Usefulness of Phase Analysis to Differentiate Ischemic and ...

Circulation Journal Vol.78, January 2014

Circulation JournalOfficial Journal of the Japanese Circulation Societyhttp://www.j-circ.or.jp

n the management of patients with heart failure (HF) due to left ventricular (LV) systolic dysfunction, accurate di-agnosis of the underling etiology is of paramount impor-

tance.1–3 Among such patients, coronary artery disease (CAD) should be detected correctly because CAD is the most com-mon etiology and is treated effectively by coronary revascular-ization.4–6 In this regard, non-invasive examinations, such as stress myocardial single-photon emission computed tomography (SPECT) and cardiac magnetic resonance imaging (MRI), have been reported as being useful in differentiating between isch-emic and non-ischemic etiologies in HF patients.3,7–11

Recently, a novel technique has been developed to evaluate LV mechanical dyssynchrony using phase analysis on gated SPECT.12,13 Using this method, the indication for cardiac resyn-chronization therapy (CRT) may be assessed in patients with HF.14–16 No study, however, as far as we are aware, has at-tempted to analyze LV mechanical dyssynchrony on SPECT after stress and at rest to detect CAD as an etiology of HF due

to LV systolic dysfunction.17,18 Although stress myocardial SPECT with conventional perfusion analysis is useful to iden-tify CAD in HF patients,3,7–10 the addition of phase analysis to stress SPECT may yield not only diagnostic improvement, but also further clinical utility, such as LV dyssynchrony assess-ment. Thus, we retrospectively evaluated the usefulness of LV mechanical dyssynchrony as assessed on phase analysis com-pared with conventional gated SPECT to identify ischemic etiology in HF patients.

MethodsSubjectsThe subjects consisted of 41 consecutive patients (30 men and 11 women; mean age, 62±14 years) who were admitted to hospital due to an initial episode of congestive HF resulting from LV systolic dysfunction between January 2009 and May 2011. The diagnosis of congestive HF required definite evi-

I

Received February 7, 2013; revised manuscript received August 7, 2013; accepted September 2, 2013; released online October 30, 2013 Time for primary review: 11 days

Department of Cardiology, Tokyo Medical University, Tokyo, JapanMailing address: Yuko Igarashi, MD, Department of Cardiology, Tokyo Medical University, 6-7-1 Nishi-Shinjuku, Shinjuku-ku, Tokyo

160-0023, Japan. E-mail: [email protected] doi: 10.1253/circj.CJ-13-0137All rights are reserved to the Japanese Circulation Society. For permissions, please e-mail: [email protected]

Usefulness of Phase Analysis to Differentiate Ischemic and Non-Ischemic Etiologies of Left Ventricular Systolic

Dysfunction in Patients With Heart FailureYuko Igarashi, MD; Taishiro Chikamori, MD; Satoshi Hida, MD; Hirokazu Tanaka, MD;

Chie Shiba, MD; Yasuhiro Usui, MD; Tuguhisa Hatano, MD; Akira Yamashina, MD

Background: The detection of significant coronary artery disease (CAD) in patients with heart failure (HF) from left ventricular (LV) systolic dysfunction is crucial. We evaluated the usefulness of LV mechanical dyssynchrony as as-sessed by phase analysis compared with conventional gated single-photon emission computed tomography to identify ischemic etiology in patients with HF.

Methods and Results: Forty-one consecutive patients who were initially admitted to hospital due to HF resulting from systolic dysfunction were evaluated. All patients underwent cardiac catheterization. LV mechanical dyssyn-chrony was evaluated using SyncTool™ to obtain the phase SD and histogram bandwidth. The changes in phase SD and histogram bandwidth with stress were calculated. The summed stress score, summed difference score, and changes in phase SD and histogram bandwidth with stress were greater in 26 patients with CAD than in 15 patients without CAD (P=0.001 and P=0.01). On multivariate analysis a phase SD of >14° (odds ratio [OR], 16.7) and a summed stress score of >17 (OR, 8.0) best differentiated LV dysfunction of ischemic and non-ischemic etiologies, with a sensitivity of 89% and a specificity of 87% (χ2=20), compared with summed stress score only (sensitivity, 46%; specificity, 87%; χ2=4.5).

Conclusions: The addition of phase analysis to conventional perfusion analysis enables better differentiation of the etiology of HF in patients with systolic dysfunction. (Circ J 2014; 78: 141 – 150)

Key Words: Heart failure; Left ventricular dyssynchrony; Phase analysis; Quantitative gated single-photon emission computed tomography

ORIGINAL ARTICLEHeart Failure


142 IGARASHI Y et al.

image was divided into 17 segments.20 The radiopharmaceuti-cal accumulation in the myocardium was visually evaluated by 2 cardiologists who were blinded to clinical data using a 5-grade scale: 0, normal; 1, slight reduction of uptake; 2, mod-erate reduction of uptake; 3, severe reduction of uptake; or 4, absence of radioactive uptake. The total of the scores for all the segments during exercise and at rest was designated as the summed stress score (SSS) and the summed rest score (SRS), respectively. The summed difference score (SDS) was defined as SSS minus SRS.20 Disagreements in image interpretation were resolved by consensus after extensive discussion be-tween 2 of the 3 experts (S.H., Y.I., or T.C.).

