- Email: [email protected]
Electromechanical mapping for detecting myocardial viability and ischemia in patients with severe ischemic cardiomyopathy
Electromechanical Mapping for Detecting Myocardial Viability and Ischemia in Patients With Severe Ischemic Cardiomyopathy Habib Samady, MD, Yi-Hwa Liu, PhD, C. Joon Choi, MD, PhD, Michael Ragosta, MD, Steven E. Pfau, MD, Michael W. Cleman, MD, Eric R. Powers, MD, Christopher M. Kramer, MD, Frans J. Th. Wackers, MD, PhD, George A. Beller, MD, and Denny D. Watson, PhD This study was designed to evaluate several electromechanical mapping parameters for assessment of myocardial viability and inducible ischemia as defined by dipyridamole single-photon emission computed tomographic (SPECT) imaging at rest in patients with severe ischemic cardiomyopathy. Unipolar voltage, normalized unipolar voltage, bipolar voltage, and fragmentation were compared with tracer uptake at rest and reversibility on stress or rest quantitative technetium-99m sestamibi SPECT imaging in 32 patients with severe ischemic cardiomyopathy (left ventricular ejection fraction 0.24 ⴞ 0.08). In dysfunctional myocardial segments, logistic regression showed unipolar voltage, normalized unipolar voltage, and bipolar voltage to be predictive of viable myocardium (>60% tracer uptake at rest) and
was significantly higher in viable than in nonviable segments (p <0.01). A unipolar voltage of >7.1 mV was the best predictor of viable myocardium. In dysfunctional viable segments, unipolar voltage was significantly higher in reversible than in fixed segments (p <0.001), and a unipolar voltage of >8.5 mV had optimal power for identifying reversibility on dipyridamole SPECT imaging. We conclude that in patients with severe ischemic cardiomyopathy, unipolar voltage can identify viable from nonviable myocardium and reversible from fixed viable defects as defined by dipyridamole technetium-99m sestamibi SPECT imaging. 䊚2003 by Excerpta Medica, Inc. (Am J Cardiol 2003;91:807– 811)
e hypothesized that in patients with profound ischemic left ventricular (LV) dysfunction, W electromechanical mapping (EMM)-derived electric
patients with peripheral vascular disease, aortic stenosis, moderate or greater mitral regurgitation, unstable ischemic syndromes, recent (⬍1 month) myocardial infarction, atrial fibrillation, LV thrombus, and a degree of LV dysfunction out of proportion to the extent of coronary disease. The Investigational Review Board of the University of Virginia and Yale University approved the protocol, and all patients signed informed consent. Myocardial perfusion imaging: After an overnight fast, patients underwent Tc-99m sestamibi imaging at rest. Twenty-eight patients also underwent stress Tc99m sestamibi imaging with dipyridamole. In patients who underwent imaging at rest only, SPECT imaging was performed 45 to 60 minutes after a 30-mCi injection of Tc-99m sestamibi. In the 28 patients who underwent imaging at rest and during stress, SPECT imaging at rest was performed 45 to 60 minutes after a 10-mCi injection of Tc-99m sestamibi. After a 2- to 3-hour delay, 0.56 mg/kg of dipyridamole was infused intravenously over 4 minutes. Thirty millicuries of Tc-99m sestamibi were injected 7 minutes after the start of the dipyridamole infusion. An electrocardiogram, heart rate, and blood pressure were recorded every minute. SPECT imaging and quantitative processing were performed with techniques previously reported from our laboratory.1 Reconstructed images were divided into 14 segments/patient as follows: apical slices were
and mechanical parameters would detect myocardium with inducible ischemia and preserved viability as assessed by dipyridamole technetium-99m (Tc-99m) sestamibi single-photon emission computed tomographic (SPECT) myocardial perfusion imaging at rest.
METHODS
Study design: Thirty-two patients with severe ischemic LV dysfunction and coronary artery disease (CAD) referred for cardiac catheterization underwent EMM using the Biosense NOGA system (BiosenseWebster, Diamond Bar, California) and Tc-99m sestamibi SPECT imaging. Exclusion criteria included
From the Cardiovascular Division, Department of Medicine, and the Department of Radiology, University of Virginia Health System, Charlottesville, Virginia; and Section of Cardiovascular Medicine, Yale University School of Medicine, New Haven, Connecticut. This study was supported by research grants from Bristol-Myers Squibb Medical Imaging, North Billerica, Massachusetts, and Biosense Webster, Diamond Bar, California. Manuscript received August 13, 2002; revised manuscript received and accepted December 17, 2002. Address for reprints: Habib Samady, MD, Cardiovascular Division, Department of Medicine, University of Virginia Health System, PO Box 800158, Charlottesville, Virginia 22908-0158. E-mail: [email protected]. ©2003 by Excerpta Medica, Inc. All rights reserved. The American Journal of Cardiology Vol. 91 April 1, 2003
0002-9149/03/$–see front matter doi:10.1016/S0002-9149(03)00012-2
807
FIGURE 1. A, quantitative Tc-99m sestamibi SPECT image during stress and at rest in a patient with a reversible defect at the apex and anteroseptal segments of the midventricle and a predominantly fixed defect in the basal septum. B, quantitative Tc99m sestamibi SPECT image during stress and at rest in a patient with a nonviable fixed defect at the apex and anteroseptal segments.
