Abstract
Heterogeneous perfusion in left bundle branch block (LBBB) has been demonstrated by 99mTc-methoxyisobutylisonitrile (MIBI) SPECT. Locally different contraction is also associated with LBBB. Quantitative analysis of myocardial SPECT is influenced by partial-volume effects depending on systolic wall thickening. Therefore, partial-volume effects may mimic perfusion heterogeneity in LBBB. Methods: Fifteen patients with nonischemic dilated cardiomyopathy and LBBB underwent resting 15O-water PET, 99mTc-MIBI SPECT, and gated 18F-FDG PET for analysis of wall thickening. Myocardial blood flow corrected for rate–pressure product (corrMBF), 99mTc-MIBI uptake, and wall thickening were determined in 4 left ventricular wall areas. In 14 patients, M-mode echocardiographic recordings were available for comparison. Results: Homogeneous distribution was found for corrMBF (1.09 ± 0.41 to 1.19 ± 0.31 mL × g−1 × min−1). 99mTc-MIBI uptake and wall thickening were heterogeneous (P < 0.0001), with the lowest values septal (99mTc-MIBI, 65% ± 10%; wall thickening, 16% ± 14%) and the highest lateral (99mTc-MIBI, 84% ± 5%; wall thickening, 55% ± 17%). Similar relationships in systolic wall thickening were observed by M-mode echocardiography (anteroseptal, 20% ± 11%; posterolateral, 37% ± 18%; P < 0.001). Conclusion: Heterogeneity of 99mTc-MIBI uptake in LBBB corresponds to differences in wall thickening and does not reflect distribution of corrMBF. Supplementary analysis of wall thickening is recommended when assessing 99mTc-MIBI SPECT in LBBB.
Studies using 201Tl or 99mTc-methoxyisobutylisonitrile (MIBI) with SPECT have demonstrated diminished septal perfusion in patients with complete left bundle branch block (LBBB) even in the absence of coronary artery disease (1,2). It has been postulated that a higher partial-volume effect and therefore a smaller count increase during systole because of reduced systolic wall thickening in the septum may mimic this hypoperfusion (3). PET with 15O-labeled water (15O-water) enables absolute quantification of myocardial blood flow (MBF) without being affected by the partial-volume effect through the introduction of perfusable tissue fraction (PTF) (4,5).
The aim of this study was to compare the regional distribution of left ventricular (LV) perfusion (measured with 99mTc-MIBI SPECT) and MBF (measured with 15O-water PET) in patients with nonischemic dilated cardiomyopathy and LBBB. The focus of attention was the comparative assessment of regional myocardial wall thickening as estimated with gated 18F-FDG PET and M-mode echocardiography.
MATERIALS AND METHODS
Patients
We investigated 15 patients with dilated cardiomyopathy and complete LBBB. Dilated cardiomyopathy was diagnosed by a reduced LV function in cine ventriculography with exclusion of coronary artery disease by coronary angiography. All patients were in heart failure of functional class III according to the New York Heart Association. Nuclear medicine studies were performed within 2 d. Patient characteristics are listed in Table 1.
Clinical Characteristics
15O-Water PET
Two dynamic resting 15O-water PET studies were started with the infusion of 700 MBq of 15O-water. Each study used an ECAT EXACT 922/47 PET scanner (Siemens-CTI) and consisted of 24 frames (14 × 5 s, 3 × 10 s, 3 × 20 s, and 4 × 30 s). A 15-min period between the 2 studies was used for a transmission scan. Corrections for dead time, attenuation, scatter, and decay were applied, and final images were reconstructed using filtered backprojection (Hann; cutoff, 0.4 Nyquist).
MBF derived from the 15O-water PET studies was quantified using pixelwise modeling software (PMOD; University Hospital Zurich) (6). Because unprocessed 15O-water PET images do not contain sufficient anatomic information for cardiac reorientation and myocardial segmentation, this software applies a factor analysis to the water data that allows reconstruction of 2 image sets: one representing myocardium and another representing blood pool. These factor images were then used for anatomic reference, that is, reorientation of the dynamic 15O-water PET data and segmentation into the septal, anterior, lateral, and posterior walls, excluding the apex. Regions of interest were also drawn within the left and right ventricles to obtain blood time–activity curves as input function (7) and to correct for left and right ventricular spillover (8). Arterial and myocardial tissue time–activity curves were fitted to a single-tissue compartment tracer kinetic model (4,9) to give values of regional MBF (mL × g−1 × min−1) free of the partial-volume effect through the introduction of PTF (4,5). Mean values for PTF were also obtained for each myocardial wall area.
