other 19 cardiac phases to obtain 20 geometrically identical batches of multiplanar reconstructions corresponding to snapshots in 20 phases of the cardiac cycle.
The parallel plane was oriented along the short axis of the left ventricle (i.e., parallel to the closed mitral valve) and included a small part of the left atrium, the mitral leaflets, commissures, annulus, and the subvalvular apparatus including the tendinous cords and papillary muscles (Fig. 1A). The parallel sections were reconstructed with a section thickness of 1 mm and an increment of 0.5 mm, resulting in a mean of 173 images (range, 160-190; 20 reconstructions each; total number of images, 3,200-3,800).
|
|
|
MDCT Image Analysis and Readout
The reconstructed data were presented for readout on a workstation
(Leonardo, Siemens Medical Solutions) using a dedicated commercially available
software tool for interactive cardiac functional analysis (Syngo Argus 2.0,
Siemens Medical Solutions). With the aid of this software, the whole volume
could be viewed in the cine mode, thus providing a real-time impression of the
moving mitral valve. For the whole volume of the valve apparatus to be
covered, the images of the cine-mode video sequence are arranged so that each
single slice was shown throughout an entire cardiac cycle before scrolling to
the next slice. On these cine-mode videos, artifacts and image quality were
analyzed by two independent radiologists, each with 7 years of experience in
cardiovascular radiology. The velocity of the videotape presentation could be
determined individually by the two reviewers. Both reviewers were allowed to
individually adjust window center and window level settings for image
analysis. For documentation, at least two videos (one in the parallel plane
and one in the perpendicular plane) of the valve in the closed, open, and
transitional phases were stored (video format; avi, frame rate; 25/sec,
resolution; 463 x 463 pixels, compression; Cinepak Codec).
Contrast Inflow Artifacts
The two reviewers assessed the degree of artifacts deriving from contrast
material inflow into the superior vena cava and right atrium throughout the
cardiac cycle by using the following scores: grade 1, severe artifacts,
nondiagnostic image quality; grade 2, fair artifacts, severely compromised
image quality; grade 3, few artifacts with slight compromise of image quality;
and grade 4, no or few artifacts, no compromise in image quality. The
artifacts were assessed during videotape presentation of the parallel and
perpendicular planes. Because artifacts may vary during the cardiac cycle and
may depend on the imaging plane, the most severe artifact level encountered
was recorded. Only artifacts compromising assessment of the valve were
considered.
Synchronization Artifacts
Artifacts deriving from eventual misregistration between the
software-detected ECG signal and cardiac motion, identified as one or numerous
parallel and straight lines along which the cardiac contours show abrupt
steplike distortions, were rated by both reviewers on a 4-point Likert scale:
grade 1, severe artifacts preventing assessment of the valve; grade 2, fair
artifacts severely compromising its assessment; grade 3, few artifacts with
slightly compromised assessment; and grade 4, no artifacts. Synchronization
artifacts were assessed during videotape presentation in the parallel and
perpendicular planes. The most severe artifact level encountered was
recorded.
Image Quality of the Mitral Valve Apparatus in the Cine Mode
For image quality analysis, the mitral valve apparatus was divided into the
following components: leaflets, zone of leaflet apposition during valve
closure, commissures, annulus, tendinous cords, and papillary muscles. The
image quality of MDCT data was assessed in the parallel and perpendicular
planes and included the evaluation of all valve components during mid-systole
(i.e., closed valve), mid-diastole (i.e., open valve), and during the phases
in between. The two phases between mid-systole and mid-diastole were analyzed
together and were termed "transitional phases," defined as the
phases of rapid valve motion between mid-systole and mid-diastole and vice
versa. Both reviewers independently assessed the image quality for each of
these anatomic components on a 4-point Likert scale. Grade 1 indicated bad
image quality, which meant that morphologic information was not obtained.