divided into inferoapical and anteroapical; midventricular and basal slices were divided into 6 equal regions (anterior, anterolateral, inferolateral, inferior, inferoseptal, and anteroseptal). Quantitative determination of segmental Tc-99m sestamibi uptake was performed using the University of Virginia software program (Figure 1).1 In 20 patients, electrocardiogram-gated stress SPECT images were performed using the standard 8 frames/cardiac cycle. SPECT images were read by an experienced reader (blinded to clinical data) as interpretable or artifact. Electromechanical mapping: EMM was performed within 1 week of the SPECT imaging. The EMM system 808 THE AMERICAN JOURNAL OF CARDIOLOGY姞
VOL. 91
has been previously described in detail.2 At the end of cardiac catheterization, the diagnostic femoral sheath was exchanged for a 45-cm, 8Fr sheath. Systemic heparin (70 U/kg) was administered. The mapping catheter was advanced into the left ventricle and points were acquired when the catheter tip was stable on the endocardium; this occurred after the documentation of local activation time stability, location stability, loop stability, and cyclelength stability. An interpolation threshold of 40 mm was set between adjacent points. The catheter was removed from the left ventricle when all endocardial regions were represented on the reconstructed map. For each patient, color-coded unipolar voltage and linear shortening maps and their corresponding “bull’s-eye” maps were generated. EMM images were divided into 14 segments corresponding to SPECT images. The EMM bull’s-eye maps were generated when a fixed cylindric polar reference coordinate map was defined with anatomic reference points acquired at end-diastole. The center of mass of the reconstructed LV chamber was automatically calculated by the system from the set of endocardial points that were sampled. The long axis of the left ventricle was defined as the line connecting the apex and the center of mass. The long axis was divided into 3 slices (apex, midventricle, and base [Figures 2 and 3]). Unipolar voltage, bipolar voltage, fragmentation, and linear shortening data points were averaged for each of the 14 myocardial segments. Normalized unipolar voltage was calculated by expressing unipolar voltage as a percentage of the highest segmental unipolar voltage for that patient. Mean unipolar voltage was calculated from the 14 individual segment voltages for each patient.
Definition of normal, viable, reversible, and nonviable myocardium: Myocardial seg-
ments were defined as being normal when linear shortening by EMM was normal (ⱖ8.9%, ⬍2 SDs below the mean of normal).3 Dysfunctional segments (linear shortening ⬍8.9%) with ⱖ60% Tc-99m sestamibi uptake at rest were considered viable, and dysfunctional segments with ⬍60% tracer uptake at rest were considered nonviable.4 Viable segments were further divided into reversible (rest–stress tracer uptake ⱖ10%) and fixed (rest–stress tracer uptake ⬍10%). As a validation of EMM functional assessment, in 20 patients with EMM and gated SPECT data, the proportion of dysfunctional segments by EMM were compared with APRIL 1, 2003
FIGURE 2. A, EMM map of the patient whose SPECT images are shown in Figure 1A. The map demonstrates hypokinesia in the apical, anteroseptal (AS), and anterobasal segments (right) with preserved voltage in the apex and mid-anteroseptal segments and diminished voltage in the basal septum (left). The lowest values are represented in red and the highest values in purple. B, The bull’s-eye demonstrates a unipolar voltage of >8.5 mV in the apical and anteroseptal segments that is consistent with inducible ischemia on SPECT and diminished voltage at the base suggestive of scar or mitral valve plane. A ⴝ anterior; I ⴝ inferior; IS ⴝ inferoseptal; L ⴝ lateral; P ⴝ posterior.
that by quantitative gated SPECT imaging (thickening fractions ⬍10% were considered dysfunctional by gated SPECT). Data analysis: Myocardial segments were matched by location for both modalities. Segments with artifact on SPECT myocardial perfusion imaging or no data points on EMM were excluded from this analysis. Fourteen segments per patient were used for univariate and logistic regression analyses, which compared EMM to SPECT data. For continuous correlations of EMM and SPECT variables, data from adjacent shortaxis segments from the mid-ventricular and base slices were merged, yielding 8 segments/patient (2 apical, 3 midventricular, and 3 basal). Statistical analysis: Continuous data are expressed as mean ⫾ SD. EMM variables were compared in normal, reversible, viable fixed, and nonviable groups using analysis of variance. When appropriate, a post-hoc comparison of paired variables was made using Bonferroni’s modification. Using Pearson’s correlation, a continuous relation was sought between EMM variables and Tc99m sestamibi uptake at rest. For dysfunctional segments, logistic regression and receiver-operating characteristic (ROC) curve analyses were performed to determine the predictive power and optimal discriminatory
FIGURE 3. A, EMM map of the patient whose SPECT images are shown in Figure 1B. The map demonstrates akinesia in the apical, anteroseptal (AS), and anterobasal segments (right) with correspondingly decreased voltage in the apex and anteroseptal segments (left). B, the bull’s eye shows voltage in the apical and anteroseptal segments to be <7.1 mV, consistent with nonviable myocardium. The normalized unipolar voltage is shown within brackets. RA ⴝ right arm; other abbreviations as in Figure 2.
values of the EMM variables for viable myocardium and reversibility during dipyridamole SPECT imaging. A p value of ⬍0.05 was considered significant.