Each reconstructed short-axis ECG-gated SPECT image was processed using the QGS program developed by Germano et al, to automatically calculate the LV end-diastolic volume (EDV), LV end-systolic volume (ESV) and LVEF.21 Changes in LV volume with ATP loading were calculated as EDV after stress minus EDV at rest, or ESV after stress minus ESV at rest. Each reconstructed short-axis ECG-gated SPECT image was pro-cessed, and LV functional parameters (LVEDV, LVESV, and LVEF) were automatically obtained as described by Germano et al.22 In addition, a 3-D count distribution was extracted from each of the LV short-axis datasets and subjected to a Fourier analysis, which generated a phase distribution (0–360°) span-ning the entire R-R interval and was displayed on a polar map and histogram (Figure 1). Two indices related to LV dyssyn-chrony were automatically extracted using SyncTool™ in con-junction with the Emory Cardiac Toolbox (Emory University/Syntermed, Atlanta, GA, USA).12 Manual base parameter place-ment was used to define the margin of the cardiac base only if low-frequency noise located at the basal end of the ventricles was observed.23 The indices used were histogram bandwidth, which marked the range of degrees of the cardiac cycle during which 95% of the myocardium initiated contraction, and phase SD, which represented the SD of the phase distribution (Figure 1). Furthermore, changes in phase SD and histogram bandwidth with stress were calculated as follows: phase SD after stress minus phase SD at rest or histogram bandwidth after stress minus histogram bandwidth at rest.19

Coronary AngiographyMultidirectional coronary angiography was performed once dur-

dence of pulmonary edema or congestion on chest X-ray film in addition to clinical symptoms. All of the patients underwent coronary angiography and pharmacologic stress myocardial perfusion imaging with phase analysis. LV systolic dysfunc-tion was defined as LV ejection fraction (EF) <40% on left ventriculogram or gated SPECT. Patients with first acute myo-cardial infarction diagnosed on clinical presentation, electro-cardiographic (ECG) changes or elevated plasma enzyme ac-tivity data, atrial fibrillation, valvular heart disease, or those receiving ventricular pacing or CRT, were excluded. Repeat admissions due to HF of known etiology, such as documented CAD or cardiomyopathies, were not included in this study. Written informed consent for invasive coronary angiography was obtained from all the participants. This retrospective study was approved by the Ethics Committee of Tokyo Medical University (No. 2123).

Stress Myocardial Perfusion ImagingStress myocardial perfusion imaging was performed using a 1-day protocol during the admission after HF improvement. Patients were requested not to consume caffeine for 12 h before the test. Adenosine triphosphate (ATP) (0.16 mg · kg−1 · min−1) was given i.v. for 6 min.19 Three minutes after the ATP, i.v. 99 mTc-sestamibi (259 MBq) was given. Imaging was started 30 min after the 99 mTc-sestamibi (259 MBq). After 4 h, the pa-tients were given 99 mTc-sestamibi (777 MBq) while at rest. Thirty minutes later, ECG-gated myocardial SPECT was per-formed.

Data were acquired with a 3-detector gamma camera (Prism 3000XP; Picker, Cleveland, OH, USA) in 360° arcs (in 6°-wide directions, taking 30 s per projection, 20 times). A low-energy high-resolution parallel multi-hole collimator was used. The matrix size of the projection images was 64×64. When obtaining ECG-gated images, the R-R interval was divided by the R wave trigger into 8 equal portions. End-diastolic and end-systolic myocardial perfusion images were obtained using this method. All patients were in sinus rhythm during image acquisition. SPECT images were reconstructed from the data with a data processor (Odyssey VP; Picker) combined with a Butterworth filter (order 8; cut-off frequency 0.25 cycles/pixel) and a ramp filter.

According to a previously described method, each SPECT

Figure 1. Representative phase polar map and phase histogram. 3-D count distribution was extracted from each of the left ven-tricular short-axis datasets and subjected to a Fourier analysis, which generated a phase distribution (0–360°) spanning the entire R-R interval. Phase SD or histogram bandwidth were automatically calculated. Standard deviation 62.5°; histogram bandwidth 238.0°.


143Phase Analysis in HF

to compare the means of continuous variables, and contingency tables were analyzed using the chi-squared test. Receiver oper-ating characteristic (ROC) curves were analyzed to determine the optimal cut-offs for SSS, SRS, SDS, changes in phase SD with stress, and histogram bandwidth with stress to distinguish patients with ischemic and non-ischemic heart disease. For these cut-off points, sensitivity and specificity were calculated using standard formulas. Univariate analysis was conducted

ing admission, within 2 weeks of scintigraphy in all patients, using the Judkins’ method. The degree of coronary artery ste-nosis was visually rated according to the criteria of the Ameri-can Heart Association.24 Significant stenosis was considered to be present when ≥75% narrowing of the diameter was observed.