Grade 2 indicated adequate image quality, which meant that all morphologically
relevant information was obtained. Grade 3 indicated good image quality, which
meant that all morphologic information was obtained with good anatomic
differentiation of the valve apparatus. Grade 4 indicated excellent image
quality, which meant that all morphologic information was obtained with
excellent anatomic differentiation of the valve apparatus.
Best Phase for Reconstruction in Systole and Diastole
A single reviewer determined the best phase for reconstruction of the MDCT
data for the visualization of the open and closed (i.e., mid-systole and
mid-diastole) valve leaflets in both reconstruction planes. This readout was
done after pausing the cine-mode videotape and scrolling through the 20
cardiac phases until finding the best phase for the visualization of the
leaflets.
Image Noise and Contrast Enhancement
The image noise and contrast enhancement were measured by a radiologist
with 3 years of experience in cardiovascular imaging. Image noise (i.e., the
SD of the attenuation) was measured using a circular region of interest (ROI)
(mean diameter, 11 mm; range, 9-13 mm) placed in the air adjacent to the left
ventrolateral chest wall at the level of the left atrium and ventricle. The
same radiologist also assessed the absolute degree of contrast enhancement (in
Hounsfield units) and the SD by placing circular ROIs of the same size in the
left atrium and left ventricle. Intracardiac SD measurements do not represent
true noise measurements because they contain additional information (i.e.,
inhomogeneities caused by contrast medium distribution and corpuscular
constituents of blood). The contrast-enhanced blood represents the background
against which the valve is imaged. Therefore, the stochastic inhomogeneity of
this background is of interest when assessing image quality of the mitral
valve. All measurements were performed in mid-diastole at 65% of the R-R
interval in the axial source images.
Radiation Exposure
The radiation exposure of MDCT was calculated using a commercially
available computer program (WinDose [version 2.1a],
Scanditronix-Well-höfer Dosimetrie) that is based on Monte Carlo
calculations for anthropomorphic mathematic phantoms
[17].
Echocardiography
Intraoperative transesophageal echocardiography is performed routinely at
our institution by cardiac anesthesiologists using a standardized database
reporting system [18]. All 37
patients were premedicated with oral benzodiazepines (flunitrazepam or
midazolam). In the operating theater, anesthesia was induced, the patient's
trachea was intubated, and the lungs were mechanically ventilated. After
insertion of a central venous catheter and a pulmonary artery catheter (if
considered necessary), the latex-sheathed transesophageal echocardiography
probe was inserted, usually with an Esmarch maneuver or with help of direct
laryngoscopy. A multiplane 5-MHz transesophageal echocardiography probe (Sonos
5500, Philips Medical Systems) equipped with pulsed wave, continuous wave, and
color Doppler capabilities was used. All transesophageal echocardiography
examinations were performed by the same experienced echocardiographer and
included B- and M-mode echocardiography combined with color Doppler
examination. According to the guidelines of the American Society of
Echocardiography [4], a
standard transesophageal echocardiography examination includes the assessment
of the biventricular systolic and diastolic function, regional wall motion,
valvular function, and aortic anatomy. A search for a shunt—in
particular, a patent foramen ovale—was performed. The mitral valve was
examined using bidimensional echocardiography and color Doppler imaging in
four mid-esophageal (0°, 60°, 90°, and 120°) and two
transgastric (0° and 90°) views
[4]
(Fig. 1C).
Overall Image Quality of Transesophageal Echocardiography
The transesophageal echocardiography videotapes were analyzed by the same
echographer who performed the transesophageal echocardiography examination.
They were assessed in terms of the overall image quality of mitral valve
visualization with the same 4-point Likert scale as described earlier.
