| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Ann Thorac Surg 2003;76:1576-1585
© 2003 The Society of Thoracic Surgeons
a Department of Radiology, Cleveland, OH, USA
b Department of Thoracic and Cardiovascular Surgery, Cleveland, OH, USA
c Department of Cardiovascular Medicine, Cleveland, OH, USA
d Department of Biomedical Engineering, Cleveland, OH, USA
e Department of Biostatistics and Epidemiology, The Cleveland Clinic Foundation, Cleveland, Ohio, USA
Accepted for publication May 28, 2003.
* Address reprint requests to Dr White, Section of Cardiovascular Imaging, Division of Radiology, Desk HB6, The Cleveland Clinic Foundation, 9500 Euclid Ave, Cleveland, OH 44195, USA.
e-mail: whiter{at}ccisd1.ccf.org
 |
Abstract |
|---|
 Top Abstract IntroductionPatients and methodsResultsCommentReferences |
|---|
METHODS: Twenty-four dilated cardiomyopathy patients underwent cardiac magnetic resonance imaging (MRI) before PLV, and 3 and 12 months after surgery. Left ventricular (LV) circumferential shortening and wall stress were computed at three short-axis levels. Exploratory outcome analysis grouped patients according to the timing of adverse cardiac events postsurgery.
RESULTS: LV mass and volume were decreased at each postsurgical time point (all p < 0.01). At 3 months, regional wall stress was reduced at all short-axis levels; but by 12 months stress was reduced from baseline only at the apex. Circumferential shortening was increased significantly at both postsurgical time points at each level. On average, septal shortening was negative (stretching) before surgery, but increased significantly, and was positive, postsurgery. Exploratory outcome analysis found that negative values of basal septum circumferential shortening before surgery increased the probability of event-free survival beyond 6 months.
CONCLUSIONS: Regional heterogeneity of LV myocardial function, associated with dilated cardiomyopathy, was diminished after PLV but was also related to patient outcome. MRI with tissue tagging is useful for assessing the efficacy of surgical therapies for congestive heart failure.
|
Introduction |
|---|
|
TopAbstractIntroductionPatients and methodsResultsCommentReferences |
|---|
Partial left ventriculectomy (PLV) was pioneered by Batista and coworkers [7] as an alternative or bridge to transplantation that involves excising a portion of the left ventricle (LV) to reduce wall tension and normalize geometry. However, clinical results with this technique were variable with few a priori predictors of outcome [8–11]; and PLV is no longer considered a viable therapeutic option.
Despite this, surgical modification of the LV has remained popular [12]. Furthermore, although previous studies have reported the effects of surgical modification by PLV on clinical indices of LV performance [8, 13, 14], few have examined the functional consequences of PLV on LV myocardial mechanics in patients [2].
We hypothesized that presurgical measures of LV morphology and function could predict outcome following PLV. We used magnetic resonance imaging (MRI) [15] to characterize LV anatomy and performance in dilated cardiomyopathy patients before PLV and, in an exploratory analysis, related these results to LV response and clinical outcome following surgery. This effort included analysis of the effects of PLV on LV myocardial mechanics up to 1 year after surgery.
|
Patients and methods |
|---|
|
TopAbstractIntroductionPatients and methodsResultsCommentReferences |
|---|
Patients underwent repeat MRI at approximately 3 months (denoted PLV3, n = 9, 97 ± 22 days) and 1 year (denoted PLV12, n = 6, 374 ± 49 days) after surgery.
Fourteen patients did not undergo follow-up MRI, for either clinical reasons, including death or placement of a left ventricular assist device (n = 8); prohibitive travel distance (n = 5); or new contraindications such as permanent pacemaker (n = 1). Of the 10 patients with a postsurgical MRI, 4 were imaged only at PLV3, 1 was imaged only at PLV12 and 5 were imaged at both. Only 1 patient was receiving inotropic therapy at PLV3, none at PLV12.
All data obtained for this study were reviewed and approved by the local Institutional Review Board. MRI examinations were clinically indicated and conducted using an Institutional Review Board approved protocol with waiver of individual consent.
During the same period, 36 patients, approved for transplantation and eventually treated by PLV, failed to undergo presurgical MRI due to hemodynamic instability, anxiety, claustrophobia, or MRI contraindications. No differences existed between the MRI and non-MRI groups in age, LV end-diastolic midcavity diameter, LV ejection fraction (EF), right ventricular EF or mitral regurgitation severity. However, the non-MRI group contained 33% NYHA functional class III and 67% NYHA functional class IV patients, which differed significantly from the MRI group (p = 0.03).