Statistical AnalysisResults are expressed as mean ± SD. Student’s t-test was used

Table 1. Patient Characteristics vs. Presence of CAD

CAD (+) (n=26)

CAD (–) (n=15) P-value

Age (years) 65±13 56±15 NS

Men 20 (77) 10 (67) NS

Hypertension 20 (77) 10 (67) NS

Hyperlipidemia 15 (58) 4 (27) NS

Diabetes mellitus 16 (62) 8 (53) NS

Smoking 20 (77) 11 (73) NS

Medication

Statin 14 (54) 4 (27) NS

Calcium-channel blocker 5 (19) 2 (13) NS

ACEI/ARB 22 (85) 14 (93) NS

β-blocker 18 (69) 12 (80) NS

Diuretics 24 (92) 13 (87) NS

Warfarin 9 (35) 4 (27) NS

QRS duration (ms) 111±20 122±25 NS

Coronary angiography

1-vessel CAD 0 N/A

2-vessel CAD 8 (31) N/A

3-vessel CAD 18 (69) N/A

Myocardial perfusion

SSS 18.9±9.1 13.1±8.6 0.05　　 SRS 15.4±8.9 11.9±9.7 NS

SDS 3.5±2.8 1.2±2.5 0.01　　Functional analysis at rest

EDV (ml) 238±111 236±97 NS

ESV (ml) 161±84 159±81 NS

EF (%) 33±8 29±8 NS

Functional analysis after stress

EDV (ml) 244±108 252±98 NS

ESV (ml) 175±90 172±59 NS

EF (%) 30±8 31±10 NS

Post-stress increase

EDV (ml) 6±22 16±21 NS

ESV (ml) 14±23 13±35 NS

EF (%) –3±8 2±6 NS

Phase analysis at rest

Phase SD (°) 50.0±18.0 52.5±13.2 NS

Histogram bandwidth (°) 155.0±61.2 180.8±64.8 NS

Phase analysis after stress

Phase SD (°) 68.5±18.2 60.0±13.0 NS

Histogram bandwidth (°) 207.2±58.3 196.5±55.6 NS

Post-stress increase

Phase SD (°) 18.0±10.9 7.3±6.1 0.001

Histogram bandwidth (°) 52.4±46.1 15.7±37.2 0.01　　

Data given as mean ± SD or n (%).ACEI, angiotensin-converting enzyme-inhibitor; ARB, angiotensin receptor blocker; CAD, coronary artery disease; EDV, end-diastolic volume; EF, ejection fraction; ESV, end-systolic volume; SDS, summed difference score; SRS, summed rest score; SSS, summed stress score.



tically significant difference. Statistical computations were per-formed using SPSS 11.0 (SPSS, Chicago, IL, USA) and MedCalc 11.4 (MedCalc Software, Mariakerke, Belgium).

ResultsClinical CharacteristicsThe mean age of the 41 patients was 62±14 years; there were 30 men and 11 women. No significant differences were ob-served in the prevalence of coronary risk factors and medica-tions used before SPECT between the 26 patients with CAD

with the logistic regression method, and stepwise multivariate analysis was conducted with the multiple logistic regression method using statistically significant variables in the univariate analysis. We performed linear discriminant analysis with step-wise variable selection with Wilks’ lambda, which is the ratio of the within-groups sum of squares to the total sum of squares, to assess the potential to correctly identify ischemic etiology in HF patients, using independent variables in the multivariate analysis. A Bayes rule with equal prior probability was used for identification, and results are presented as sensitivity, specific-ity and accuracy. P<0.05 was considered to represent a statis-

Figure 2. Cut-off points for perfusion parameters to identify ischemic etiology in patients with heart failure were defined as summed stress score >17, summed rest score >3, and summed difference score >0. The area under the curve (AUC) was 0.68 for the summed stress score, 0.61 for the summed rest score, and 0.74 for the summed difference score.



ans for SSS, SRS, and SDS were 9, 3, and 5, respectively. SSS (18.9±9.1 vs. 13.1±8.6; P=0.05) and SDS (3.5±2.8 vs. 1.2±2.5; P=0.01) were significantly greater in the 26 patients with CAD than in the 15 patients without CAD (Table 1).

LVEDV, LVESV, and LVEF at rest and after stress, or post-stress increases were not significantly different between the 26 patients with CAD and the 15 patients without CAD.

In all of the 41 patients, the means of phase SD and histo-gram bandwidth at rest were 50.9±16.2° and 164.5±63.0°, respectively, and after ATP loading the increases in phase SD and histogram bandwidth were 14.1±10.7° and 39.0±46.2°, respectively. Based on the indices of LV dyssynchrony for the CRT response defined as a phase SD >43° by Henneman et al,14 30 patients were considered as responders. Similarly, judging from a histogram bandwidth >135° which was defined as CRT response, 28 patients were considered as responders.14 In addition, 26 patients fulfilled both criteria for the indices of LV dyssynchrony for the CRT response.

Among 13 patients with a QRS duration of ≥120 ms, 12 patients had the indices of LV dyssynchrony fulfilling the CRT response defined by either criteria mentioned here. Among 28 patients with a QRS duration of <120 ms, 18 patients had the indices of LV dyssynchrony for the CRT response defined as

and the 15 patients without CAD (Table 1). The average LVEDV, LVESV, and LVEF were 237±105 ml, 161±82 ml, and 32±8%, respectively, among the total 41 patients. The mean QRS duration was 115±22 ms. No significant difference was found in the QRS duration between ischemic patients and non-ischemic patients (111±20 vs. 122±25 ms; P=NS). A QRS duration of ≥120 ms was observed in 13 patients and a QRS duration of <120 ms in 28 patients. Left bundle branch block was observed in 7 patients; 4 patients had non-ischemic car-diomyopathy and 3 patients had ischemic cardiomyopathy. Right bundle branch block was observed in only 1 patient. All patients underwent coronary angiography; 2-vessel CAD was found in 8 patients (1patient had LMT lesion) and 3-vessel CAD in 18, (4 patients had LMT lesion), and an insignificant lesion in the remaining 15 patients (Table 1). Among 26 pa-tients with ischemic LV dysfunction, percutaneous coronary intervention was performed in 16 patients, coronary artery bypass grafting in 8 patients, and the remaining 2 patients were treated medically after the final diagnosis based on the gated SPECT and coronary angiography.

