| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Thoracic Imaging |
1 From the Departments of Radiology (K.F.J.K., S.L., H.U.K., M.B.P., M.T.); Cardiothoracic and Vascular Surgery (E.M., T.K.); and Medical Biometry, Epidemiology and Informatics (F.K.); Johannes Gutenberg-University Mainz, Langenbeckstrasse 1, D-55131 Mainz, Germany; and Department of Radiology, German Cancer Research Center, Heidelberg, Germany (S.L., H.U.K.). From the 1999 RSNA scientific assembly. Received June 6, 2003; revision requested August 21; final revision received November 28; accepted January 5, 2004. Supported in part by Schering, Berlin, Germany. Contract grant sponsor: Deutsche Forschungsgemeinschaft; contract grant number, FOR 474/1. Address correspondence to K.F.J.K. (e-mail: kreitner@radiologie.klinik.uni-mainz.de).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
MATERIALS AND METHODS: Thirty-four patients with CTEPH were examined before and after pulmonary thromboendarterectomy (PTE). For morphologic assessment, contrast material–enhanced MR angiography was used; for assessment of hemodynamics, velocity-encoded gradient-echo sequences and cine gradient-echo sequences along the short axis of the heart were performed. Contrast-enhanced MR angiography was compared with selective digital subtraction angiography (DSA) for depiction of central thromboembolic material and visualization of the pulmonary arterial tree. Functional analysis included calculation of left and right ventricular ejection fractions and peak velocities, net forward volumes per heartbeat, and blood volume per minute in the left and right pulmonary arteries and ascending aorta. Flow measurements were compared with invasively measured mean pulmonary arterial pressure (MPAP) and pulmonary vascular resistance (PVR) measurements. Nonparametric Wilcoxon and sign tests were used for statistical analysis.
RESULTS: MR angiography revealed typical findings of CTEPH (intraluminal webs and bands, vessel cutoffs, and organized central thromboemboli) in all patients. It depicted pulmonary vessels up to the segmental level in all cases. For subsegmental arteries, DSA revealed significantly more patent vessel segments than did MR angiography (733 versus 681 segments, P < .001). MR angiography revealed technical success of surgery in 33 of 34 patients. Patients had reduced right ventricular ejection fractions and pulmonary peak velocities that significantly increased after PTE (P < .001 for both). Right ventricular ejection fraction had good correlation with PVR (r = 0.6) and MPAP (r = 0.7). The postoperative decrease in MPAP correlated well with the increase in right ventricular ejection fraction (r = 0.8). Postoperatively, there was complete reduction of a preoperatively existing bronchosystemic shunt volume in 33 of 34 patients.
CONCLUSION: Breath-hold MR imaging techniques enable morphologic and semiquantitative functional assessment of patients with CTEPH.
© RSNA, 2004
Index terms: Digital subtraction angiography, 564.1243 • Hypertension, pulmonary, 564.783 • Magnetic resonance (MR), vascular studies • Pulmonary arteries, flow dynamics, 564.12144 • Pulmonary arteries, MR, 564.12142, 564.12144 • Pulmonary arteries, thrombosis, 564.813
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
This aberrant path of obstruction and reopening occurs in repeated cycles over many years. During the first years, patients may be asymptomatic before mean pulmonary arterial pressure (MPAP) and pulmonary vascular resistance (PVR) in the pulmonary arteries rise above 25 mm Hg or 130 dynes · sec · cm–5, respectively. These critical values are reached after occlusion of approximately 60% of the total diameter of the pulmonary arterial vascular bed. The longstanding increase in MPAP to greater than 30 mm Hg results in cor pulmonale and ends in right-sided heart failure, which has a 5-year survival rate of only 30% (1,3).