Because MBF is closely related to the rate–pressure product (arterial systolic blood pressure × heart rate), MBF corrected for the rate–pressure product was calculated as corrected MBF (corrMBF = MBF/rate–pressure product × 104). Arterial blood pressure and heart rate were automatically monitored during the 15O-water PET studies with a Dinamap 8100 monitor (Critikon Corp.). Results of both 15O-water PET studies were averaged for every patient.
To measure the precision of the PET measurement in the single study session, we calculated the repeatability coefficient according to the method of Bland and Altman (10). This coefficient is defined as 1.96 × the SD of the differences and is recommended by the British Standards Institution (11). The repeatability coefficient is also given as a percentage of the average value of the 2 measurements.
99mTc-MIBI SPECT
Myocardial perfusion SPECT was performed 60 min after injection of 424 ± 17 MBq of 99mTc-MIBI, with a light meal taken by the patient after tracer application. Data were acquired with a dual-head γ-camera (Solus; Philips Medical Systems). Acquisition parameters and attenuation- and scatter-corrected reconstruction in a 128 × 128 matrix were described in detail elsewhere (12). Briefly, emission was performed in 3 independent energy windows: 140 ± 14 keV for emission, 120 ± 6 keV for scatter detection, and 90 ± 11 keV for backscatter detection. Datasets of windows 1 and 2 were processed to obtain a scatter-corrected dataset, which was then reconstructed using a Butterworth filter (cutoff, 0.7 Nyquist; order, 5; matrix, 128 × 128) and processed with the dataset of window 3 (filtered backprojection, ramp) to obtain a final segmented attenuation- and scatter-corrected transaxial dataset.
99mTc-MIBI images were reoriented according to the LV axes. A volumetric vector sampling method and a 25-segment model were used to display relative 99mTc-MIBI uptake as percentage of the segment with maximum activity. Mean 99mTc-MIBI uptake of the 4 LV wall areas (septal, anterior, lateral, and posterior)—each consisting of 6 segments—was calculated as the mean value of the respective 6 segments for every patient. The remaining apical segment was not included in the wall area analysis.
The transaxial resolution of the SPECT system was measured with a line source filled with 99mTc-pertechnetate. The line source was placed centrally in a physiologic thorax phantom (septal area of the simulated myocardium). The full width at half maximum was 24.9 mm after acquisition and reconstruction with the same parameters used in the patient studies.
Gated 18F-FDG PET
All patients received 250 mg of acipimox 2 h before administration of 18F-FDG and 50 g of glucose orally 1 h before. Gated 18F-FDG PET (ECAT EXACT 922/47; Siemens-CTI) was performed 60 min after intravenous administration of 284 ± 30 MBq of 18F-FDG with 8 gates per R-R interval. The acquisition time was 30 min for emission (2-dimensional mode) and 15 min for transmission (68Ge/68Ga rod sources). Attenuation-corrected gated images were reconstructed using an iterative algorithm (ordered-subsets expectation maximization, 16 subsets, 6 steps). The transaxial resolution of the PET system measured with the 18F-FDG–filled line source in the thorax phantom was 10.9 mm in full width at half maximum.
The pixel size of gated 18F-FDG images was changed to an isotropic 5.79 mm. These data were reoriented and analyzed for regional myocardial wall thickening and LV volumes using Quantitative Gated SPECT (Cedars-Sinai Medical Center) (13). Wall thickening was estimated in the 25-segment model and expressed as the percentage increase from end-diastolic myocardial thickness. Mean wall thickening of the 4 LV wall areas was calculated analogously to the determination of relative 99mTc-MIBI uptake values.
Echocardiography
Echocardiography recordings obtained within 1 mo before the nuclear medicine studies were available for 14 patients. Regional wall thickness for the anteroseptal and posterolateral walls was retrospectively measured from parasternal M-mode recordings without knowledge of the nuclear medicine results. End-diastole was identified by the onset of the QRS complex; end-systole was defined as the smallest LV diameter.