Statistical Analysis
Statistical analysis was performed using commercially available software
(SSPS 11.5, Statistical Package for the Social Sciences) for Windows
(Microsoft). Interobserver agreement between both reviewers who evaluated the
MDCT data was calculated using kappa statistics. According to Landis and Koch
[19], a kappa value of zero
indicates poor agreement, a kappa value of 0.01-0.20 indicates slight
agreement, a kappa value of 0.21-0.40 indicates fair agreement, a kappa value
of 0.41-0.60 indicates moderate agreement, a kappa value of 0.61-0.80
indicates good agreement, and a kappa value of 0.81-1.00 indicates excellent
agreement. The differences between the two reviewers' image quality ratings
for each anatomic structure shown on cine-mode MDCT were compared using the
Wilcoxon's signed rank test. The same test also was used to compare the image
quality of the individual mitral valve components between the parallel and
perpendicular reconstruction planes. We considered p values of less
than 0.05 to indicate statistically significant differences.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Contrast Inflow Artifacts
Artifacts related to the inflow of contrast material into the superior vena
cava and right atrium were rated as not present (grade 4) in 25 (68%) by
reviewer 1 and in 24 (65%) of 37 patients by reviewer 2. Reviewer 1 and
reviewer 2 rated inflow artifacts as few (grade 3) in 11 (30%) and 12 (32%)
patients, respectively. These reviewers rated image quality as fair (grade 2)
in one (3%), and no reviewer rated inflow artifacts as severe (grade 1) in any
patient.
Synchronization Artifacts
Artifacts related to possible misregistration of the MDCT data with the ECG
signal were rated as not present (grade 4) in 22 (59%) by reviewer 1 and in 20
(54%) of 37 patients by reviewer 2. Reviewer 1 and reviewer 2 rated
synchronization artifacts as few (grade 3) in 13 (35%) and 14 (38%) patients,
respectively. The reviewers rated synchronization artifacts as fair (grade 2)
in two (5%) and three (8%) of 37 patients, respectively. In none of the
patients were synchronization artifacts rated as severe (grade 1).
Image Quality of the Mitral Valve Apparatus in the Cine Mode
Valve leaflets—In the parallel plane, image quality for
showing the leaflets during mid-systole was rated by reviewer 1 as 1.4
± 0.5 and by reviewer 2 as 1.5 ± 0.6 (p =0.180, 
= 0.758); during the transitional phases, as 1.1 ± 0.3 and 1.1 ±
0.4 (p = 0.157, = 0.773); and during mid-diastole, as 2.1
± 0.5 and 2.0 ± 0.7 (p = 0.705, = 0.677),
respectively. In the perpendicular plane, image quality during mid-systole was
rated by both reviewers as 3.9 ± 0.3 (p = 0.839, =
0.790); during the transitional phases, as 2.3 ± 0.6 and 2.4 ±
0.6 (p = 0.056, = 0.759), respectively; and during
mid-diastole, by both as 3.7 ± 0.5 (p = 1.000,
=0.828).
Leaflet apposition during mid-systole—The image quality for
showing the leaflet apposition zone was rated in the parallel plane by
reviewer 1 as 2.4 ± 0.8 and by reviewer 2 as 2.5 ± 0.8
(p = 0.705, = 0.736); in the perpendicular plane, it was
rated as 3.9 ± 0.3 and 3.8 ± 0.4 (p = 0.083, =
0.689), respectively.
Commissures—In the parallel plane, image quality for showing
commissures during mid-systole was rated by both reviewers as 1.3 ± 0.5
(p = 0.157, = 0.886); during the transitional phases, as 1.0
± 0.2 and 1.1 ± 0.3 (p = 0.157, = 0.643),
respectively; and during mid-diastole, by both as 1.5 ± 0.6 (p
= 0.414, = 0.720). In the perpendicular plane, image quality during
mid-systole was rated as 3.7 ± 0.5 and 3.8 ± 0.5 (p
=0.414, = 0.759); during the transitional phases, as 1.7 ± 0.6
and 1.8 ± 0.6 (p = 0.157, = 0.762); and during
mid-diastole, as 3.2 ± 0.7 and 3.3 ± 0.8 (p = 0.096,
= 0.912), respectively.