The primary treatment for CTEPH is surgical pulmonary thromboendarterectomy (PTE), which leads to a permanent improvement in pulmonary hemodynamics (4–9). The technical feasibility and success of the surgical procedure mainly depend on the localization of the thromboembolic material: Surgical accessibility is possible if the organized thrombi are not located distal to the lobar arteries or to the origin of the segmental vessels and a safe dissection plane for endarterectomy can therefore be developed (1,3–5).
The combination of right ventricle catheterization and selective pulmonary digital subtraction angiography (DSA) is still regarded as the standard of reference with respect to establishing the diagnosis, assessing the severity of the disease, and determining the technical feasibility of surgery for CTEPH (2,3,8,10). However, cross-sectional imaging modalities such as single- or multisection helical computed tomography (CT) are becoming increasingly important because they display the extent of the central thromboembolic material. Although CT is highly sensitive for detecting proximal and subsegmental thrombi (11–13), it fails to yield sufficient quantitative information about the severity of functional impairment (11–15).
Magnetic resonance (MR) imaging seems to be a promising modality for overcoming the restrictions of CT and DSA. Thus, the purpose of this study was to evaluate the potential of breath-hold MR imaging techniques in the morphologic and functional assessment of patients with CTEPH before and after surgery.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
Selective DSA and Right Ventricle Catheterization
Selective bilateral DSA was performed before surgery by only one examiner (M.B.P., with 7 years of experience in DSA), who used an Integris 3000 imager (Philips, Eindhoven, the Netherlands). A 7-F pigtail catheter (Supertorque Plus; Cordis, Miami, Fla) was selectively placed in the main right or left pulmonary arteries with electrocardiographic surveillance by using a transfemoral or transcubital approach (10,16). DSA was performed during deep inspiration, and three projections were obtained per side in the posteroanterior, 30° anterior oblique, and lateral projections; the frame rate was 12 per second. The contrast agent bolus consisted of 25 mL of the nonionic contrast medium iopamidol (Solutrast 300; Altanapharma, Konstanz, Germany) administered with a flow rate of 12.5 mL/sec. The total contrast agent amount was about 175 mL.
Preoperatively, right ventricle catheterization was performed after DSA in the angiographic suite by using a 7.5-F Swan-Ganz catheter (Edwards Lifesciences, Irvine, Calif) that was placed in the main pulmonary artery (M.B.P.) for determination of MPAP and PVR. Postoperative values of MPAP and PVR were obtained in the intensive care unit 2–3 days after surgery by two experienced examiners (E.M., T.K.).
MR Imaging
MR imaging was performed with a four-element body phased-array coil and a standard 1.5-T whole-body imager (Magnetom Vision; Siemens Medical Systems, Erlangen, Germany) that was equipped with 25 mT/m gradients that enabled a minimal rise time of 300 µsec. All examinations were performed during an inspiratory breath hold. An oxygen supply with a flow of 4 L/min through a nasal tube was required during imaging for augmenting the breath-hold capacity in six patients who had marked dyspnea in resting conditions.
MR Angiography
For MR angiography, we used a three-dimensional (3D) radiofrequency spoiled fast low-angle shot sequence with the following parameters: repetition time msec/echo time msec, 4.6/1.8; flip angle, 30°; 6/8–7/8 rectangular field of view, 420–480 mm; image matrix, 200–235 x 512 pixels; and receiver bandwidth, 488 Hz per pixel. Cardiac gating was not used because of the high number of phase-encoding steps. The field of view was varied between 420 and 480 mm, according to the body shape of each patient, to avoid wraparound artifacts as the patients’ arms were placed beside the body. An 80% partial Fourier technique was applied in the phase-encoding direction, and the k-space sampling was performed linearly and sequentially—starting on the shorter side of k-space—to allow the central lines of k-space to be sampled after 37.5% of the acquisition time. The 3D volume had a mean thickness of 112 mm and was subdivided into 32 partitions. Owing to an asymmetric k-space data sampling from –
to +
, zero-filled from –1 to +1 in the section-selection direction, the resulting effective section thickness was 2.7 mm. In-plane resolution ranged between 1.50 x 0.78 and 1.78 x 0.94 mm2. We used a coronal plane for the 3D slab, which was prescribed on the basis of a transverse localizing image that showed the lower lobe arteries that were included in the 3D volume. Acquisition time was either 23 or 27 seconds, depending on the number of phase-encoding steps.