ScintigraphyAmong the conventional scintigraphic parameters, the medi-

Table 2. Scintigraphy Parameters and Detection of Ischemic Etiology in HF

Sensitivity (%) Specificity (%) Accuracy (%)

SSS >17 46 87 61

SRS >3 100 27 73

SDS >0 89 53 76

Increase in phase SD >14° 65 93 76

Increase in bandwidth >24° 69 67 68

HF, heart failure. Other abbreviations as in Table 1.

Figure 3. Cut-off points for increase in phase SD and increase in histogram bandwidth to detect ischemic etiology in patients with heart failure were defined as increase in phase SD >14° and increase in histogram bandwidth >24°. The area under the curve (AUC) was 0.77 for the increase in phase SD and 0.74 for histogram bandwidth.



Univariate Analysis for Detection of Ischemic Etiology in HFTo detect ischemic etiology in patients with HF using myocar-dial perfusion analysis, we applied ROC curve analysis to this group of patients. The cut-off points for the detection of isch-emic etiology in patients with HF were >17 for SSS, >3 for SRS, and >0 for SDS (Figure 2). The respective sensitivities, specificities, and accuracies in the detection of ischemic etiol-ogy in patients with HF were 46%, 87%, and 61% with SSS, 100%, 27%, and 73% with SRS, and 89%, 53%, and 76% with SDS, respectively (Table 2). ROC curve analysis demon-strated that changes in the indices of LV dyssynchrony could

phase SD >43°, or 17 patients had a histogram bandwidth >135°. In addition, 15 patients fulfilled both criteria for the indices of LV dyssynchrony for CRT response. No significant differences were observed in the histogram bandwidth at rest and after stress between the patients with CAD and the patients without CAD. The increases in phase SD and the increases in histogram bandwidth with stress, however, were greater in the 26 patients with CAD than in the 15 patients without CAD (18.0±10.9° vs. 7.3±6.1°, P=0.001 and 52.4±46.1° vs. 15.7± 37.2°, P=0.01; Table 1).

Table 3. Detection of Ischemic Etiology in HF: Univariate and Multivariate Analysis

VariablesUnivariate Analysis Multivariate Analysis

OR (95% CI) P-value OR (95% CI) P-value

Age (years) 1.1 (1.0–1.1)　　 NS

Men 2.8 (0.5–15.0) NS

Hypertension 1.7 (0.4–6.8)　　 NS

Hyperlipidemia 3.8 (0.9–15.0) NS

Diabetes mellitus 1.4 (0.4–5.1)　　 NS

Smoking 1.2 (0.3–5.2)　　 NS

QRS duration (ms) 1.0 (0.9–1.0)　　 NS

Post-stress increase in EDV 1.0 (0.9–1.0)　　 NS

Post-stress increase in ESV 1.0 (0.9–1.0)　　 NS

Post-stress increase in EF 0.9 (0.8–1.0)　　 NS

SSS >17 5.6 (1.0–29.8) 0.05　　 8.0 (1.0–64.2) 0.05

SDS >0 8.8 (1.8–42.3) 0.007 3.6 (0.5–27.8) 0.2　　Increase in histogram bandwidth >24° 4.5 (1.2–17.5) 0.03　　 2.2 (0.3–13.6) 0.4　　Increase in phase SD >14° 26.4 (3.0–234.4) 0.003 16.7 (1.5–188.2) 0.02

CI, confidence interval; OR, odds ratio. Other abbreviations as in Tables 1,2.

Figure 4. (A) Comparison of diagnostic value with the addition of phase analysis in the detection of ischemic etiology in patients with heart failure (HF). (B) Improved diagnostic value with the addition of phase analysis to perfusion analysis in the detection of ischemic etiology in patients with HF.



Figure 5. (A) Representative images of perfusion and phase analysis in a 62-year-old woman with coronary artery disease (CAD). Coronary angiogram showed 3-vessel CAD with severe narrowing of the right coronary artery and left circumflex artery, and inter-mediate stenosis of the left anterior descending coronary artery. Fixed perfusion defects can be observed in the apex and infero-postero-lateral segments, with summed stress score (SSS), summed rest score (SRS), and summed difference score (SDS) of 31, 31, and 0, respectively. Phase distribution images (Left, after ATP stress; Right, at rest) showed that the increase in phase SD after ATP stress was 15.9° and the increase in histogram bandwidth was 83°. (B) Representative images of perfusion and phase analysis in a 63-year-old woman without CAD. Reversible perfusion defects can be observed in the apex and infero-posterior segments, with SSS, SRS, and SDS of 15, 10, and 5, respectively. Phase distribution images (Left, after ATP stress; Right, at rest) showed that the increase in phase SD after ATP stress was 2.8° and the increase in histogram bandwidth was 9°.