Mitral annulus—In the parallel plane, image quality for
showing the mitral annulus during mid-systole was rated as 3.4 ± 0.5
and 3.5 ± 0.5 (p = 0.054, = 0.612); during the
transitional phases, as 2.7 ± 0.5 and 2.8 ± 0.5 (p =
0.206, = 0.474); and during mid-diastole, as 3.1 ± 0.6 and 3.3
± 0.6 (p =0.059, = 0.688), respectively. In the
perpendicular plane, image quality during mid-systole was rated as 3.9
± 0.3 and 3.9 ± 0.4 (p =0.317, = 0.791); during
the transitional phases, as 3.1 ± 0.8 and 3.2 ± 0.7 (p
= 0.109, = 0.489); and during mid-diastole, as 3.3 ± 0.9 and
3.4 ± 0.7 (p = 0.206, = 0.607), respectively.
Tendinous cords—In the parallel plane, image quality for
showing the tendinous cords during mid-systole was rated by both reviewers as
1.3 ± 0.6 (p = 0.655, = 0.647); during the
transitional phases, as 1.1 ± 0.3 and 1.1 ± 0.4 (p =
0.157, = 0.650); and during mid-diastole, as 1.1 ± 0.3 and 1.1
± 0.3 (p = 1.000, = 0.821), respectively. In the
perpendicular plane, image quality during mid-systole was rated as 1.5
± 0.8 and 1.6 ± 0.9 (p = 0.083, = 0.861);
during the transitional phases, by both as 1.2 ± 0.5 (p
=1.000, = 0.875); and during mid-diastole, as 1.3 ± 0.7 and 1.3
± 0.6 (p = 0.157, = 0.664), respectively.
Papillary muscles—In the parallel plane, image quality for
showing the papillary muscles during mid-systole was rated by reviewer 1 as
3.7 ± 0.5 and by reviewer 2 as 3.6 ± 0.5 (p = 0.065,
= 0.745); during the transitional phases, as 2.4 ± 0.6 and 2.6
± 0.6 (p =0.059, = 0.680); and during mid-diastole, as
3.6 ± 0.5 and as 3.7 ± 0.5 (p = 0.102, =
0.696), respectively. In the perpendicular plane, image quality during
mid-systole was rated by both as 3.4 ± 0.5 (p = 0.655,
= 0.761); during the transitional phases, as 2.3 ± 0.5 and 2.4 ±
0.5 (p = 0.058, =0.655); and in mid-diastole, as 3.7 ±
0.4 and 3.6 ± 0.5 (p = 0.069, = 0.728),
respectively.
Table 1 summarizes the breakdown of findings of both reviewers regarding image quality for showing the individual mitral valve components using cine-mode 16-MDCT. Because of nonsignificant differences and high interobserver agreement, the mean image quality scores calculated from both reviewers are listed. Figure 2 shows the images in 10% step reconstructions of the R-R interval in the parallel short-axis plane, and Figure 3 shows the images in the perpendicular long-axis reconstruction.
|
|
|
Comparison Between Parallel and Perpendicular Planes
Image quality for the visualization of the leaflets, apposition zone, and
commissures was better in the perpendicular plane than the parallel plane in
all cardiac phases (p < 0.0001, p < 0.0001, p
< 0.05, respectively). The mitral annulus was better visualized in the
perpendicular plane than the parallel plane during mid-systole (p
< 0.01) and the transitional phases (p < 0.05) and differed
nonsignificantly between planes during mid-diastole (p = 0.217).
Image quality for the visualization of tendinous cords was superior in the
perpendicular plane during mid-systole (p <0.05) and mid-diastole
(p < 0.01) and differed nonsignificantly during the transitional
phases (p = 0.116). Image quality for papillary muscle visualization
was superior in the parallel plane in mid-systole and during the transitional
phases (p < 0.0001) and differed nonsignificantly in mid-diastole
(p = 0.197).