Each data acquisition was enhanced by administering 20 mL of gadopentetate dimeglumine (Magnevist; Schering, Berlin, Germany) independently from patients’ body weight. The contrast agent was injected through a 22-gauge needle into an antecubital vein at a constant flow rate of 2 mL/sec by using an MR imaging–compatible power injector (Spectris; Medrad, Pittsburgh, Pa). Contrast agent administration was immediately followed by flushing with 20 mL of saline solution with the same flow rate. The mean patient weight was 71.8 kg ± 11.5 (range, 49–98 kg), which resulted in a mean gadopentetate dimeglumine dose of 0.14 mmol per kilogram of body weight ± 0.02 (range, 0.1–0.2 mmol/kg).
For timing the delay between contrast agent administration and the start of the acquisition, we used the test bolus method described elsewhere (17).
Cine MR Imaging
Short-axis cine MR acquisitions were planned to be perpendicular to the interventricular septum as it was depicted on horizontal and vertical long-axis views. Both ventricles were entirely imaged from the base to the apex in at least seven consecutive double-oblique short-axis sections by using a segmented electrocardiographically gated two-dimensional fast low-angle shot sequence with echo sharing (echo time, 4.8 msec; temporal resolution, 50 msec; section thickness, 8 mm; flip angle, 20°; lines per heartbeat, nine; matrix size, 162 x 256; field of view, 350 x 350 mm; gap, 2 mm). At least 13 frames, covering the first 600 msec of the cardiac cycle, were acquired.
Phase-Contrast MR Imaging
For flow measurements, an electrocardiographically gated velocity-encoded k-space segmented gradient-echo sequence (fast low-angle shot; echo time, 4.8 msec; flip angle, 30°; section thickness, 6 mm; lines per heartbeat, five) with a temporal resolution of 110 msec was used. The image matrix was 100 x 256 pixels. With a field of view of 300 mm, the in-plane resolution was 2.63 x 1.37 mm2. Owing to the use of k-space segmentation with a segmentation factor of five and 100 phase-encoding steps, 20 R-R intervals were required for complete sampling of k-space data. Flow measurements were planned perpendicularly in a double-oblique plane in the ascending aorta and in the right and left pulmonary arteries at a point closely after the bifurcation of the main pulmonary artery. There were no stenoses caused by the underlying disease in the pulmonary arteries at the level of the flow measurements in any patient. The velocity-encoding range was set to 250 cm/sec for the aorta and to 150 cm/sec for the right and left pulmonary arteries.
Image Analysis
MR angiography.—Two observers (K.F.J.K., S.L.) reviewed the contrast material–enhanced MR angiographic data sets on a commercially available workstation (Siemens Medical Systems) in consensus and without knowledge of the patients’ identities or the results of selective pulmonary DSA. They had 6 and 3 years of experience in reading MR angiograms, respectively. They reviewed source images, interactively created multiplanar reformations, and a series of seven maximum intensity projections that were generated at 15° rotational increments from a left anterior oblique view through a frontal view to a right anterior oblique view.
Overall image quality was scored by using a scale in which a score of 1 represented excellent; a score of 2, good; a score of 3, fair; a score of 4, poor; and a score of 5, nondiagnostic. An image quality rating was determined by averaging the rating for the enhancement of the pulmonary arteries (1 = excellent, 2 = good, 3 = fair, and 4 = poor); the rating for the presence of motion, susceptibility, and/or wraparound artifacts (1 = none, 2 = mild, 3 = moderate, 4 = severe); and the rating for the enhancement of pulmonary veins and the superior vena cava (1 = no enhancement, 2 = mild enhancement, 3 = moderate enhancement, 4 = greater enhancement than pulmonary arteries).