the etiology of HF became very low. These reports are consis-tent with the present results in that the sensitivity of ATP myocardial SPECT to detect CAD was >89%, either with a significant defect at rest or significant ischemia on myocardial SPECT (Table 2). The previous studies and the present one indicate that myocardial SPECT is sufficiently sensitive to detect CAD in HF patients with LV systolic dysfunction be-cause these patients usually have a large myocardial infarction with subsequent remodeling or a combination of a small infarc-tion and extensive ischemia and/or myocardial hibernation.36

Previous observations and the present ones show that the specificity of perfusion imaging for CAD is modest,3,7–10 given that even in non-ischemic dilated cardiomyopathy there may be significant territories of fibrosis.37 The indices of LV dys-synchrony, however, improved diagnostic specificity, and the combination of a noteworthy SSS and a significant increase in phase SD after ATP loading was most important in detecting CAD in the present HF patients.

Recently, the pattern of late gadolinium enhancement on MRI has been shown to be a powerful tool to detect CAD in HF patients because myocardial infarction almost always starts from the subendocardium.38 In contrast, late gadolinium enhancement is usually observed in the mid-layer in non-ischemic dilated cardiomyopathy.38 The low spatial resolution of myocardial SPECT compared with MRI hampers the diag-nostic utility of this non-invasive method for the evaluation of HF patients. The present novel approach, however, of adding phase analysis to stress SPECT may yield not only diagnostic improvement, but also further clinical utility to automatically obtain the important indices of LV dyssynchrony, which may influence the subsequent management of HF patients, includ-ing CRT.

LV Mechanical Dyssynchrony Index Changes Due to CAD in HFTwo different mechanisms may generate LV dyssynchrony: temporal delay and contractile disparity.39 In contrast to pa-tients with temporal delay due to a wide QRS, LV regional disparities in contractility are regarded as a predominant mech-anism for dyssynchrony in CAD patients.39,40 When activation occurs, one part of the LV wall, which has normal myocardi-um under a normal coronary artery, contracts more strongly than the other with myocardial ischemia or fibrosis. This strong part of the LV wall pushes out the weaker part, which contracts in early diastole.39,41,42 Delayed contraction of the weak wall, imposed by exercise-induced ischemia or stunning, is thought to be shown by phase SD using SyncTool™ as de-layed onset of mechanical contraction.12,39 Thus, if the extent of ischemic myocardium were greater, the indices of LV me-chanical dyssynchrony would also become greater, particu-larly in patients with extensive CAD, such as those in the current study.40

In contrast to exercise stress, although ATP usually does not induce actual ischemia, an increased myocardial blood flow, which is induced by pharmacologic vasodilation in a normal vascular territory, results in increased myocardial con-traction.19,43 Therefore, a difference in contractility may occur after ATP loading, particularly in patients with severe LV dysfunction due to multi-vessel CAD or non-ischemic cardio-myopathy.

Clinical ImplicationsThe uniqueness of the present non-invasive approach (using phase analysis for patients with HF due to systolic dysfunction) is its diagnostic value, not only for detecting HF etiology but

be considered to indicate ischemic etiology if a >14° increase in phase SD and a >24° increase in histogram bandwidth were observed after exercise (Figure 3). The sensitivities, speci-ficities, and accuracies in the detection of ischemic etiology in patients with HF were 65%, 93%, and 76% with changes in phase SD, and 69%, 67%, and 68% with changes in histogram bandwidth, respectively (Table 2).

Multivariate Analysis for Detection of Ischemic Etiology in HFTo detect ischemic etiology in patients with HF, we performed logistic regression analysis by entering 4 variables that were statistically significant on univariate analysis (Table 3). The 2 most important variables in the identification of ischemic eti-ology in patients with HF were an increase in phase SD >14° after ATP loading and SSS >17 (Table 3). Linear discriminant analysis using SSS yielded 46% sensitivity, 87% specificity, and 61% accuracy (global χ2=4.5; Figure 4). We repeated discriminant analysis using SSS and mechanical dyssynchro-ny indices, which were statistically significant on multivariate analysis. This showed that the combination of an increase in phase SD after ATP loading and SSS best identified ischemic etiology in patients with HF, with a higher sensitivity of 89% and a similar specificity of 87% and an accuracy of 88% (global χ2=20.0), than SSS alone (Figure 4). Typical cases are shown in Figure 5.

DiscussionImportance of CAD in HF With LV Systolic DysfunctionHF morbidity and mortality have been increasing.25–27 Al-though the treatment for HF due to LV diastolic dysfunction remains unclear,28–30 that for HF due to LV systolic dysfunc-tion has been reported.6,31–35 We elucidated whether indices of LV mechanical dyssynchrony assessed on phase analysis using SyncTool™ have greater diagnostic value than conven-tional perfusion analysis for CAD detection.

Main FindingsThe indices of LV mechanical dyssynchrony had greater diag-nostic value than clinical and perfusion variables in the detec-tion of CAD in HF patients. Compared with the indices of LV mechanical dyssynchrony at rest, a >14° increase in phase SD or a >24° increase in bandwidth after ATP loading had a sen-sitivity of 65% to 69% and a specificity of 93% to 67% in CAD detection in 41 patients with a first episode of HF due to LV systolic dysfunction. On multivariate analysis, the combination of an increase in phase SD >14° after ATP loading and SSS >17 best identified CAD as an etiology of LV systolic dysfunc-tion, with a higher sensitivity of 89% and a similar specificity of 87%, compared with SSS alone (sensitivity, 46%; specific-ity, 87%). Although previous studies reported the usefulness of stress myocardial scintigraphy with conventional perfusion analysis in CAD detection in HF patients,3,7–10 the present study has shown that the novel addition of phase analysis to stress SPECT could yield increased diagnostic value. Among potential confounding factors for LV dyssynchrony, reduced LVEF had no effect on these indices.