Comparison of Systole, Diastole, and Transitional Phases
Overall image quality for the visualization of the valve components in the
parallel plane was superior during mid-diastole when compared with mid-systole
(p < 0.001) and the transitional phases (p < 0.0001)
and was superior during mid-systole when compared with transitional phases
(p < 0.0001). In the perpendicular plane, the apparatus was better
visualized during mid-systole than during mid-diastole (p <
0.0001) and the transitional phases (p < 0.0001) and was better
visualized during mid-diastole than during the transitional phases (p
< 0.0001).
|
Image Noise and Contrast Enhancement
The overall image noise measured in the air adjacent to the chest wall was
19 ± 10 H. The mean contrast enhancement in the left atrium and left
ventricle was 321 ± 71 H and 335 ± 74 H, respectively. The image
noise in the left atrium and left ventricle was 35 ± 28 H and 39
± 32 H, respectively.
Radiation Exposure
Calculated estimated effective radiation doses based on Monte Carlo
calculations for anthropomorphic mathematic phantoms
[17] were 12.0 mSv for men and
15.8 mSv for women.
Overall Image Quality of Transesophageal Echocardiography
The overall image quality of the transesophageal echocardiography
examination was rated by the echographer as adequate (grade 2) in five (14%),
as good (grade 3) in 13 (35%), and as excellent (grade 4) in 19 (51%) of 37
patients (mean, 3.6 ± 0.6). No transesophageal echocardiography
examination was rated as yielding images of bad quality (grade 1). Figures
5 and
6 show the typical
transesophageal echocardiography appearance of the normal mitral valve in the
transgastric short-axis and midesophageal four-chamber long-axis views during
eight phases equally distributed throughout the cardiac cycle.
|
|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Dynamic Morphology of the Mitral Valve
Accurate imaging of mitral valve components requires a high temporal and
spatial resolution because of the rapid and complex motion of the valve. In
addition to the intrinsic movements of the individual valve components, the
plane of the whole valve itself moves during diastole upward into the left
atrium and toward the left ventricular apex during systole
[21,
22].
The valve is closed during mid-systole, and the anterior meets the posterior leaflet to form an arc-shaped zone of apposition [23]. This apposition zone could be visualized with cine-mode MDCT in the parallel and perpendicular planes with good and excellent image quality, respectively. When the valve starts to open, the leaflet curvature flattens, becomes reversed, and moves into the left ventricle [24]. After maximal opening of the valve, the leaflet edges exhibit a to-and-fro movement until another less forceful opening impulse occurs with atrial contraction. Valve closure starts with leaflet bulging toward the atrium at its attachment point to the annulus. With cine-mode MDCT, we were able to visualize the closed and open leaflets and commissures in the perpendicular plane with a good to excellent image quality. Because of the fast movements of leaflets and commissures in the transitional phases, however, image quality during this part of the cardiac cycle was significantly reduced. Image quality for showing the leaflets and commissures was significantly lower in the parallel plane than in the perpendicular plane, which can be explained by the anatomic orientation of the structures being almost parallel to the short-axis plane. In addition, the continuous and fast change of the valve's plane toward the atrium and in the opposite direction toward the ventricular apex hampers imaging in the short axis, where the structures continuously and rapidly leave the actual imaging plane. In contrast, on perpendicular images the up- and downward movements of the valve apparatus together with the intrinsic valve movements can be visualized on a single reconstruction plane.