Images were further analyzed for the presence of organized adherent thrombotic material in the main pulmonary artery and the right, left, lobar, and segmental arteries; abrupt vessel cutoffs; intraluminal webs and bands; and abnormal proximal-to-distal tapering of the vessels. This analysis was restricted to the level of the segmental arteries. At the same time, the pulmonary vasculature was separated into different vessel segments, and the number of patent vessel segments was counted up to the subsegmental level. Finally, the most proximal beginning of the central thromboembolic material was evaluated and compared with the most proximal beginning of this material on DSA images and at surgery (as the definitive standard of reference).
Analysis of postoperative MR angiograms consisted of the detection of reopened segmental and subsegmental vessels and an overall assessment of the restoration of pulmonary vasculature (ie, whether the pulmonary vasculature showed complete normalization, partial normalization, or no change). The diameters of the right and left pulmonary arteries were determined pre- and postoperatively.
Functional imaging.—Volumetry was performed with short-axis cine MR images to determine the right and left ventricular ejection fractions. For contour tracing and for evaluation of the end-systolic and end-diastolic short-axis views (performed by one skilled observer [S.L.]), an evaluation program (ARGUS V2.3; Siemens Medical Systems) on an independent satellite console was used. Window and level settings were optimized for best image contrast between the myocardium and the ventricular lumen on a representative midventricular image and were then applied to all images. If necessary, the window and level settings were optimized for individual images. The end-diastolic images were always the images obtained immediately after the R wave. The left ventricular and right ventricular end-systolic images were selected as the images on which the areas of the respective cardiac chambers were the smallest. Ejection fraction was calculated as stroke volume divided by end-diastolic volume and was automatically expressed as a percentage. The motion of the interventricular septum (ie, whether it was normal or paradoxical) was analyzed.
The calculation of flow was performed by using the same software used for the analysis of cine MR images. The vessels were segmented manually by the same observer (S.L.) by drawing a region of interest on all magnitude images. The region of interest varied between 2.4 and 2.7 cm in diameter.
The software automatically measured the flow in the appropriate areas on the velocity-encoded images and yielded the net forward volume (in milliliters) per heartbeat (R-R interval) after subtraction of the reverse volume (in milliliters per R-R interval) from the forward volume (in milliliters per R-R interval). The blood volume (in milliliters) per minute was calculated by multiplying the heart rate by the net forward volume per heartbeat. The maximum peak velocity (in centimeters per second) of the flowing blood was displayed automatically. The above measurements were performed for the left and right pulmonary arteries and the ascending aorta. The total pulmonary arterial flow was calculated by adding the flow volumes of the left and the right pulmonary arteries. For all calculations, a phase background correction was applied. Flow measurements were compared with invasively measured MPAP and PVR measurements.
Statistical Analysis
Parameters are reported as means ± SDs. The nonparametric Wilcoxon test (ie, the Mann-Whitney U test) was used to evaluate the correlation between pre- and postsurgical hemodynamic data. The nonparametric sign test was used in the evaluation of paired values (eg, pulmonary arterial diameter). The Pearson correlation coefficient was used for correlation analysis. We regarded the number of 35 patients as sufficient for primary analysis. A power analysis was performed because differences between DSA and MR angiography with respect to the rate of depiction of patent subsegmental arteries were expected. Before the start of the experiment, DSA was expected to depict a mean of 1.5 more subsegmental arteries than MR angiography, with an SD of at least 2. If a statistical power of 80% is assumed, the minimum number of patients required for intraindividual comparison on the basis of a t test would be 16.