RESULTS
Patient characteristics: The study population consisted of 25 men and 7 women (mean age 65 ⫾ 14 years). Sixty percent of patients had a prior myocardial infarction, mean Canadian Cardiovascular Society classification angina score was 2 ⫾ 1, and mean New York Heart Association heart failure score was 3 ⫾ 1. Cardiac catheterization revealed critical 3-vessel CAD in 26 patients, 2-vessel CAD in 6 patients, and mean ejection fraction of 0.24 ⫾ 0.08 by contrast left ventriculography. EMM and SPECT imaging characteristics: All 32 patients underwent EMM and SPECT imaging at rest, which yielded 448 matched myocardial segments for viability assessment. After automated stability criteria editing, 84 ⫾ 38 points were sampled per patient map. Volumetric measurements by EMM revealed an enddiastolic volume of 199 ⫾ 74 ml (end-diastolic volume index 101 ⫾ 38), end-systolic volume of 150 ⫾ 65 ml (end-systolic volume index 76 ⫾ 33), and LV ejection fraction of 0.26 ⫾ 0.08. Mapping time was 47 ⫾ 20 minutes. There were no complications related to the
CORONARY ARTERY DISEASE/ ELECTROMECHANICAL MAPPING FOR VIABILITY ASSESSMENT
809
TABLE 1 Electromechanical Data in 433 Myocardial Segments Divided According to Rest Function Defined by Electromechanical Mapping (EMM) and Viability Status According to Radionuclide Single-photon Emission Computed Tomographic (SPECT) Imaging
Normal (n ⫽ 152) Viable (n ⫽ 215) Nonviable (n ⫽ 66) p Value Overall analysis of variance Normal vs viable segments Normal vs nonviable segments Viable vs nonviable segments
Linear Shortening (%)
Unipolar Voltage (mV)
Normalized Unipolar Voltage (%)
Bipolar Voltage (mV)
13.6 ⫾ 3.6 3.6 ⫾ 3.9 2.4 ⫾ 4.2
10.4 ⫾ 4.4 8.7 ⫾ 4.1 6.8 ⫾ 3.2
67 ⫾ 23 58 ⫾ 24 48 ⫾ 25
3.8 ⫾ 2.7 3.2 ⫾ 2.6 2.3 ⫾ 1.9
⬍0.001 ⬍0.05 ⬍0.05 NS
⬍0.001 ⬍0.05 ⬍0.05 ⬍0.05
⬍0.001 ⬍0.05 ⬍0.05 ⬍0.05
⬍0.001 ⬍0.05 ⬍0.05 NS
TABLE 2 Electromechanical Data in 370 Myocardial Segments Divided According to Rest Function and Perfusion Status by Radionuclide Single-photon Emission Computed Tomographic (SPECT) Imaging Linear Shortening (%) Normal (n ⫽ 131) Reversible (n ⫽ 35) Fixed viable (n ⫽ 147) Nonviable (n ⫽ 57) p Value Overall analysis of variance Reversible vs nonviable segments Reversible vs fixed viable segments Fixed viable vs nonviable segments
13.6 3.7 3.5 2.2
⫾ ⫾ ⫾ ⫾
3.6 4.4 3.7 4.4
⬍0.001 ⬍0.05 NS ⬍0.05
10.4 11.1 8.2 6.8
⫾ ⫾ ⫾ ⫾
4.6 4.8 3.9 3.3
⬍0.001 ⬍0.05 ⬍0.05 ⬍0.05
EMM procedure. Of 448 matched myocardial segments, 433 segments had adequate EMM and SPECT data for analysis. Of the 433 segments, 152 had normal function and 281 were dysfunctional by EMM. Of the 264 myocardial segments with EMM and gated SPECT functional data, 170 (64.4%) were dysfunctional by EMM and 168 (63.6%) by gated SPECT (p ⫽ NS). Of the dysfunctional segments, 215 (77%) were viable and 66 (23%) were nonviable. In 28 patients who underwent SPECT imaging at rest and during stress, 370 of 392 myocardial segments had adequate EMM and SPECT data for analysis. Of the 370 segments, 131 had normal function and 239 were dysfunctional. Of the dysfunctional segments, 35 (15%) were reversible, 147 (61%) were fixed viable, and 57 (24%) were nonviable. EMM for myocardial viability assessment: There was a significant correlation between Tc-99m sestamibi uptake at rest and both unipolar voltage and normalized unipolar voltage for all myocardial segments (r ⫽ 0.34 and 0.31, respectively; p ⬍0.001 for both). This correlation was best at the apex for unipolar and normalized unipolar voltage (r ⫽ 0.58 and 0.52, respectively; p ⬍0.001 for both). No continuous correlation was demonstrated between tracer uptake at rest and either bipolar voltage or fragmentation. Of the dysfunctional segments, unipolar voltage, normalized unipolar voltage, and bipolar voltage were predictive of viable myocardium during SPECT imaging (p ⬍0.001 for each variable). Fragmentation did not predict viable myocardium (p ⫽ NS) and therefore was not included in further analysis. Table 1 lists values for unipolar voltage, normalized 810 THE AMERICAN JOURNAL OF CARDIOLOGY姞
Unipolar Voltage (mV)
VOL. 91
Normalized Unipolar Voltage (%) 67 65 57 48
⫾ ⫾ ⫾ ⫾
23 24 23 26
0.003 ⬍0.05 NS ⬍0.05
Bipolar Voltage (mV) 3.9 3.2 3.1 2.3
⫾ ⫾ ⫾ ⫾
2.7 2.6 2.2 2.0
0.052 NS NS ⬍0.001
unipolar voltage, and bipolar voltage in normal, viable, and nonviable segments. Unipolar voltage, normalized unipolar voltage, and bipolar voltage variables were significantly greater in normal than dysfunctional segments, whether the segments were viable or nonviable. Furthermore, unipolar voltage and normalized unipolar voltage were significantly higher in viable than in nonviable segments. The best EMM predictor of viability was a unipolar voltage of ⱖ7.1 mV, which yielded a sensitivity of 70%, specificity of 60%, and an area under the ROC curve of 0.69 ⫾ 0.04 (p ⬍0.001). A normalized unipolar voltage of ⱖ53% had optimal value for detecting viable myocardium, with a sensitivity of 64%, a specificity of 60%, and an area under the ROC curve of 0.65 ⫾ 0.04 (p ⬍0.001). Mean segmental unipolar voltage per patient correlated with global LV ejection fraction (r ⫽ 0.41, p ⬍0.001). EMM for inducible ischemia assessment: Table 2 lists electromechanical data in myocardial segments according to perfusion defects, with the normal segments as controls. Unipolar voltage was significantly greater in reversible than in fixed viable segments, and the 2 groups had higher unipolar voltage than nonviable segments. Normalized unipolar voltage was significantly higher in reversible than in nonviable segments; it was also higher in fixed viable segments than in nonviable segments. However, unlike unipolar voltage, normalized unipolar voltage was not significantly different between reversible and fixed viable segments. There were no significant differences in bipolar voltage between reversible and fixed viable segments (p ⫽ NS). Unipolar voltage was the best EMM predictor of APRIL 1, 2003
inducible ischemia on SPECT imaging, with a voltage of ⱖ8.5 mV having a sensitivity of 71%, a specificity of 60%, and an area under the ROC curve of 0.65 ⫾ 0.05 (p ⫽ 0.002). Normalized unipolar voltage only showed a trend toward predicting inducible ischemia (p ⫽ 0.052).