The mitral annulus contributes to timely, efficient, and competent valve closure and unimpeded left ventricular filling [25]. The annulus moves toward the left atrium during late diastole (i.e., left atrial filling), remains immobile during systole (i.e., mitral valve closure), and descends toward the left ventricular apex during isovolumic contraction and ventricular ejection [22]. Subsequently, the annulus moves little during isovolumic relaxation but then exhibits a rapid recoil back toward the left atrium in early diastole (i.e., rapid blood flow into the ventricle). This annular movement pattern explains the excellent dynamic cine-mode visualization of the annulus when the valve is closed and the good visualization when the valve is open. As could also be expected from this motion pattern, long-axis perpendicular planes provided a superior visualization of the mitral annulus throughout the entire cycle.
The tendinous cords are fine, fibrous, stringlike structures that attach the ventricular surface or the free edge of the leaflets to the papillary muscles [23]. Normal valves have a spectrum of cordal support, and uniformity of cordal attachments to the leaflets is uncommon [26]. The thickness of the tendinous cords in normal mitral valves ranges between 0.4 and 1.2 mm [27]. This small size and anatomic variability together with the rapid movement may contribute to the insufficient quality of cine-mode MDCT images in both the parallel and perpendicular planes.
The papillary muscles comprise the muscular components of the mitral apparatus. The anterolateral papillary muscle is commonly single, whereas the posteromedial usually has multiple heads [23]. Papillary muscle contraction pulls the two leaflets toward one another and thereby promotes valve closure. Both papillary muscles closely mimic left ventricular dynamics—that is, the muscles shorten during ejection, lengthen during diastole, and minimally change their length during isovolumic relaxation periods in systole [28]. With cine MDCT, we were able to visualize the papillary muscles with a good to excellent image quality when the valve was open and when it was closed, and only a slight reduction of image quality occurred in the transitional phases. This is explained primarily by the large dimensions of papillary muscles compared with the other components of the mitral valve. The slightly better visualization of papillary muscles in the parallel compared with the perpendicular reconstruction is due to the orientation of the muscle bellies, running almost perpendicularly through the short-axis plane and thus being better visualized.
Technical Considerations in Cardiac CT
The scanner used in this study has a gantry rotation time of 375 msec. When
data from a 180° gantry rotation are used for image reconstruction, the
temporal resolution is 187.5 msec. In case of segmented adaptive cardiac
reconstruction from two cardiac cycles, the temporal resolution is 93.75 msec.
Besides the advantage of this algorithm to provide a narrow slice sensitivity
profile [16], it has the
disadvantage that the temporal resolution depends on the heart rate. In the
scanner used in this study, the pitch is fixed and does not depend on the
heart rate or switching of the reconstruction algorithm from one- to
two-segmented reconstruction. Therefore, the applied radiation does not depend
on the reconstruction algorithm. In other scanner types with automatic pitch
adjustment, the use of this algorithm may result in longer scanning times and
higher radiation exposure. We were able to obtain good to excellent results
for imaging the mitral valve in the closed and open states but inferior
results for imaging the mitral valve during the transitional states. The
current restrictions may cease when applying faster imaging options in the
future. This may be achieved either by faster gantry rotation or by image
reconstruction using data from more than two R-R intervals. The latter
approach is technically easier, but higher temporal resolution is achieved at
the expense of a higher level of interpolation, causing inferior spatial
resolution.
Given the mean field of view of 211 mm in the present study, the spatial in-plane voxel size was 0.41 mm. The craniocaudal voxel size (i.e., z-axis resolution) was 1 mm. The 3D voxel size was, therefore, nonisotropic. However, this spatial resolution allowed the imaging of all components of the mitral valvular apparatus except the tendinous cords in the appropriate plane with a good to excellent image quality. This might be explained by our limited z-axis resolution because the orientation of the tendinous cords is almost parallel to the orientation of the original axial data set from which the reconstructions were performed. This results in a significant partial volume effect with the surrounding contrast medium and may lead—in combination with the mentioned anatomic variability, small dimensions, and rapid movements—to a confounded image quality.