Comparison With Previous StudiesMyocardial perfusion imaging may be regarded as an alterna-tive initial step to coronary angiography for CAD detection in HF patients.3,7–10 Using planar thallium imaging, 6 studies reported a very high sensitivity of these imaging techniques to detect CAD.36 These observations indicated that if the perfu-sion image showed a normal pattern, the likelihood of CAD as



References 1. Soman P, Lahiri A, Mieres JH, Calnon DA, Wolinsky D, Beller GA,

et al. Etiology and pathophysiology of new-onset heart failure: Eval-uation by myocardial perfusion imaging. J Nucl Cardiol 2009; 16: 82 – 91.

2. Yatteau RF, Peter RH, Behar VS, Bartel AG, Rosati RA, Kong Y. Ischemic cardiomyopathy: The myopathy of coronary artery disease. Natural history and results of medical versus surgical treatment. Am J Cardiol 1974; 34: 520 – 525.

3. Dunn RF, Uren RF, Sadick N, Bautovich G, McLaughlin A, Hiroe M, et al. Comparison of thallium-201 scanning in idiopathic dilated cardiomyopathy and severe coronary artery disease. Circulation 1982; 66: 804 – 810.

4. Elefteriades JA, Tolis G Jr, Levi E, Mills LK, Zaret BL. Coronary artery bypass grafting in severe left ventricular dysfunction: Excel-lent survival with improved ejection fraction and functional state. J Am Coll Cardiol 1993; 22: 1411 – 1417.

5. Ragosta M, Beller GA, Watson DD, Kaul S, Gimple LW. Quantita-tive planar rest-redistribution 201Tl imaging in detection of myocar-dial viability and prediction of improvement in left ventricular func-tion after coronary bypass surgery inpatients with severely depressed left ventricular function. Circulation 1993; 87: 1630 – 1641.

6. Velazquez EJ, Lee KL, Deja MA, Jain A, Sopko G, Marchenko A, et al. Coronary-artery bypass surgery in patients with left ventricular dysfunction. N Engl J Med 2011; 364: 1607 – 1616.

7. Bulkley BH, Hutchins GM, Bailey I, Strauss HW, Pitt B. Thallium 201 imaging and gated cardiac blood pool scans in patients with isch-emic and idiopathic congestive cardiomyopathy. A clinical and patho-logic study. Circulation 1977; 55: 753 – 760.

8. Saltissi S, Hockings B, Croft DN, Webb-Peploe MM. Thallium-201 myocardial imaging in patients with dilated and ischemic cardiomy-opathy. Br Heart J 1981; 46: 290 – 295.

9. Tauberg SG, Orie JE, Bartlett BE, Cottington EM, Flores AR. Use-fulness of thallium-201 for distinction of ischemic from idiopathic dilated cardiomyopathy. Am J Cardiol 1993; 71: 674 – 680.

10. Chikamori T, Doi YL, Yonezawa Y, Yamada M, Seo H, Ozawa T. Value of dipyridamole thallium-201 imaging in noninvasive differ-entiation of idiopathic dilated cardiomyopathy from coronary artery disease with left ventricular dysfunction. Am J Cardiol 1992; 69: 650 – 653.

11. McCrohon JA, Moon JCC, Prasad SK, McKenna WJ, Lorenz CH, Coats AJS, et al. Differentiation of heart failure related to dilated car-diomyopathy and coronary artery disease using gadolinium-enhanced cardiovascular magnetic resonance. Circulation 2003; 108: 54 – 59.

12. Chen J, Garcia EV, Folks RD, Cooke CD, Faber TL, Tauxe EL, et al. Onset of left ventricular mechanical contraction as determined by phase analysis of ECG-gated myocardial perfusion SPECT imaging: Development of a diagnostic tool for assessment of cardiac me-chanical dyssynchrony. J Nucl Cardiol 2005; 12: 687 – 695.

13. Garcia EV, Faber TL, Cooke CD, Folks RD, Chen J, Santana C. The increasing role of quantification in clinical nuclear cardiology: The Emory approach. J Nucl Cardiol 2007; 14: 420 – 432.

14. Henneman MM, Chen J, Dibbets-Schneider P, Stokkel MP, Bleeker GB, Ypenburg C, et al. Can LV dyssynchrony as assessed with phase analysis on gated myocardial perfusion SPECT predict response to CRT? J Nucl Med 2007; 48: 1104 – 1111.

15. Henneman MM, Chen J, Ypenburg C, Dibbets P, Bleeker GB, Boersma E, et al. Phase analysis of gated myocardial perfusion sin-gle-photon emission computed tomography compared with tissue Doppler imaging for the assessment of left ventricular dyssynchrony. J Am Coll Cardiol 2007; 49: 1708 – 1714.