Data were collected with Excel 9.0 (Microsoft, Redmond, Wash). Statistical analyses were performed with the SPSS-PC software, version 9.0 (SPSS, Chicago, Ill). P 
.05 was considered to indicate a statistically significant difference on a local level.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
Preoperative MR angiograms revealed typical findings for CTEPH in all patients: 162 vessel segments contained thromboembolic material that was adherent to the pulmonary arterial wall. Intraluminal webs and bands were seen in 180 vessel segments, and abnormal proximal-to-distal tapering and abrupt vessel cutoffs were seen in 221 vessels. A thorough analysis of source images and the creation of multiplanar reformations were most important for exact assessment of morphologic findings, while maximum intensity projections provided an overview and an impression of the arterial vascular tree that were comparable to those provided by the DSA images (Figs 1, 2).
|
|
|
|
|
|
|
Postoperatively, unchanged pulmonary arterial vasculature was observed in one (3%) of the 34 patients; this correlated with unchanged pulmonary hemodynamics (MPAP and PVR). The remaining 33 patients had 144 reopened segmental arteries (mean, four reopened segmental arteries; range, 1–13 reopened segmental arteries). Complete normalization of pulmonary arterial vasculature was not documented in any case (Figs 2b, 2c). There was a significant reduction in the diameter of the main pulmonary arteries: The mean diameter of the right pulmonary artery decreased from 2.7 cm ± 0.4 preoperatively to 2.4 cm ± 0.4 postoperatively (P < .001, sign test). There was also a reduction in the mean diameter of the left pulmonary artery—from 2.7 cm ± 0.3 preoperatively to 2.5 cm ± 0.4 postoperatively (P = .001, sign test).
Functional Imaging
Cine MR measurements could be evaluated in 31 of 34 cases. Limited breath-holding capabilities in two patients and electrocardiographic misregistration in one patient precluded further evaluation. One patient could not hold his breath long enough to enable us to perform flow measurements; thus, data from this examination were excluded from the preoperative evaluation. Preoperatively, the MPAP was available for comparison in all patients except one in whom the value was not documented. Nine of 34 postoperative cine MR data sets contained motion or electrocardiographic misregistration artifacts. All flow measurements were eligible for evaluation. PVR values were available in 26 patients, and MPAP values were available in 28 patients.
Hemodynamic values observed at MR imaging before and after PTE, and the statistical significance of the changes in them, are listed in Table 1. The patients had markedly reduced right ventricular ejection fractions that returned to normal values postoperatively. The net forward volume in the pulmonary artery was significantly (P < .001) lower than that in the aorta and increased significantly after PTE. Preoperatively, the net forward volume was nearly equally distributed between the right and left pulmonary arteries. The postoperative increase in total pulmonary arterial net forward volume was mainly due to the increase in net forward volume in the right pulmonary artery—net forward volume in the left pulmonary artery was almost unchanged. Before PTE, the maximum peak velocities in the pulmonary arteries were markedly low. After surgery, they increased significantly without reaching a normal range. The interventricular septum showed paradoxical movement in all patients preoperatively. After surgery, the movement of the septum returned to normal in 19 (68%) of the 28 patients who could be evaluated for this finding (Fig 3).
|
|
|
The invasively determined values of MPAP and PVR are listed in Table 2; they decreased significantly after surgery.
|
|
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
Morphologic Assessment
The results of the present study suggest that an optimized contrast-enhanced MR angiographic technique (18–21) can enable diagnostically sufficient image quality in all patients and can delineate typical angiographic findings such as intraluminal webs and bands, abrupt vessel cutoffs, and abnormal proximal-to-distal tapering. Because it is a cross-sectional imaging modality, contrast-enhanced MR angiography proved to be superior even to selective DSA for assessment and delineation of the proximal extent of organized thrombotic material. Contrast-enhanced MR angiography depicted all patent vessel segments down to the level of the segmental arteries and depicted 93% of patent subsegmental arteries.