DISCUSSION This is the first study evaluating LV EMM for detection of myocardial viability and inducible ischemia, as defined by dipyridamole SPECT myocardial perfusion imaging, in patients with severe ischemic cardiomyopathy. A unipolar voltage of ⱖ7.1 mV and a normalized unipolar voltage of ⱖ53% had optimal discriminatory power for detecting viable myocardium. A unipolar voltage of ⱖ8.5 mV had the best predictive value for detecting inducible ischemia. EMM parameters for the detection of myocardial viability: We found that a unipolar voltage of 7.1 mV
was best able to discriminate viable from nonviable myocardium in dysfunctional myocardial segments in patients with profound LV dysfunction (LV ejection fraction 0.24 ⫾ 0.08). Our findings are consistent with those of other investigators who, in patients with less severe LV dysfunction, found a unipolar voltage of 6.5 mV to discriminate viable from nonviable myocardium as defined by positron emission tomography5,6 and thallium-201.7,8 Two of these studies7,8 excluded patients with a LV ejection fraction of ⬍0.30 (mean LV ejection fraction 0.40 ⫾ 0.48), and a third study6 was predominantly in patients with single-vessel CAD (LV ejection fraction 0.49 ⫾ 0.15). Therefore, our data extend the applicability of EMM to a patient population with profound ischemic LV dysfunction, in whom the assessment of myocardial viability is most relevant. The specificity of unipolar voltage for detecting viable myocardium is only fair (60%). This may reflect the limitation of Tc-99m sestamibi for detecting viable myocardium in the inferobasal segments due to diaphragmatic attenuation artifact.9 We found the correlation between unipolar voltage and Tc-99m sestamibi on SPECT to be weakest in the basal segments (r ⫽ 0.20, p ⫽ 0.001) where both technologies have limitations (fibrous mitral valve plane decreasing voltage on EMM and diaphragmatic attenuation causing artifact on SPECT), and best at the apex (r ⫽ 0.58, p ⬍0.001). Further studies investigating the predictive value of the 2 technologies in this patient population for recovery of ventricular function are warranted. We observed a continuous relation between mean segmental unipolar voltage per patient and global ventricular function, which suggests that patients with profound LV dysfunction had globally diminished voltage, even in myocardial segments remote from the demonstrable pathology by SPECT. This could reflect ventricular remodeling and may explain why absolute voltage was found to be a better discriminator of viability and ischemia than normalized voltage.
Unlike previous investigators,10 we found unipolar voltage to be a better indicator of viability than bipolar voltage. However, Wolf et al11 investigated a canine model of myocardial infarction where exhaustive and detailed sampling may have been more practical than in the clinical setting. In addition, unipolar voltage measures voltage in a wider field than bipolar voltage, thus mirroring the larger resolution of SPECT imaging, which we used as the “gold” standard. EMM parameters for the detection of myocardial ischemia: This is the first study demonstrating the
relation between preserved voltage and inducible ischemia in patients with severe ischemic LV dysfunction. Unipolar voltage was significantly higher in reversible viable than in fixed viable segments, with a unipolar voltage of ⱖ8.5 mV being the best discriminator of inducible ischemia. Fuchs et al8 found a similar unipolar voltage of 9.0 mV to discriminate inducible ischemia in patients with preserved ventricular function (LV ejection fraction ⫽ 0.48 ⫾ 0.11). Acknowledgment: The investigators thank Jennifer Hunter, RN, Linda Snyder, RN, and Katherine Bunger, RN, for their invaluable assistance, and Jerry Curtis, MA, for his superb editorial assistance.
1. Smith WH, Kastner RJ, Calnon DA, Segalla D, Beller GA, Watson DD. Quantitative gated SPECT imaging: a counts-based method for display and measurement of regional and global ventricular systolic function. J Nucl Cardiol 1997;4:451–463. 2. Kornowski R, Hong MK, Leon MB. Comparison between left ventricular electromechanical mapping and radionuclide perfusion imaging for detection of myocardial viability. Circulation 1998;98:1837–1841. 3. Kornowski R, Hong MK, Gepstein L, Goldstein S, Ellahham S, Ben-Haim SA, Leon MB. Preliminary animal and clinical experiences using an electromechanical endocardial mapping procedure to distinguish infarcted from healthy myocardium. Circulation 1998;98:1116 –1124. 4. Udelson JE, Coleman PS, Metherall J, Pandian NG, Gomez AR, Griffith JL, Shea NL, Oates E, Konstam MA. Predicting recovery of severe regional ventricular dysfunction. Comparison of resting scintigraphy with 201Tl and 99mTcsestamibi. Circulation 1994;89:2552–2561. 5. Botker HE, Lassen JF, Hermansen F, Wiggers H, Sogaard P, Kim WY, Bottcher M, Thuesen L, Pedersen AK. Electromechanical mapping for detection of myocardial viability in patients with ischemic cardiomyopathy. Circulation 2001;103:1631–1637. 6. Koch KC, vom Dahl J, Wenderdel M, Nowak B, Schaefer WM, Sasse A, Stellbrink C, Buell U, Hanrath P. Myocardial viability assessment by endocardial electroanatomic mapping: comparison with metabolic imaging and functional recovery after coronary revascularization. J Am Coll Cardiol 2001;38:91–98. 7. Gyongyosi M, Sochor H, Khorsand A, Gepstein L, Glogar D. Online myocardial viability assessment in the catheterization laboratory via NOGA electroanatomic mapping: quantitative comparison with thallium-201 uptake. Circulation 2001;104:1005–1011. 8. Fuchs S, Hendel RC, Baim DS, Moses JW, Pierre A, Laham RJ, Hong MK, Kuntz RE, Pietrusewicz M, Bonow RO, et al. Comparison of endocardial electromechanical mapping with radionuclide perfusion imaging to assess myocardial viability and severity of myocardial ischemia in angina pectoris. Am J Cardiol 2001;87:874 –880. 9. Soufer R, Dey HM, Ng CK, Zaret BL. Comparison of sestamibi single-photon emission computed tomography with positron emission tomography for estimating left ventricular myocardial viability. Am J Cardiol 1995;75:1214 –1219. 10. Dilsizian V, Bonow RO. Current diagnostic techniques of assessing myocardial viability in hibernating and stunned myocardium. Circulation 1993;87:1–20. 11. Wolf T, Gepstein L, Dror U, Hayam G, Shofti R, Zaretzky A, Uretzky G, Oron U, Ben-Haim SA. Detailed endocardial mapping accurately predicts the transmural extent of myocardial infarction. J Am Coll Cardiol 2001;37:1590 – 1597.