Statistical Analysis
Statistical analyses were performed with SPSS 11 software (SPSS Inc.). Data are expressed as mean ± SD. Differences of values between septal, anterior, lateral, and posterior myocardium were assessed with the 1-way ANOVA followed by a post hoc Bonferroni analysis. Parameters derived from echocardiography were tested using a t test for paired samples. P < 0.05 was considered significant.
RESULTS
At rest, corrMBF was homogeneously distributed in the LV myocardium, without significant differences between the respective wall areas (Fig. 1A). The mean PTF of the posterior wall (0.73 ± 0.11 g × mL−1) was higher than that of the anterior wall (0.63 ± 0.06 g × mL−1, P < 0.05) and that of the septum (0.59 ± 0.13 g × mL−1, P < 0.01). PTF values of the lateral wall (0.68 ± 0.08 g × mL−1), of the septum, and of the anterior wall did not differ significantly from one another. The repeatability coefficients for both 15O-water PET measurements are shown in Table 2. The analysis revealed similar values for the septal, anterior, and posterior wall areas, with a slightly higher repeatability in the lateral wall.
Distribution of LV corrMBF (A), 99mTc-MIBI uptake (B), and wall thickening (C) as determined by gated 18F-FDG PET in 15 patients with LBBB without coronary artery disease.
Repeatability of MBF Measurements
In contrast to the corrMBF values, 99mTc-MIBI uptake proved to be heterogeneously distributed in the 4 LV wall areas (P < 0.0001; Fig. 1B). Mean 99mTc-MIBI uptake was significantly higher in the lateral wall (84% ± 5%) than in the remaining 3 wall areas. The lowest 99mTc-MIBI uptake was observed in the septum (65% ± 10%).
Analysis of regional wall thickening by gated 18F-FDG PET demonstrated a significantly heterogeneous contraction pattern of the left ventricle (P < 0.0001; Fig. 1C). In analogy to the distribution pattern of relative 99mTc-MIBI uptake, wall thickening was highest in the lateral wall (55% ± 17%, P < 0.0001 compared with the remaining wall areas) and lowest in the septum (16% ± 14%).
These findings were substantiated by the echocardiographic M-mode measurements, which demonstrated a comparable regional difference in wall thickening for the corresponding anteroseptal and posterolateral walls (Fig. 2).
End-diastolic (diastole) and end-systolic (systole) wall thickness (A) and systolic wall thickening in percentage of the anteroseptal and posterolateral wall (B) in 14 patients with LBBB without coronary artery disease. Values were determined from parasternal echocardiographic M-mode recordings.
DISCUSSION
It is a well-known phenomenon that patients with complete LBBB exhibit a heterogeneous distribution pattern of myocardial perfusion as estimated with 201Tl or 99mTc-MIBI SPECT. Stress-induced septal hypoperfusion is frequently observed, but diminished septal perfusion, compared with the lateral free wall, also occurs in resting studies even in the absence of coronary artery disease (1,2). Downregulated perfusion secondary to diminished oxygen demand, increased intramyocardial pressure during diastole, fibrotic alterations, or pronounced partial-volume effects due to reduced wall thickening have been discussed as the pathophysiologic correlates of these observations (3,14–17).
Partial-volume effects are of particular relevance in myocardial SPECT studies. Because of the limited spatial resolution of γ-cameras, systolic thickening and diastolic relaxation of the myocardium entail continuous changes in the recovery coefficient during acquisition. Therefore, the myocardial counts actually measured depend considerably on systolic thickening, as marked thickening leads to a higher recovery coefficient and subsequently to more myocardial counts.
The main findings of our study are that, in patients with dilated cardiomyopathy and LBBB, resting MBF and its LV distribution pattern as estimated with the partial-volume–independent method of 15O-water PET are within the range of values observed in healthy controls, without evidence of regional heterogeneity (18). However, 99mTc-MIBI uptake and wall thickening as estimated by gated 18F-FDG PET differ significantly between the respective wall areas in an identical pattern. This finding was confirmed by M-mode echocardiography from corresponding regions.