Image quality of cardiac-gated data sets generally requires that data acquisition be synchronized to the periodic motion of the heart. It is thus necessary to accurately time the data acquisition within each heartbeat to consistently capture the same phase of the cardiac cycle from one heartbeat to the next [29]. This usually is done by reconstructing data in predefined percent steps of the R-R interval. However, the percent approach to sample the data from two consecutive heartbeats may result in inconsistencies in the position of the valve components in patients with sinus arrhythmia. Reconstructing data using the absolute gating option from the R peak could possibly improve image quality of cardiac CT in this particular group of patients when compared with the relative reconstruction approach.
Cardiac MDCT commonly relies on the ECG signal reflecting the properties of the electrical conduction system and only indirectly mirroring the mechanical performance of the heart. Synchronization artifacts usually are more pronounced because of the much faster heart motion during systole and the transitional phases when compared with diastole. Because image reconstruction for CT angiography of the coronary arteries is commonly performed in diastole, the issue of synchronization artifacts throughout the cycle did not play a major role in cardiac imaging until now.
The relatively short acquisition time with 16-MDCT scanners theoretically permits timing of the contrast medium-saline chaser bolus in a way that the left ventricle is maximally opacified, whereas the contrast medium has already been washed out from the right heart. The acquisition protocol in our study was aimed at this effect, but the results were not satisfying in all patients. Contrast medium-related artifacts did not cause severe problems in any patient, but the combination of artifacts of different origins might exponentiate and cause a critically bad image quality. Shorter acquisition times in the future might allow better bolus timing and thus may completely eliminate contrast medium-related artifacts.
Image noise is a major problem in cardiac-gated MDCT. The temporal resolution is optimized at least partially at the expense of increasing the image noise. Except for increasing the radiation dose (which is not legitimate due to ethical grounds), the user has no means for affecting image noise. For the assessment of the mitral valve, the amount of image noise in the left ventricular and atrial cavities is crucial. The difference between noise outside the heart and that measured in the left atrium and ventricle in the present study shows that the structural inhomogeneity of contrast-enhanced blood represents a considerable additional source of noise.
The best percent phase of reconstruction as determined in the current study can be used as an orientation to which phase the open and closed mitral leaflets can be visualized with the best image quality. These phases then may be used to assess anatomic details of the valves in the respective phases. However, we do not propose limiting mitral valve imaging to these two phases but suggest assessing the valve in the dynamic cine mode. Motion has been shown to be a powerful informative diagnostic cue, because motion perception influences the perception of structure [30]. Elements of structure may be perceived on the basis of motion even if the static images alternated to generate the motion do not themselves contain the information. Moreover, authors of a number of experiments have argued that perception of motion is antecedent to that of object quality [31]. It is therefore not surprising that motion in diagnostic low-contrast images, such as the motion of the valve during the transitional phases, is still informative.
Limitations
We acknowledge the following limitations of this study. We did not address
the time issue for postprocessing the MDCT data. Lacking specialized software
for multiphase image reconstruction and 4D presentation of the data set, we
were forced to reconstruct the multiplanar reconstruction data for every part
of the cardiac cycle individually. Future software changes will shorten this
postprocessing time and possibly allow time-effective assessment of cardiac
valves directly from the initial axial CT data set.
Another drawback for the use of CT is the applied radiation dose inherent with the technique. During ECG-gated MDCT, data acquisition requires an overlapping helical pitch and continuous X-ray exposure, which results in a considerable estimated effective radiation dose. We did not apply systolic dose reduction because this study aimed at obtaining high-quality images throughout the cardiac cycle. When applicable, the mean effective radiation dose may be reduced by 48% for men and 45% for women [32]. On the other hand, the data set was used not only for imaging the mitral valve but also primarily for assessing coronary arteries, and the same data may be used to quantify ventricular ejection fractions. This integrated concept of functional cardiac imaging may at least partially justify the applied radiation dose.