CORONARY ARTERY DISEASE/ ELECTROMECHANICAL MAPPING FOR VIABILITY ASSESSMENT
811
was significantly higher in viable than in nonviable segments (p <0.01). A unipolar voltage of >7.1 mV was the best predictor of viable myocardium. In dysfunctional viable segments, unipolar voltage was significantly higher in reversible than in fixed segments (p <0.001), and a unipolar voltage of >8.5 mV had optimal power for identifying reversibility on dipyridamole SPECT imaging. We conclude that in patients with severe ischemic cardiomyopathy, unipolar voltage can identify viable from nonviable myocardium and reversible from fixed viable defects as defined by dipyridamole technetium-99m sestamibi SPECT imaging. 䊚2003 by Excerpta Medica, Inc. (Am J Cardiol 2003;91:807– 811)
e hypothesized that in patients with profound ischemic left ventricular (LV) dysfunction, W electromechanical mapping (EMM)-derived electric
patients with peripheral vascular disease, aortic stenosis, moderate or greater mitral regurgitation, unstable ischemic syndromes, recent (⬍1 month) myocardial infarction, atrial fibrillation, LV thrombus, and a degree of LV dysfunction out of proportion to the extent of coronary disease. The Investigational Review Board of the University of Virginia and Yale University approved the protocol, and all patients signed informed consent. Myocardial perfusion imaging: After an overnight fast, patients underwent Tc-99m sestamibi imaging at rest. Twenty-eight patients also underwent stress Tc99m sestamibi imaging with dipyridamole. In patients who underwent imaging at rest only, SPECT imaging was performed 45 to 60 minutes after a 30-mCi injection of Tc-99m sestamibi. In the 28 patients who underwent imaging at rest and during stress, SPECT imaging at rest was performed 45 to 60 minutes after a 10-mCi injection of Tc-99m sestamibi. After a 2- to 3-hour delay, 0.56 mg/kg of dipyridamole was infused intravenously over 4 minutes. Thirty millicuries of Tc-99m sestamibi were injected 7 minutes after the start of the dipyridamole infusion. An electrocardiogram, heart rate, and blood pressure were recorded every minute. SPECT imaging and quantitative processing were performed with techniques previously reported from our laboratory.1 Reconstructed images were divided into 14 segments/patient as follows: apical slices were
and mechanical parameters would detect myocardium with inducible ischemia and preserved viability as assessed by dipyridamole technetium-99m (Tc-99m) sestamibi single-photon emission computed tomographic (SPECT) myocardial perfusion imaging at rest.
METHODS
Study design: Thirty-two patients with severe ischemic LV dysfunction and coronary artery disease (CAD) referred for cardiac catheterization underwent EMM using the Biosense NOGA system (BiosenseWebster, Diamond Bar, California) and Tc-99m sestamibi SPECT imaging. Exclusion criteria included
From the Cardiovascular Division, Department of Medicine, and the Department of Radiology, University of Virginia Health System, Charlottesville, Virginia; and Section of Cardiovascular Medicine, Yale University School of Medicine, New Haven, Connecticut. This study was supported by research grants from Bristol-Myers Squibb Medical Imaging, North Billerica, Massachusetts, and Biosense Webster, Diamond Bar, California. Manuscript received August 13, 2002; revised manuscript received and accepted December 17, 2002. Address for reprints: Habib Samady, MD, Cardiovascular Division, Department of Medicine, University of Virginia Health System, PO Box 800158, Charlottesville, Virginia 22908-0158. E-mail: [email protected]. ©2003 by Excerpta Medica, Inc. All rights reserved. The American Journal of Cardiology Vol. 91 April 1, 2003
0002-9149/03/$–see front matter doi:10.1016/S0002-9149(03)00012-2
807
FIGURE 1. A, quantitative Tc-99m sestamibi SPECT image during stress and at rest in a patient with a reversible defect at the apex and anteroseptal segments of the midventricle and a predominantly fixed defect in the basal septum. B, quantitative Tc99m sestamibi SPECT image during stress and at rest in a patient with a nonviable fixed defect at the apex and anteroseptal segments.
divided into inferoapical and anteroapical; midventricular and basal slices were divided into 6 equal regions (anterior, anterolateral, inferolateral, inferior, inferoseptal, and anteroseptal). Quantitative determination of segmental Tc-99m sestamibi uptake was performed using the University of Virginia software program (Figure 1).1 In 20 patients, electrocardiogram-gated stress SPECT images were performed using the standard 8 frames/cardiac cycle. SPECT images were read by an experienced reader (blinded to clinical data) as interpretable or artifact. Electromechanical mapping: EMM was performed within 1 week of the SPECT imaging. The EMM system 808 THE AMERICAN JOURNAL OF CARDIOLOGY姞
VOL. 91
has been previously described in detail.2 At the end of cardiac catheterization, the diagnostic femoral sheath was exchanged for a 45-cm, 8Fr sheath. Systemic heparin (70 U/kg) was administered. The mapping catheter was advanced into the left ventricle and points were acquired when the catheter tip was stable on the endocardium; this occurred after the documentation of local activation time stability, location stability, loop stability, and cyclelength stability. An interpolation threshold of 40 mm was set between adjacent points. The catheter was removed from the left ventricle when all endocardial regions were represented on the reconstructed map. For each patient, color-coded unipolar voltage and linear shortening maps and their corresponding “bull’s-eye” maps were generated. EMM images were divided into 14 segments corresponding to SPECT images. The EMM bull’s-eye maps were generated when a fixed cylindric polar reference coordinate map was defined with anatomic reference points acquired at end-diastole. The center of mass of the reconstructed LV chamber was automatically calculated by the system from the set of endocardial points that were sampled. The long axis of the left ventricle was defined as the line connecting the apex and the center of mass. The long axis was divided into 3 slices (apex, midventricle, and base [Figures 2 and 3]). Unipolar voltage, bipolar voltage, fragmentation, and linear shortening data points were averaged for each of the 14 myocardial segments. Normalized unipolar voltage was calculated by expressing unipolar voltage as a percentage of the highest segmental unipolar voltage for that patient. Mean unipolar voltage was calculated from the 14 individual segment voltages for each patient.