Surgical procedure
Partial left ventriculectomy has been previously described [13]. Briefly, a wedge-shaped section of the LV free wall, between the papillary muscles and supplied by the circumflex artery, was excised in all patients. This was accompanied by mitral valve repair in 23 patients, DeVega tricuspid valve annuloplasty in 11 patients [16], resection and resuspension of papillary muscles in 11 patients, coronary bypass grafting in 2 patients, and mitral valve replacement in 1 patient.
Magnetic resonance imaging techniques
Imaging was performed using a 1.5 Tesla MRI scanner (Magnetom Vision, phased-array body coil; Siemens Medical Solutions, Erlangen, Germany), during which heart rate and upper extremity blood pressure were recorded (Dynamap; Critikon, Tampa, FL).
Dynamic cine image loops were acquired along the cardiac short-axis in contiguous slices from the mitral valve to LV apex, using an electrocardiogram (ECG) triggered, segmented k-space, gradient-echo protocol (FLASH, TE 4.8 msec, TR 100 msec, flip angle 20 degrees, slice thickness 8 to 10 mm, field of view 300 to 360 mm, 6/8 to 8/8 rectangular matrix starting with 256x256 lines). Temporal resolution was 50 msec with view sharing [17]. For patients capable of repetitive 6 to 10 second breath-holds, 1 signal average was used. Otherwise, three averages were used during free breathing. Horizontal long-axis, vertical long-axis, and LV outflow tract views were also acquired [15].
Dynamic tagged image loops were acquired using an ECG-triggered, segmented k-space, grid-tagged gradient-echo protocol (SPAMM, 8 mm tag spacing) [18]. Image locations and measurements were identical to the FLASH sequence, except TE 4 msec, TR 90 msec, flip angle 15 degrees, temporal resolution 45 msec.
Dynamic phase-contrast imaging was used to measure flow through the ascending aorta (TE 5 msec, TR 24 msec, flip angle 30 degrees, slice thickness 6 mm, field of view 300 to 350 mm, 6/8 to 8/8 rectangular matrix starting with 128x256 lines, and three signal averages) [19].
Data analysis
Image analysis was performed using commercially available software (Argus v2.0; Siemens Corporate Research, Princeton, NJ). Three representative LV short-axis levels were identified for analysis from each patient. Midventricle was defined using image loops in which papillary muscle insertion points were visualized throughout the cardiac cycle. The base was defined by image loops nearest the mitral annulus in which no valve plane or membranous septum were visualized. The apex was defined midway between midventricle and apex.
Cine image analysis
In short-axis cine images, LV endocardial and epicardial contours were manually traced for determination of LV mass, end-diastolic volume (EDV), end-systolic volume (ESV), stroke volume, and EF. To assess the adequacy of physiologic LV hypertrophy, the ratio of EDV to LV mass (V/M) was computed [20].
Global LV circumferential wall stress (WSG) was calculated [21] as
 |
Regional circumferential WS was computed [22] as
 |
The rate-corrected velocity of circumferential fiber shortening (Vcfc) was computed as
 |
D is percent fractional shortening computed at midventricle [23], LVET is ejection time estimated from MRI image loops, and RR is R-R interval.
Tagged image analysis
Left ventricular epicardial and endocardial contours were manually traced at end-diastole. Tag-line intersections within the myocardium were identified in systolic frames at each short-axis level. The grid overlay in each frame was reduced to triangular tissue elements, and the centroid of each relative to the slice center of mass was determined.
Circumferential line segments connecting midwall tissue element centroids were manually identified [24], with circumferential defined as perpendicular to lines connecting the tissue element centroid and slice center of mass. Circumferential shortening (%S) for each segment was defined as end-diastolic minus end-systolic length divided by end-diastolic length. Slice %S was determined by averaging line segment shortening at each short-axis level (%SBASE-ALL, %SMID-ALL, and %SAPEX-ALL). For regional analysis, the LV was divided into regions identical to those identified above for WS calculation.
Baseline tagged images were analyzed in 21 patients, with three sets not analyzed because either the examination was ended at the patient's request (n = 2), or inadequate image quality because of poor respiratory compensation (n = 1). PLV3 tagged images were analyzed in all 9 patients, but apical images were inadequate in 1 patient. PLV12 tagged images were analyzed in 5 of 6 patients.
Phase-contrast image analysis
The ascending aorta was manually delineated in systolic phase-contrast images to calculate net flow from the LV (effective stroke volume). Effective EF (effective stroke volume divided by EDV) and mitral regurgitant fraction (the difference between stroke volume and effective stroke volume divided by stroke volume) were also computed.