The incomplete delineation of subsegmental arteries at MR angiography was caused, in most cases, by incomplete coverage of the lung volume. Additionally, the lack of electrocardiographic gating hampered the evaluation of vessels in close vicinity to the heart, especially those of the lingulae. Breathing artifacts, which could be expected to occur owing to the use of acquisition times of 23 or 27 seconds, did not play a substantial role in overall image quality in the present study. One may speculate as to whether the high motivation of the patients was an important factor in our study. However, an additional oxygen supply should be provided in patients with severe dyspnea.
The results of postoperative contrast-enhanced MR angiography in our study show that complete normalization of the pulmonary vasculature is not achieved with bilateral PTE. However, with improvement of hemodynamics, a significant reduction in pulmonary arterial diameter can be observed.
To our knowledge, there are only a few studies in which MR angiographic techniques have been evaluated for the assessment of patients with CTEPH. In a study involving 34 patients, sensitivities and specificities, respectively, for detection of CTEPH were 83% and 95% for the lobar level and 72% and 94% for the segmental level (22). The same group of investigators evaluated MR angiography for the differentiation of patients with primary pulmonary hypertension from patients with CTEPH and from healthy subjects. These investigators could correctly identify all healthy subjects, and good differentiation of primary pulmonary hypertension from CTEPH was achieved (accuracy was 92% [49 of 53 patients and healthy subjects] for reader 1 and 91% [48 of 53] for reader 2) (23). In another study by this group (24), MR angiography proved to be inferior to helical CT for the preoperative assessment of CTEPH when surgical and DSA findings were used as the standard of reference. The accuracies of MR imaging ranged between 39% and 46% for central vessels and between 57% and 73% for lobar and segmental vessels (24). However, in all of these studies, contrast-enhanced two-dimensional gradient-echo fast low-angle shot sequences were used for MR angiography, and these sequences can be regarded as being out-of-date.
Three clinical studies (25–27) (all of which were prospective and blinded) have involved evaluation of the accuracy of contrast-enhanced MR angiography for detecting acute pulmonary embolism. The results of these studies showed that the technique enabled detection of acute pulmonary embolism on a patient-by-patient basis, with sensitivities and specificities, respectively, ranging from 77% (27 of 35 patients) to 100% (eight of eight patients) and from 95% (21 of 22 patients) to 98% (81 of 83 patients) (17,24,26). The interobserver agreement was substantial, with 
values between 0.75 and 0.60. However, at a segment-by-segment analysis that included the subsegmental arteries, the sensitivities decreased to values between 40% and 75%, respectively (26,27). These values are considerably lower than those obtained at spiral CT angiography of the pulmonary circulation; the diagnostic accuracy and volume coverage of spiral CT angiography has been decisively improved by the introduction of multi–detector row technology (14,28).
Given the results of our study, which showed that contrast-enhanced MR angiography had a sensitivity of 93% for the depiction of patent subsegmental arteries and a sensitivity of 100% down to the level of the segmental arteries, we suggest that selective DSA be reserved for use in selected situations—for example, when diagnostic difficulties arise in patients who have obliterations that are located distal to the segmental level. These patients could be candidates for balloon angioplasty if lung transplantation is not a therapeutic option (29).
Functional Assessment
Our results, which showed reduced right ventricular ejection fractions without impairment of left ventricular function by abnormal interventricular septal movement, are in concordance with the results of studies involving echocardiography in patients with CTEPH (30–34). Although there is still some discussion about the accuracy of echocardiography, cine MR imaging is a well-acknowledged method for determination of right ventricular ejection fraction (30,32). A basic condition for cine MR imaging is that a sufficient temporal resolution (eg, 50 msec or less) be used so that end-systolic volumes are not overestimated and ejection fractions are not therefore underestimated (35).