Definition of normal, viable, reversible, and nonviable myocardium: Myocardial seg-
ments were defined as being normal when linear shortening by EMM was normal (ⱖ8.9%, ⬍2 SDs below the mean of normal).3 Dysfunctional segments (linear shortening ⬍8.9%) with ⱖ60% Tc-99m sestamibi uptake at rest were considered viable, and dysfunctional segments with ⬍60% tracer uptake at rest were considered nonviable.4 Viable segments were further divided into reversible (rest–stress tracer uptake ⱖ10%) and fixed (rest–stress tracer uptake ⬍10%). As a validation of EMM functional assessment, in 20 patients with EMM and gated SPECT data, the proportion of dysfunctional segments by EMM were compared with APRIL 1, 2003
FIGURE 2. A, EMM map of the patient whose SPECT images are shown in Figure 1A. The map demonstrates hypokinesia in the apical, anteroseptal (AS), and anterobasal segments (right) with preserved voltage in the apex and mid-anteroseptal segments and diminished voltage in the basal septum (left). The lowest values are represented in red and the highest values in purple. B, The bull’s-eye demonstrates a unipolar voltage of >8.5 mV in the apical and anteroseptal segments that is consistent with inducible ischemia on SPECT and diminished voltage at the base suggestive of scar or mitral valve plane. A ⴝ anterior; I ⴝ inferior; IS ⴝ inferoseptal; L ⴝ lateral; P ⴝ posterior.
that by quantitative gated SPECT imaging (thickening fractions ⬍10% were considered dysfunctional by gated SPECT). Data analysis: Myocardial segments were matched by location for both modalities. Segments with artifact on SPECT myocardial perfusion imaging or no data points on EMM were excluded from this analysis. Fourteen segments per patient were used for univariate and logistic regression analyses, which compared EMM to SPECT data. For continuous correlations of EMM and SPECT variables, data from adjacent shortaxis segments from the mid-ventricular and base slices were merged, yielding 8 segments/patient (2 apical, 3 midventricular, and 3 basal). Statistical analysis: Continuous data are expressed as mean ⫾ SD. EMM variables were compared in normal, reversible, viable fixed, and nonviable groups using analysis of variance. When appropriate, a post-hoc comparison of paired variables was made using Bonferroni’s modification. Using Pearson’s correlation, a continuous relation was sought between EMM variables and Tc99m sestamibi uptake at rest. For dysfunctional segments, logistic regression and receiver-operating characteristic (ROC) curve analyses were performed to determine the predictive power and optimal discriminatory
FIGURE 3. A, EMM map of the patient whose SPECT images are shown in Figure 1B. The map demonstrates akinesia in the apical, anteroseptal (AS), and anterobasal segments (right) with correspondingly decreased voltage in the apex and anteroseptal segments (left). B, the bull’s eye shows voltage in the apical and anteroseptal segments to be <7.1 mV, consistent with nonviable myocardium. The normalized unipolar voltage is shown within brackets. RA ⴝ right arm; other abbreviations as in Figure 2.
values of the EMM variables for viable myocardium and reversibility during dipyridamole SPECT imaging. A p value of ⬍0.05 was considered significant.
RESULTS
Patient characteristics: The study population consisted of 25 men and 7 women (mean age 65 ⫾ 14 years). Sixty percent of patients had a prior myocardial infarction, mean Canadian Cardiovascular Society classification angina score was 2 ⫾ 1, and mean New York Heart Association heart failure score was 3 ⫾ 1. Cardiac catheterization revealed critical 3-vessel CAD in 26 patients, 2-vessel CAD in 6 patients, and mean ejection fraction of 0.24 ⫾ 0.08 by contrast left ventriculography. EMM and SPECT imaging characteristics: All 32 patients underwent EMM and SPECT imaging at rest, which yielded 448 matched myocardial segments for viability assessment. After automated stability criteria editing, 84 ⫾ 38 points were sampled per patient map. Volumetric measurements by EMM revealed an enddiastolic volume of 199 ⫾ 74 ml (end-diastolic volume index 101 ⫾ 38), end-systolic volume of 150 ⫾ 65 ml (end-systolic volume index 76 ⫾ 33), and LV ejection fraction of 0.26 ⫾ 0.08. Mapping time was 47 ⫾ 20 minutes. There were no complications related to the
CORONARY ARTERY DISEASE/ ELECTROMECHANICAL MAPPING FOR VIABILITY ASSESSMENT
809
TABLE 1 Electromechanical Data in 433 Myocardial Segments Divided According to Rest Function Defined by Electromechanical Mapping (EMM) and Viability Status According to Radionuclide Single-photon Emission Computed Tomographic (SPECT) Imaging
Normal (n ⫽ 152) Viable (n ⫽ 215) Nonviable (n ⫽ 66) p Value Overall analysis of variance Normal vs viable segments Normal vs nonviable segments Viable vs nonviable segments