16. Pazhenkottil AP, Buechel RR, Husmann L, Nkoulou RN, Wolfrum M, Ghadri JR, et al. Long-term prognostic value of left ventricular dyssynchrony assessment by phase analysis from myocardial perfu-sion imaging. Heart 2011; 97: 33 – 37.

17. Aljaroudi W, Koneru J, Heo J, Iskandrian AE. Impact of ischemia on left ventricular dyssynchrony by phase analysis of gated single pho-ton emission computed tomography myocardial perfusion imaging. J Nucl Cardiol 2011; 18: 36 – 42.

18. Horigome M, Yamazaki K, Ikeda U. Assessment of left ventricular dyssynchrony in patients with coronary artery disease during adenos-ine stress using ECG-gated myocardial perfusion single-photon emis-sion computed tomography. Nucl Med Commun 2010; 31: 864 – 873.

19. Hida S, Chikamori T, Tanaka H, Igarashi Y, Shiba C, Hatano T, et al. Postischemic myocardial stunning is superior to transient ischemic dilation for detecting multivessel coronary artery disease. Circ J 2012; 76: 430 – 438.

20. Igarashi Y, Chikamori T, Hida S, Tanaka H, Shiba C, Usui Y, et al. Importance of the ankle-brachial pressure index in the diagnosis of

also for automatically obtaining the important indices related to LV dyssynchrony. Currently, the standard indications for CRT in HF patients were clinical symptoms and a QRS dura-tion ≥120 ms.14–16,34,35 Based on the ECG criterion, 13 patients (32%) fulfilled this in the present study, whereas 26 patients (63%) were considered as CRT responders according to the phase analysis.14 Of 28 patients (54%) with a QRS <120 ms, 15 patients were judged as CRT responders based on the indices of LV dyssynchrony as assessed on phase analysis on myocar-dial SPECT. Thus, the present methodology may be consid-ered sufficiently sensitive for dyssynchrony assessment. This novel approach is clinically useful in the simultaneous detec-tion of HF etiology and patients requiring CRT.

Study LimitationsThis study was performed retrospectively in a relatively small number of HF patients. Long-term follow-up is required to determine whether the indices of LV dyssynchrony obtained at the time of initial diagnosis can play a role in identifying who may benefit from CRT. Therefore, a prospective ap-proach that applies stress myocardial SPECT with phase anal-ysis and invasive coronary angiography in a large HF patient group is necessary.

In this study, we used specific software to evaluate LV dys-synchrony (SyncTool™ in conjunction with the Emory Car-diac Toolbox), and did not compare the results with those ob-tained using other software. Therefore, caution should be taken when applying the present results to a similar group of patients using a different software program. Given that the diagnostic method used in the present study was limited to nuclear cardi-ology, the comparison of different modalities such as MRI and echocardiography may be necessary.

In this study, the phase analysis data were acquired at 8 frames per cardiac cycle. Higher temporal resolution may the-oretically be obtained by acquisition at 16 frames per cardiac cycle, but the current method was developed by Chen et al in the knowledge that the first-harmonic Fourier transformation can enhance automatic phase analysis when applied to the data at 8 frames per cardiac cycle because of greater count densi-ty.44,45 In addition, previous studies that obtained a high degree of intra-observer and inter-observer reproducibility of the indi-ces of LV mechanical dyssynchrony, have demonstrated the feasibility of this method.23,40

Finally, the change in LV regional wall motion after ATP stress was not evaluated in the present study. Although previ-ous studies compared post-stress LV regional wall motion change between exercise and ATP stress, post-stress myocar-dial change was rarely observed after ATP stress.46 Whether a sophisticated nuclear cardiology imaging method may have an impact on the detection of a small change in LV regional wall motion and provide further diagnostic utility such as in isch-emic cardiomyopathy, however, remains to be determined.

ConclusionsThe addition of phase analysis to conventional perfusion anal-ysis on ATP myocardial SPECT enables better differentiation of the etiology of LV systolic dysfunction in HF patients.

AcknowledgmentsWe are indebted to Associate Professor Edward F. Barroga (DVM, PhD) of the Department of International Medical Communications at Tokyo Medical University for editorial review of the English manuscript.



34. Tang AS, Wells GA, Talajic M, Arnord MO, Sheldon R, Connolly S, et al. Cardiac-resynchronization therapy for mild-to-moderate heart failure events. N Eng J Med 2010; 363: 2385 – 2395.

35. Moss AJ, Hall WJ, Cannom DS, Klein H, Brown MW, Daubert JP, et al. Cardiac-resynchronization therapy for the prevention of heart failure events. N Eng J Med 2009; 361: 1329 – 1338.

36. Udelson JE, Shafer CD, Carrio I. Radionuclide imaging in heart fail-ure: Assessing etiology and outcomes and implications for manage-ment. J Nucl Cardiol 2002; 9: S40 – S52.

37. Assomull RG, Prasad SK, Lyne J, Smith G, Burman ED, Khan M, et al. Cardiovascular magnetic resonance, fibrosis, and prognosis in dilated cardiomyopathy. J Am Coll Cardiol 2006; 48: 1977 – 1985.

38. Karamitsos TD, Francis JM, Myerson S, Selvanayagam JB, Neubauer S. The role of cardiovascular magnetic resonance imaging in heart failure. J Am Coll Cardiol 2009; 54: 407 – 424.