Finally, we did not examine patients with mitral valve abnormalities. However, understanding the normal anatomy of the constituents of the mitral valve on dynamic CT not only helps in the examination of these parts but also enhances the appreciation of anatomic variants and disease, the latter of which will be addressed in our next study. It is to be expected that the current MDCT technique is able not only to depict structural abnormalities of the mitral valve [13], but also to assess pathologic functional processes such as mitral stenosis and regurgitation or systolic anterior motion.
Conclusion
This study has prospectively shown that MDCT is feasible, accurate, and
reliable for the dynamic morphologic assessment of the normal mitral valve.
Perpendicular reconstructions in the perpendicular long-axis plane provide
superior visualization of the mitral valve components throughout the cardiac
cycle when compared with the parallel short-axis plane. A thorough examination
of the valve can be performed from the same data set that has been acquired
for the evaluation of the coronary arteries or for the quantitative assessment
of ventricular function. Familiarity with the normal dynamic morphology of the
valve is a prerequisite for diagnosing any deviation from the norm in the
future. However, the realization of the complex valvular anatomy and motion
notwithstanding, the CT assessment of mitral valve abnormalities and the
development of clinically available hardware and software with sufficient
computer power remain major challenges. These must be overcome if
multidimensional cardiac CT is to become a clinical reality.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  J. J. Chen, M. A. Manning, A. A. Frazier, J. Jeudy, and C. S. White CT Angiography of the Cardiac Valves: Normal, Diseased, and Postoperative Appearances RadioGraphics, September 1, 2009; 29(5): 1393 - 1412. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Y.-K. Guo, Z.-G. Yang, G. Ning, L. Rao, L. Dong, Y. Pen, T.-M. Zhang, Y. Wu, X.-C. Zhang, and Q.-L. Wang Isolated Mitral Regurgitation: Quantitative Assessment with 64-Section Multidetector CT--Comparison with MR Imaging and Echocardiography Radiology, August 1, 2009; 252(2): 369 - 376. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 V. Delgado, L. F. Tops, J. D. Schuijf, A. de Roos, J. Brugada, M. J. Schalij, J. D. Thomas, and J. J. Bax Assessment of mitral valve anatomy and geometry with multislice computed tomography. J. Am. Coll. Cardiol. Img., May 1, 2009; 2(5): 556 - 565. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
F. A. Flachskampf and D. Ropers Computed tomography to analyze mitral valve an answer in search of a question. J. Am. Coll. Cardiol. Img., May 1, 2009; 2(5): 566 - 568. [Full Text] [PDF]
|
|
|
|
 |
|
 B. Sundaram, S. Patel, P. Agarwal, and E. A. Kazerooni Anatomy and Terminology for the Interpretation and Reporting of Cardiac MDCT: Part 2, CT Angiography, Cardiac Function Assessment, and Noncoronary and Extracardiac Findings Am. J. Roentgenol., March 1, 2009; 192(3): 584 - 598. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 O UCAR, M VURAL, H CICEKCIOGLU, L PASAOGLU, S AYDOGDU, and S KOPARAL Multidetector CT presentation of a parachute-like asymmetric mitral valve Br. J. Radiol., November 1, 2008; 81(971): e266 - e268. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. Ryan, S. Abbara, R. R. Colen, S. Arnous, M. Quinn, R. C. Cury, and J. D. Dodd Cardiac Valve Disease: Spectrum of Findings on Cardiac 64-MDCT Am. J. Roentgenol., May 1, 2008; 190(5): W294 - W303. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. F. Tops, S. C. Krishnan, J. D. Schuijf, M. J. Schalij, and J. J. Bax Noncoronary applications of cardiac multidetector row computed tomography. J. Am. Coll. Cardiol. Img., January 1, 2008; 1(1): 94 - 106. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. I. Fann, N. B. Ingels Jr., and D. C. Miller Pathophysiology of Mitral Valve Disease Card. Surg. Adult, January 1, 2008; 3(2008): 973 - 1012. [Full Text]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