Linear Shortening (%)
Unipolar Voltage (mV)
Normalized Unipolar Voltage (%)
Bipolar Voltage (mV)
13.6 ⫾ 3.6 3.6 ⫾ 3.9 2.4 ⫾ 4.2
10.4 ⫾ 4.4 8.7 ⫾ 4.1 6.8 ⫾ 3.2
67 ⫾ 23 58 ⫾ 24 48 ⫾ 25
3.8 ⫾ 2.7 3.2 ⫾ 2.6 2.3 ⫾ 1.9
⬍0.001 ⬍0.05 ⬍0.05 NS
⬍0.001 ⬍0.05 ⬍0.05 ⬍0.05
⬍0.001 ⬍0.05 ⬍0.05 ⬍0.05
⬍0.001 ⬍0.05 ⬍0.05 NS
TABLE 2 Electromechanical Data in 370 Myocardial Segments Divided According to Rest Function and Perfusion Status by Radionuclide Single-photon Emission Computed Tomographic (SPECT) Imaging Linear Shortening (%) Normal (n ⫽ 131) Reversible (n ⫽ 35) Fixed viable (n ⫽ 147) Nonviable (n ⫽ 57) p Value Overall analysis of variance Reversible vs nonviable segments Reversible vs fixed viable segments Fixed viable vs nonviable segments
13.6 3.7 3.5 2.2
⫾ ⫾ ⫾ ⫾
3.6 4.4 3.7 4.4
⬍0.001 ⬍0.05 NS ⬍0.05
10.4 11.1 8.2 6.8
⫾ ⫾ ⫾ ⫾
4.6 4.8 3.9 3.3
⬍0.001 ⬍0.05 ⬍0.05 ⬍0.05
EMM procedure. Of 448 matched myocardial segments, 433 segments had adequate EMM and SPECT data for analysis. Of the 433 segments, 152 had normal function and 281 were dysfunctional by EMM. Of the 264 myocardial segments with EMM and gated SPECT functional data, 170 (64.4%) were dysfunctional by EMM and 168 (63.6%) by gated SPECT (p ⫽ NS). Of the dysfunctional segments, 215 (77%) were viable and 66 (23%) were nonviable. In 28 patients who underwent SPECT imaging at rest and during stress, 370 of 392 myocardial segments had adequate EMM and SPECT data for analysis. Of the 370 segments, 131 had normal function and 239 were dysfunctional. Of the dysfunctional segments, 35 (15%) were reversible, 147 (61%) were fixed viable, and 57 (24%) were nonviable. EMM for myocardial viability assessment: There was a significant correlation between Tc-99m sestamibi uptake at rest and both unipolar voltage and normalized unipolar voltage for all myocardial segments (r ⫽ 0.34 and 0.31, respectively; p ⬍0.001 for both). This correlation was best at the apex for unipolar and normalized unipolar voltage (r ⫽ 0.58 and 0.52, respectively; p ⬍0.001 for both). No continuous correlation was demonstrated between tracer uptake at rest and either bipolar voltage or fragmentation. Of the dysfunctional segments, unipolar voltage, normalized unipolar voltage, and bipolar voltage were predictive of viable myocardium during SPECT imaging (p ⬍0.001 for each variable). Fragmentation did not predict viable myocardium (p ⫽ NS) and therefore was not included in further analysis. Table 1 lists values for unipolar voltage, normalized 810 THE AMERICAN JOURNAL OF CARDIOLOGY姞
Unipolar Voltage (mV)
VOL. 91
Normalized Unipolar Voltage (%) 67 65 57 48
⫾ ⫾ ⫾ ⫾
23 24 23 26
0.003 ⬍0.05 NS ⬍0.05
Bipolar Voltage (mV) 3.9 3.2 3.1 2.3
⫾ ⫾ ⫾ ⫾
2.7 2.6 2.2 2.0
0.052 NS NS ⬍0.001
unipolar voltage, and bipolar voltage in normal, viable, and nonviable segments. Unipolar voltage, normalized unipolar voltage, and bipolar voltage variables were significantly greater in normal than dysfunctional segments, whether the segments were viable or nonviable. Furthermore, unipolar voltage and normalized unipolar voltage were significantly higher in viable than in nonviable segments. The best EMM predictor of viability was a unipolar voltage of ⱖ7.1 mV, which yielded a sensitivity of 70%, specificity of 60%, and an area under the ROC curve of 0.69 ⫾ 0.04 (p ⬍0.001). A normalized unipolar voltage of ⱖ53% had optimal value for detecting viable myocardium, with a sensitivity of 64%, a specificity of 60%, and an area under the ROC curve of 0.65 ⫾ 0.04 (p ⬍0.001). Mean segmental unipolar voltage per patient correlated with global LV ejection fraction (r ⫽ 0.41, p ⬍0.001). EMM for inducible ischemia assessment: Table 2 lists electromechanical data in myocardial segments according to perfusion defects, with the normal segments as controls. Unipolar voltage was significantly greater in reversible than in fixed viable segments, and the 2 groups had higher unipolar voltage than nonviable segments. Normalized unipolar voltage was significantly higher in reversible than in nonviable segments; it was also higher in fixed viable segments than in nonviable segments. However, unlike unipolar voltage, normalized unipolar voltage was not significantly different between reversible and fixed viable segments. There were no significant differences in bipolar voltage between reversible and fixed viable segments (p ⫽ NS). Unipolar voltage was the best EMM predictor of APRIL 1, 2003
inducible ischemia on SPECT imaging, with a voltage of ⱖ8.5 mV having a sensitivity of 71%, a specificity of 60%, and an area under the ROC curve of 0.65 ⫾ 0.05 (p ⫽ 0.002). Normalized unipolar voltage only showed a trend toward predicting inducible ischemia (p ⫽ 0.052).