There is further evidence of regional heterogeneity in systolic shortening and myocardial wall thickness in patients with asynchronous electrical activation of the left ventricle. In animal experiments during ventricular pacing, regional systolic shortening proved to be reduced in early-activated regions and increased in late-activated regions as determined with MRI tagging (19). These data could recently be confirmed by echocardiographic strain rate imaging in humans with complete LBBB, whose interventricular septum represents the early-activated region and the lateral free wall the late-activated region (20). The reduction of 99mTc-MIBI uptake in the anterior and posterior wall in our patients must also be interpreted as resulting from a relative reduction in regional wall thickening compared with the lateral wall. This reduction in wall thickening occurred mainly in the adjacent anteroseptal and posteroseptal parts of the anterior and posterior walls, respectively.
Furthermore, diminished end-diastolic wall thickness of the septum, compared with the late-activated posterior wall, has been described in LBBB, whereas end-diastolic wall thickness was identical in both wall areas in healthy controls (21). In our echocardiography analysis, we could verify a trend of smaller end-diastolic wall thickness in the anteroseptal wall compared with the posterolateral wall, which, however, did not reach statistical significance.
Therefore, at least one of these alterations in LBBB, namely reduced systolic thickening and diminished end-diastolic wall thickness in the septum and adjacent anteroseptal and posteroseptal walls, provides a pathophysiologic explanation for the heterogeneous uptake pattern in resting 201Tl or 99mTc-MIBI SPECT studies due to pronounced partial-volume effects.
Our results argue against the concept of downregulated septal perfusion caused by diminished oxygen demand, as MBF was found to be similar in all myocardial regions by 15O-water PET. This finding is underscored by data of Zanco et al., who, using 13N-NH3 PET, found preserved septal perfusion in patients with LBBB and reduced LV function but without coronary artery disease (22).
It is unlikely that the variability in the 15O-water measurements influenced our results. Calculation of repeatability coefficients from our data revealed good agreement of both measurements in each patient. The repeatability coefficients of our parameters were almost identical to previously published values obtained in similar settings (6,23).
CONCLUSION
15O-Water PET proves that resting MBF is homogeneously distributed in patients with nonischemic dilated cardiomyopathy and LBBB. The quantitative analysis of 99mTc-MIBI SPECT in these patients is significantly affected by heterogeneity in regional myocardial systolic wall thickening. Therefore, supplementary analysis and consideration of wall thickening derived from ECG-gated data is recommended in the assessment of 99mTc-MIBI SPECT in patients with LBBB. The reader should take into account that the combination of reduced septal 99mTc-MIBI uptake and reduced septal wall thickening does not necessarily represent myocardial infarction, as this combination usually occurs in patients without LV conduction disturbances (24). Reduced septal 99mTc-MIBI uptake should rather be considered the artificial result of reduced wall thickening.
Footnotes
Received Aug. 5, 2003; revision accepted Oct. 13, 2003.
For correspondence or reprints contact: Bernd Nowak, MD, Department of Nuclear Medicine, University Hospital, Aachen University of Technology, Pauwelsstrasse 30, 52074 Aachen, Germany.
E-mail: bnowak{at}ukaachen.de
REFERENCES
In this issue
Jump to section
Related Articles
Cited By...
- Effects of Epicardial and Endocardial Cardiac Resynchronization Therapy on Coronary Flow: Insights From Wave Intensity Analysis
- Myocardial Structural, Perfusion, and Metabolic Correlates of Left Bundle Branch Block Mechanical Derangement in Patients With Dilated Cardiomyopathy: A Tagged Cardiac Magnetic Resonance and Positron Emission Tomography Study
- Effects of Left Bundle-Branch Block on Cardiac Structure, Function, Perfusion, and Perfusion Reserve: Implications for Myocardial Contrast Echocardiography Versus Radionuclide Perfusion Imaging for the Detection of Coronary Artery Disease
- Is Septal Glucose Metabolism Altered in Patients with Left Bundle Branch Block and Ischemic Cardiomyopathy?
- Attenuation Correction of Myocardial SPECT Perfusion Images with Low-Dose CT: Evaluation of the Method by Comparison with Perfusion PET