39. Kass DA. An epidemic of dyssynchrony: But what does it mean? J Am Coll Cardiol 2008; 51: 12 – 17.

40. Hida S, Chikamori T, Tanaka H, Igarashi Y, Shiba C, Usui Y, et al. Diagnostic value of left ventricular dyssynchrony after exercise and at rest in the detection of multivessel coronary artery disease on single-photon emission computed tomography. Circ J 2012; 76: 1942 – 1952.

41. Skulstad H, Edvardsen T, Urheim S, Rabben SI, Stugaard M, Lyseggen E, et al. Postsystolic shortening in ischemic myocardium: Active contraction or passive recoil? Circulation 2002; 106: 718 – 724.

42. Pislaru C, Anagnostopoulos PC, Seward JB, Greenleaf JF, Belohlavek M. Higher myocardial strain rates during isovolumic relaxation phase than during ejection characterize acutely ischemic myocardium. J Am Coll Cardiol 2002; 40: 1487 – 1494.

43. Chikamori T, Doi YL, Seo H, Kawamoto A, Akagi N, Maeda T, et al. Effects of dipyridamole on left ventricular systolic and diastolic function in healthy young and elderly subjects as assessed by radio-nuclide angiography. Am J Cardiol 1994; 73: 1024 – 1029.

44. Cooke CD, Garcia EV, Cullom SJ, Faber TL, Pettigrew RI. Deter-mining the accuracy of calculating systolic wall thickening using a fast Fourier transform approximation: A simulation study based on canine and patient data. J Nucl Med 1994; 35: 1185 – 1192.

45. Chen J, Henneman MM, Trimble MA, Bax JJ, Borges-Neto S, Iskandrian AE, et al. Assessment of left ventricular mechanical dys-synchrony by phase analysis of ECG-gated SPECT myocardial per-fusion imaging. J Nucl Cardiol 2008; 15: 127 – 136.

46. Tanaka H, Chikamori T, Hida S, Usui Y, Harafuji K, Igarashi Y, et al. Comparison of post-exercise and post-vasodilator stress myocardial stunning as assessed by electrocardiogram-gated single-photon emis-sion computed tomography. Circ J 2005; 69: 1338 – 1345.

coronary artery disease in women with diabetes without anginal pain. Circ J 2011; 75: 2206 – 2212.

21. Germano G, Kavanagh PB, Waechter P, Areeda J, Van Kriekinge S, Sharir T, et al. A new algorithm for the quantitation of myocardial perfusion SPECT. I: Technical principles and reproducibility. J Nucl Med 2000; 41: 712 – 719.

22. Germano G, Kiat H, Kavanagh PB, Moriel M, Mazzanti M, Su HT, et al. Automatic quantification of ejection fraction from gated myo-cardial perfusion SPECT. J Nucl Med 1995; 36: 2138 – 2147.

23. Trimble MA, Velazquez EJ, Adams GL, Honeycutt EF, Pagnanelli RA, Barnhart HX, et al. Repeatability and reproducibility of phase analysis of gated single-photon emission computed tomography myo-cardial perfusion imaging used to quantify cardiac dyssynchrony. Nucl Med Commun 2008; 29: 374 – 381.

24. AHA committee report. A reporting system on patients evaluated for coronary artery disease. Circulation 1975; 51: 5 – 34.

25. Shah RU, Tsai V, Klein L, Heidenreich PA. Characteristics and out-comes of very elderly patients after first hospitalization for heart failure. Circ Heart Fail 2011; 4: 301 – 307.

26. Fang J, Mensah GA, Croft JB, Keenan NL. Heart failure-related hospitalization in the U.S, 1979 to 2004. J Am Coll Cardiol 2008; 52: 428 – 434.

27. Haldeman GA, Croft JB, Giles WH, Rashidee A. Hospitalization of patients with heart failure: National hospital discharge survey, 1985 to 1995. Am Heart J 1999; 137: 352 – 360.

28. Vasan RS, Levy D. Defining diastolic heart failure. Circulation 2000; 101: 2118 – 2121.

29. Zile MR, Gaasch WH, Carroll JD, Feldman MD, Aurigemma GP, Schaer GL, et al. Heart failure with a normal ejection fraction. Is measurement of diastolic function necessary to make the diagnosis of diastolic heart failure? Circulation 2001; 104: 779 – 782.

30. Massie BM, Carson PE, McMurray JJ, Komajda M, McKelvie R, Zile MR, et al. Irbesartan in patients with heart failure and preserved ejection fraction. N Engl J Med 2008; 359: 2456 – 2467.

31. The CONSENSUS Trial Study Group. Effects of enalapril on mor-tality in severe congestive heart failure: Results of the cooperative north Scandinavian enalapril survival study (CONSENSUS). N Engl J Med 1987; 316: 1429 – 1435.

32. The SOLVD Investigators. Effects of enalapril on survival in patients with reduced left ventricular ejection fraction and congestive heart failure. N Engl J Med 1991; 325: 293 – 302.

33. Packer MP, Bristow MR, Cohn JN, Colucci WS, Fowler MB, Edward BS, et al. The effect of carvedilol on morbidity and mortality in pa-tients with chronic heart failure. N Engl J Med 1996; 334: 1349 – 1355.

Usefulness of Phase Analysis to Differentiate Ischemic and ...

Documents

Transcript of Usefulness of Phase Analysis to Differentiate Ischemic and ...