DISCUSSION This is the first study evaluating LV EMM for detection of myocardial viability and inducible ischemia, as defined by dipyridamole SPECT myocardial perfusion imaging, in patients with severe ischemic cardiomyopathy. A unipolar voltage of ⱖ7.1 mV and a normalized unipolar voltage of ⱖ53% had optimal discriminatory power for detecting viable myocardium. A unipolar voltage of ⱖ8.5 mV had the best predictive value for detecting inducible ischemia. EMM parameters for the detection of myocardial viability: We found that a unipolar voltage of 7.1 mV
was best able to discriminate viable from nonviable myocardium in dysfunctional myocardial segments in patients with profound LV dysfunction (LV ejection fraction 0.24 ⫾ 0.08). Our findings are consistent with those of other investigators who, in patients with less severe LV dysfunction, found a unipolar voltage of 6.5 mV to discriminate viable from nonviable myocardium as defined by positron emission tomography5,6 and thallium-201.7,8 Two of these studies7,8 excluded patients with a LV ejection fraction of ⬍0.30 (mean LV ejection fraction 0.40 ⫾ 0.48), and a third study6 was predominantly in patients with single-vessel CAD (LV ejection fraction 0.49 ⫾ 0.15). Therefore, our data extend the applicability of EMM to a patient population with profound ischemic LV dysfunction, in whom the assessment of myocardial viability is most relevant. The specificity of unipolar voltage for detecting viable myocardium is only fair (60%). This may reflect the limitation of Tc-99m sestamibi for detecting viable myocardium in the inferobasal segments due to diaphragmatic attenuation artifact.9 We found the correlation between unipolar voltage and Tc-99m sestamibi on SPECT to be weakest in the basal segments (r ⫽ 0.20, p ⫽ 0.001) where both technologies have limitations (fibrous mitral valve plane decreasing voltage on EMM and diaphragmatic attenuation causing artifact on SPECT), and best at the apex (r ⫽ 0.58, p ⬍0.001). Further studies investigating the predictive value of the 2 technologies in this patient population for recovery of ventricular function are warranted. We observed a continuous relation between mean segmental unipolar voltage per patient and global ventricular function, which suggests that patients with profound LV dysfunction had globally diminished voltage, even in myocardial segments remote from the demonstrable pathology by SPECT. This could reflect ventricular remodeling and may explain why absolute voltage was found to be a better discriminator of viability and ischemia than normalized voltage.
Unlike previous investigators,10 we found unipolar voltage to be a better indicator of viability than bipolar voltage. However, Wolf et al11 investigated a canine model of myocardial infarction where exhaustive and detailed sampling may have been more practical than in the clinical setting. In addition, unipolar voltage measures voltage in a wider field than bipolar voltage, thus mirroring the larger resolution of SPECT imaging, which we used as the “gold” standard. EMM parameters for the detection of myocardial ischemia: This is the first study demonstrating the
relation between preserved voltage and inducible ischemia in patients with severe ischemic LV dysfunction. Unipolar voltage was significantly higher in reversible viable than in fixed viable segments, with a unipolar voltage of ⱖ8.5 mV being the best discriminator of inducible ischemia. Fuchs et al8 found a similar unipolar voltage of 9.0 mV to discriminate inducible ischemia in patients with preserved ventricular function (LV ejection fraction ⫽ 0.48 ⫾ 0.11). Acknowledgment: The investigators thank Jennifer Hunter, RN, Linda Snyder, RN, and Katherine Bunger, RN, for their invaluable assistance, and Jerry Curtis, MA, for his superb editorial assistance.
1. Smith WH, Kastner RJ, Calnon DA, Segalla D, Beller GA, Watson DD. Quantitative gated SPECT imaging: a counts-based method for display and measurement of regional and global ventricular systolic function. J Nucl Cardiol 1997;4:451–463. 2. Kornowski R, Hong MK, Leon MB. Comparison between left ventricular electromechanical mapping and radionuclide perfusion imaging for detection of myocardial viability. Circulation 1998;98:1837–1841. 3. Kornowski R, Hong MK, Gepstein L, Goldstein S, Ellahham S, Ben-Haim SA, Leon MB. Preliminary animal and clinical experiences using an electromechanical endocardial mapping procedure to distinguish infarcted from healthy myocardium. Circulation 1998;98:1116 –1124. 4. Udelson JE, Coleman PS, Metherall J, Pandian NG, Gomez AR, Griffith JL, Shea NL, Oates E, Konstam MA. Predicting recovery of severe regional ventricular dysfunction. Comparison of resting scintigraphy with 201Tl and 99mTcsestamibi. Circulation 1994;89:2552–2561. 5. Botker HE, Lassen JF, Hermansen F, Wiggers H, Sogaard P, Kim WY, Bottcher M, Thuesen L, Pedersen AK. Electromechanical mapping for detection of myocardial viability in patients with ischemic cardiomyopathy. Circulation 2001;103:1631–1637. 6. Koch KC, vom Dahl J, Wenderdel M, Nowak B, Schaefer WM, Sasse A, Stellbrink C, Buell U, Hanrath P. Myocardial viability assessment by endocardial electroanatomic mapping: comparison with metabolic imaging and functional recovery after coronary revascularization. J Am Coll Cardiol 2001;38:91–98. 7. Gyongyosi M, Sochor H, Khorsand A, Gepstein L, Glogar D. Online myocardial viability assessment in the catheterization laboratory via NOGA electroanatomic mapping: quantitative comparison with thallium-201 uptake. Circulation 2001;104:1005–1011. 8. Fuchs S, Hendel RC, Baim DS, Moses JW, Pierre A, Laham RJ, Hong MK, Kuntz RE, Pietrusewicz M, Bonow RO, et al. Comparison of endocardial electromechanical mapping with radionuclide perfusion imaging to assess myocardial viability and severity of myocardial ischemia in angina pectoris. Am J Cardiol 2001;87:874 –880. 9. Soufer R, Dey HM, Ng CK, Zaret BL. Comparison of sestamibi single-photon emission computed tomography with positron emission tomography for estimating left ventricular myocardial viability. Am J Cardiol 1995;75:1214 –1219. 10. Dilsizian V, Bonow RO. Current diagnostic techniques of assessing myocardial viability in hibernating and stunned myocardium. Circulation 1993;87:1–20. 11. Wolf T, Gepstein L, Dror U, Hayam G, Shofti R, Zaretzky A, Uretzky G, Oron U, Ben-Haim SA. Detailed endocardial mapping accurately predicts the transmural extent of myocardial infarction. J Am Coll Cardiol 2001;37:1590 – 1597.
CORONARY ARTERY DISEASE/ ELECTROMECHANICAL MAPPING FOR VIABILITY ASSESSMENT
811