For MR imaging flow measurements, we used established sequences (31,36,37) with sufficient numbers of voxels inside the vessel of interest and sufficient numbers of heart phases per cardiac cycle (38,39). With respect to maximum peak velocities, a lower velocity-encoding range than 150 cm/sec seems to be preferable for further studies. However, as shown by Figure 4, peak velocities may exceed 100 cm/sec, so an individual adjustment of the velocity-encoding range may be necessary.
The equal distribution of blood flow between the right and left lungs changed dramatically after surgery, with 62.8% of total pulmonary blood flow going to the right side. By correlating this finding with postoperative contrast-enhanced MR angiographic results, we found that two times more segments were reopened on the right side than on the left side. This high degree of difference can be explained by the nature of PTE, at which thromboendarterectomy is first performed on the right side and then performed on the left side.
The results of our study showed a significant difference between the blood flow through the pulmonary arteries and the blood flow through the ascending aorta in patients with CTEPH. This flow difference can be explained by the bronchosystemic shunt volume that is caused by dilatation of bronchial arteries that originate from the aorta and supply the lung parenchyma (40,41). In a previous study by Ley et al (41), there was a significant correlation between the cross-sectional area of the bronchial arteries as determined at helical CT and the shunt volume between the systemic arterial and pulmonary venous circulations as determined with phase-contrast MR imaging flow measurements. After surgery, there was a complete resolution of this bronchosystemic shunt volume.
Analogous to our results, the results of a study by Kondo et al (36) showed that increased resistance in the pulmonary circulation occurs along with decreases in flow velocity, blood flow through the pulmonary arteries, and right ventricular ejection fraction.
However, probably because of the low temporal resolution of 110 msec, the velocity-encoding technique that we used did not enable sufficiently direct estimation of MPAP or PVR compared with invasive measurements. Echo-view sharing that combines k-space data from two cardiac phases for creation of a virtual phase—leading to better temporal resolution—was not available with our MR imaging unit. Thus, there was only a fair to good linear correlation between the decrease in maximum peak velocity in the pulmonary arteries and MPAP (r = 0.6) and between the decrease in the right ventricular ejection fraction and MPAP (r = 0.7). However, comparison of the difference between the pre- and postoperative right ventricular ejection fractions and the difference between the pre- and postoperative MPAPs revealed a better correlation (r = 0.8), indicating that the decrease in MPAP can be estimated accurately with our methods.
Limitations
We acknowledge several limitations of our study: First, there was a strong selection bias, in that all patients included in the study were potential candidates for surgery. Thus, no further statement is possible with respect to the specificity of our techniques—especially contrast-enhanced MR angiography—or to the ability of our techniques to enable the differentiation of different causes of pulmonary hypertension.
Second, the incomplete coverage of the lung volume with the 3D slab used at contrast-enhanced MR angiography and the lack of electrocardiographic triggering prevented the depiction of all patent subsegmental arteries and the display and evaluation of subsegmental vessels in close vicinity to the heart. Some investigators have therefore proposed the acquisition of two separate volumes in the sagittal plane—one for each lung (27,42). The advantage of sagittal imaging lies in the reduced imaging volume, which permits shortening of the data acquisition time. The latter can be further reduced by implementing parallel imaging techniques (43). However, the acquisition of two 3D data sets requires the twofold application of contrast agents (ie, one contrast agent would have to be administered twice).
Third, all measurements were performed during a deep inspiratory breath hold, so our absolute numbers for blood flow are probably lower than they would have been in freely breathing patients (44).
Fourth, with a temporal resolution of 110 msec at velocity-encoded MR imaging, detection of flow curve alterations caused by increased MPAP is not possible. Mousseaux et al (45), by using a temporal resolution of 24 msec, demonstrated that in patients with increased resistancies of the pulmonary vasculature, the maximum peak velocity was reached earlier than it was in healthy volunteers. Abolmaali et al (46) reported good correlations between absolute and relative acceleration times, as well as between peak flow velocities in the pulmonary trunk and MPAP, in an animal model. Their velocity-encoding sequence had a temporal resolution of less than 10 msec.