Statistical analysis
A univariate exploratory analysis of all measured variables (Table 1)
was performed with two types of dependent variables: (1) time until PLV-failure event (defined as heart failure related death or relisting for transplantation), and (2) a categorization of outcome as either PLV-failure event occurrence within 6 months after PLV, or no event within 6 months. Cox proportional hazards regression was used to test whether each potential predictor was related to the time until event. Logistic regression was used to test whether each potential predictor was related to outcome group.
|
p < 0.10 was marginally significant). |
Results |
|---|
|
TopAbstractIntroductionPatients and methodsResultsCommentReferences |
|---|
|
The average myocardium removed during surgery was 96 ± 51 g, representing 23% ± 12% of LV mass.
Post-PLV LV morphology and function
Early post-PLV phase
At PLV3, LV mass and EDV were significantly reduced from baseline, as were heart rate and V/M (Table 3).
LV function (EF, Vcfc) was significantly improved; effective EF, mitral regurgitant fraction and cardiac index were unchanged.
|
|
|
|
|
|
|
Changes in WS and %S between baseline and PLV3 correlated weakly at the base (r = -0.25, p = 0.51) and midventricle (r = -0.31, p = 0.41), but more strongly at the apex (r = -0.71, p = 0.05). At PLV12, changes in WS and %S were virtually uncorrelated at the base (r = -0.12, p = 0.84) and midventricle (r = -0.04, p = 0.95), and were only moderately correlated at the apex (r = -0.54, p = 0.35).
Outcome analysis
Follow-up findings
Fifteen of 24 patients (63%) had an adverse cardiac-related event during follow-up. Mean time to event was 370 days (range 3 to 972). Events included 4 deaths and 11 patients relisted for transplantation. In the remaining 9 patients with no cardiac events, the mean follow-up time was 973 days (range 365 to 1642).
Measurement versus time to event analysis
In the univariate analysis of time until event, only WSMID-LATERAL was statistically significant (p = 0.02), with greater baseline WSMID-LATERAL in those patients with a cardiac-related event within 6 weeks post-PLV (p = 0.001, Table 5).
Three variables were marginally significant: body surface area (BSA, p = 0.09), %SBASE-SEPTUM (p = 0.07), and %SBASE-INFERIOR (p = 0.08).
|
Seven of 24 patients (29%) had a cardiac-related event within the first 6 months. Based on this classification, only %SBASE-SEPTUM was a statistically significant predictor (p = 0.03), indicating that negative baseline %SBASE-SEPTUM was associated with event-free survival beyond 6 months post-PLV. %SBASE-SEPTUM was -2.1% in patients with no event within 6 months, but was + 4.1% in patients with an event. Also, baseline NYHA functional class was marginally significant (p = 0.07); 59% of patients with no event within 6 months were NYHA class III (the remaining class IV) before surgery, compared to 29% of patients with an event within 6 months.
|
Comment |
|---|
|
TopAbstractIntroductionPatients and methodsResultsCommentReferences |
|---|
Baseline LV function in dilated cardiomyopathy
Marked regional heterogeneity of LV myocardial function in dilated cardiomyopathy has been shown previously [2–5]. Using a subset of patients from the current study, Young and colleagues [2] have demonstrated significant regional variations in circumferential shortening in dilated cardiomyopathy, including septal stretching. They also reported regionally heterogeneous wall stress, but concluded that this alone could not account for regional variations in function.
Bach and coworkers [3] have reported heterogeneous LV function in dilated cardiomyopathy patients, with relatively increased lateral wall function, similar to the current study. They also revealed a direct relationship between regions of preserved LV function and higher regional oxidative metabolism, leading them to conclude that local loading conditions were not solely responsible for the LV functional heterogeneity.
Negative curvature of the septum can result from elevated right ventricular pressure [25] and might lead to a misinterpretation of septal contraction as lengthening. In the current study, right ventricular pressure measured during roentgenogram angiography was elevated on average: 49 ± 13/8 ± 4 mm Hg (systolic/diastolic). However, MRI did not illustrate reversal of septal curvature in any patient. Also, there was no correlation between RV pressure and circumferential shortening in the septum (r values from -0.11 to 0.13 depending on short-axis level). Therefore, we believe that septal stretching at baseline was independent of RV effects.
Left ventricular function after PLV
Due to the nature of PLV, LV mass, EDV and ESV were significantly reduced following surgery; however, stroke volume and cardiac index did not improve.
Both Vcfc and EF improved post-PLV, indicating that in patients who responded well to the procedure (ie, cardiac-related event free for 1 year), sustained improvement of global LV function was possible. These results contradict a previous report in which cardiac index was improved at 18 months post-PLV [9], but had deteriorated by 24 months, indicating that improved LV function at 1 year post-PLV does not guarantee sustained improvement. However, our results agree with our institution's general experience at one year post-PLV [14], suggesting that in terms of LV function our patients did not deviate substantially from the population from which they were sampled.