Finally, we did not derive MPAP from measurements of peak velocity in the regurgitant flow jet in the accompanying tricuspid insufficiency. For correct assessment of MPAP, the central venous pressure that can be estimated on the basis of the width of the proximal inferior vena cava has to be added.
In summary, breath-hold MR imaging techniques enable noninvasive morphologic and partial functional assessment of patients with CTEPH. In our study, contrast-enhanced MR angiography depicted characteristic angiographic findings in all patients, and its results showed very good correlation with those of selective DSA. Postoperatively, contrast-enhanced MR angiography enables documentation of the technical success of PTE. Cine and phase-contrast MR imaging examinations enable sufficient characterization of the impairment of function in the right side of the heart. Blood flow ratios between the right and left pulmonary arteries seem helpful for indicating the side most affected by the thromboembolic disease. However, MR imaging cannot actually replace invasive preoperative determination of PVR and MPAP. It may play an important role in postsurgical follow-up and may serve as the noninvasive modality of choice for documentation of morphologic and functional improvement after surgery.
|
FOOTNOTES |
|---|
Abbreviations: CTEPH = chronic thromboembolic pulmonary hypertension, DSA = digital subtraction angiography, MPAP = mean pulmonary arterial pressure, PTE = pulmonary thromboendarterectomy, PVR = pulmonary vascular resistance, 3D = three-dimensional
Author contributions: Guarantors of integrity of entire study, K.F.J.K., S.L., M.T.; study concepts, K.F.J.K., S.L., H.U.K., E.M.; study design, K.F.J.K., S.L.; literature research, K.F.J.K., S.L., H.U.K., E.M., F.K.; clinical studies, K.F.J.K., T.K., M.B.P., E.M.; data acquisition, K.F.J.K., T.K., M.B.P., H.U.K., E.M.; data analysis/interpretation, K.F.J.K., S.L.; statistical analysis, S.L., F.K.; manuscript preparation, K.F.J.K., S.L.; manuscript definition of intellectual content, K.F.J.K., S.L., H.U.K., E.M., F.K.; manuscript editing, K.F.J.K., S.L., M.T.; manuscript revision/review, S.L., E.M., H.U.K., F.K., M.T.; manuscript final version approval, all authors
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  H. J. Reesink, I. I. Tulevski, J. T. Marcus, F. Boomsma, J. J. Kloek, A. Vonk Noordegraaf, and P. Bresser Brain Natriuretic Peptide as Noninvasive Marker of the Severity of Right Ventricular Dysfunction in Chronic Thromboembolic Pulmonary Hypertension Ann. Thorac. Surg., August 1, 2007; 84(2): 537 - 543. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Sanz, P. Kuschnir, T. Rius, R. Salguero, R. Sulica, A. J. Einstein, S. Dellegrottaglie, V. Fuster, S. Rajagopalan, and M. Poon Pulmonary Arterial Hypertension: Noninvasive Detection with Phase-Contrast MR Imaging Radiology, April 1, 2007; 243(1): 70 - 79. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
F. D. Rubens, M. Bourke, M. Hynes, D. Nicholson, M. Kotrec, M. Boodhwani, M. Ruel, C. J. Dennie, and T. Mesana Surgery for Chronic Thromboembolic Pulmonary Hypertension--Inclusive Experience From a National Referral Center Ann. Thorac. Surg., March 1, 2007; 83(3): 1075 - 1081. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. J. Reesink, J. T. Marcus, I. I. Tulevski, S. Jamieson, J. J. Kloek, A. V. Noordegraaf, and P. Bresser Reverse right ventricular remodeling after pulmonary endarterectomy in patients with chronic thromboembolic pulmonary hypertension: Utility of magnetic resonance imaging to demonstrate restoration of the right ventricle J. Thorac. Cardiovasc. Surg., January 1, 2007; 133(1): 58 - 64. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