Wall Stress–Circumferential shortening
Consistent with a fundamental purpose of the surgery, we have demonstrated significant reductions in WSG after PLV using a thin-walled ellipsoidal model [21]. However, regional WS was heterogeneous post-PLV, with WSBASE-ALL unchanged from baseline at PLV3, and both WSBASE-ALL and WSAPEX-ALL unchanged at PLV12.
Furthermore, only at the apex did a reduction in WS consistently lead to an increase in %S. At the base and midventricle, in contrast, decreases in WS were only weakly correlated with increases in %S at PLV3 and were virtually uncorrelated at PLV12.
Predictors of outcome after PLV
Previous studies have indicated that myocardial cell diameter [9], myocardial fiber diameter [10], infarction of both papillary muscles, and myocardial infarction at the surgical resection borders [11] are related to patient outcome after PLV.
However, this study has demonstrated a relationship between presurgical LV myocardial mechanics and outcome. In the current study, septal stretching was present before PLV at the base in 10 patients, at midventricle in 12 patients, and at the apex in 11 patients; while 8 patients demonstrated septal stretching at all levels. Three months after surgery, only 1 patient exhibited septal stretching, and did so at all short-axis levels. Twelve months after surgery, no patient exhibited septal stretching at any level.
We believe that recovery of septal function post-PLV is our most significant finding because it implies that the presence of viable, noncontracting myocardium in the septum before surgery might help compensate for the loss of contractile tissue in the lateral wall. However, hearts without this contractile reserve were unable to compensate for the loss of lateral wall myocardium.
Our finding that NYHA functional class was related to outcome with marginal significance was contradictory to previous studies [9, 10], and is probably related to our sample size.
Study limitations
The most notable limitation of the current study was the small number of patients studied post-PLV. However, many patients were unavailable for MRI after surgery. Post-PLV examinations were biased towards patients in relatively better health and with more favorable outcomes than PLV patients in general. Furthermore, in some patients portions of the LV wall were only a single tag width thick, adding uncertainty to tagging measurements.
The outcome analysis had a small sample size but many potential predictors. Therefore, it would be unwise to derive prediction rules for outcome based on these results. However, our results suggest that either WSMID-LATERAL or %SBASE-SEPTUM (ROC area 0.862) have the potential to predict outcome, and that %SBASE-INFERIOR may also be important. Additional studies would be needed before definite conclusions could be reached.
Lastly, in some patients we had difficulty defining the anterior/inferior wall junction after PLV, which along with the inclusion of scar tissue from this region in our analysis, might have negatively impacted our results.
Conclusions
Regional heterogeneity in systolic LV function associated with dilated cardiomyopathy is diminished post-PLV, but is related to patient outcome. Specifically, myocardial stretching in the basal septum is associated with better outcomes. Although it would be difficult to extend these findings to PLV candidates in general due to our small patient population, we believe these findings are significant because they suggest that noninvasive, quantitative evaluation of regional LV function by MRI, particularly with tissue tagging, is useful for predicting patient outcome following PLV. However, because PLV is no longer performed regularly, a more comprehensive study is not possible. It remains to be seen whether regional variations in LV myocardial mechanics will be predictive of outcome in patients undergoing newly developed surgical treatments for end-stage heart failure.
|
References |
|---|
|
TopAbstractIntroductionPatients and methodsResultsCommentReferences |
|---|
Related Article
This article has been cited by other articles:
|
|
 |
|
  M. J.W. Gotte, T. Germans, I. K. Russel, J. J.M. Zwanenburg, J. T. Marcus, A. C. van Rossum, and D. J. van Veldhuisen Myocardial Strain and Torsion Quantified by Cardiovascular Magnetic Resonance Tissue Tagging: Studies in Normal and Impaired Left Ventricular Function J. Am. Coll. Cardiol., November 21, 2006; 48(10): 2002 - 2011. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. D. Buckberg and the RESTORE Group Form versus disease: optimizing geometry during ventricular restoration Eur. J. Cardiothorac. Surg., April 1, 2006; 29(Suppl_1): S238 - S244. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. Ascione, P. Wilde, and G. D. Angelini Left ventricular volume reduction MMCTS, June 28, 2005; 2005(0628): 760. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ANN THORAC SURG | ASIAN CARDIOVASC THORAC ANN | EUR J CARDIOTHORAC SURG |
| J THORAC CARDIOVASC SURG | ICVTS | ALL CTSNet JOURNALS |
= baseline; 
= PLV3. (ANT = anterior; FREE = free wall; INF = inferior; LAT = lateral; PLV3 = partial left ventriculectomy after 3 months; SEP = septum; WS = wall stress.)
