Nuclear tomographic phase analysis: Localization of accessory conduction pathway in patients with Wolff-Parkinson-White syndrome

Nuclear tomographic phase analysis: Localization of accessory conduction in patients with Wolff-Parkinson-White syndrome

pathway

The purpose of this study was to evaluate the usefulness of tomographic phase analysis in detecting the site of the accessory conduction pathway (ACP) in patients with Wolff-Parkinson-White (WPW) syndrome. Gated emission computed tomography and planar gated blood pool scintigraphy were performed in 20 patients with WPW syndrome, 14 with delta waves and six without delta waves (two intermittent types and four concealed types). The abnormal initial contractions in both planar and tomographic phase images were compared with the sites of ACPs confirmed by epicardial mapping and surgery. The atrioventricular ring was divided into eight segments on each side, and the identification of the initial phase in the segment in which the ACP was located, or that adjacent to it, was considered to be the correct diagnosis. In planar phase analysis, the abnormal initial phase was identified correctly in 8 of 14 patients (57%), whereas in tomographic phase analysis, the site of the ACP was detected in 12 of 14 patients (86%). Tomographic phase analysis can be a helpful adjunctive method in patients with WPW syndrome. (AM HEART J lOg:8Og, 1985.)

Kenichi Nakajima, M.D., Hisashi Bunko, M.D., Akira Tada, M.D., Norihisa Tonami, M.D., Kinichi Hisada, M.D., Takuro Misaki, M.D., and Takashi Iwa, M.D. Kanazawa, Japan

The recent development of computer technology has made possible the detection of new parameters’ of cardiac function in gated blood pool studies.‘x2 Phase imaging derived from gated cardiac studies is a functional imaging of the timing of contractions in cardiac chambers and has been applied to conduction anomalies, such as bundle branch block and Wolff-Parkinson-White (WPW) syndrome.3-12 Generally, there has been relatively good correlation between conduction and abnormal wall motion. In patients with WPW syndrome, correlation between the site of initial activation and phase analysis has been studied.6-‘1 However, in our experience, based on a comparison with surgically proved accessory conduction pathway (ACP), there was a high frequency of correct identification of the side of preexcitation, but the ability to precisely pinpoint the From the Department of Nuclear Medicine and the First Surgery, School of Medicine, Kanazawa University. Received accepted

for publication Nov. 16, 1984.

Reprint requests: Kenichi tine, School of Medicine, zawa, 920. Japan.

July

19, 1984;

Nakajima, Kanazawa

revision

received

Department Oct.

M.D., Department of Nuclear University, Takara-machi 13-1,

of

9. 1984; .MediKana-

location of the ACP was limited.‘O Overlap of the blood pool was considered one of the major problems in detecting the site of the initial contraction. To solve this problem, we have attempted to apply Fourier analysis to gated emission computed tomography. l2 The purpose of this study was to evaluate the usefulness of tomographic phase analysis in detecting the site of the ACP in WPW syndrome. METHODS

Patients. Twenty patients with WPW syndrome (ages 16 to 54 years), who underwent surgical division of the ACP, were studied by gated blood pool scintigraphy and gated blood pool tomography. Routine 12-leadECGs were obtained before and after surgery in each patient. Two patients with intermittent type and four with concealed type showedno delta waves during radionuclide studies. When the patients were classifiedaccording to the surgically confirmed side of the ACP, eight patients were right cardiac type (including four right posterior septal types and one anterior septal type, patients Nos. 4 to 8 in Table I) and six were left cardiac type. There were no complications other than conduction anomalies in any of the patients. Epicardial mapping and surgery. Epicardial mapping wasperformed in all patients with WPW syndrome during 809

Nakajima et al.

810

American

1. Localizations of ACPs were confirmed by epicardial mapping and surgery. Numbers correspond to the patient numbers in Table I. MV = mitral valve; TV = tricuspid valve; Ao = aorta; PA = pulmonary artery; CS = coronary sinus. Fig.

I. Comparison of phase analysis and location of

Table

ACPs

Patient

Site of ACP

PlClfZW

cardiac type (including septal type) rP rPS rAL rL rP 4 rPS rPS rPS rP 5 6 rPS rPL 7 rPS rP rAS rA 8 Left cardiac type 9 1L 10 1L 1 (base) 11 1PL 12 1AL IL 13 1PL 1PL 14 IPL r(apex)/lP No delta waves during radionuclide study 15 1PL (intermittent) 16 rPS (intermittent) 17 rPS (concealed) rA/r(apex) 18 rP (concealed) rL/lA 19 1L (concealed) IA r(apex)/lPL 20 1L (concealed)

Tomographic

Right

1 2 3

rPS rAL rP/lL rP/lP rP rP rPS rAS 1L 1AL 1PL IAL 1PL IL r(apex)/l(apex) r(apex)/lL r(apex) rL/lA IA rA

ACE’ = accessory conduction pathway; r = right side; 1 = left side; identification possible; others abbreviations as in Fig. 2.

-

= no

surgery, as previously described.‘“,“’ By means of an electrode catheter, which had six pairs of bipolar electrodes with a 5 mm interelectrode distance, activation on the cardiac surface wasanalyzed by computerized recorder (SiemensElema Mingograph 82). The activation time

April, 1985 Heart Journal

was measuredfrom the beginning of the delta wave to the main deflection of the epicardial electrogram. The earliest epicardial activation was found along the atrioventricular groove, which indicates the ventricular insertion site of the ACP. Retrograde atria1 activation wasalso examined during reciprocating tachycardia. Thus, the location of the ACP was determined at the site of the earliest activation. The localizations of the ACPs are summarizedin Fig. 1. The direct incision was made at the site of the ACP along the anulusand wasextended to the adjacent area in both directions. After transection of the ACP, the ECG changedto a normal pattern with narrow a QRS complex and no delta waves, indicating the disappearance of ventricular preexcitation. Planar gated blood pool study. Equilibrium gated blood pool studies were performed with the useof 20 mCi (740 MBq) Tc-99m red blood cells (in vivo labeling by stannous pyrophosphate). A large field of view scintillation camera, equipped with a high-resolution, parallelhole collimator, waspositioned in the 30 to 40-degree(best septal) left anterior oblique projection with 10 to 20degreecaudaltilt and 35degree right anterior oblique and left lateral projections. The blood pool data were acquired in 24 frames/cardiac cycle in 64 X 64 matrices and stored in one of the two nuclear medicine computer systems (Ohionuclear VIP460 or Shimadzu Scintipac 2400s).Usually from 300,000to 500,000counts were collected in each frame. Gated emission computed tomography. The emission computed tomography system in this study consisted of dual heads of large field of view sintillation cameras, equipped with high-resolution, parallel-hole collimators, interfaced to a minicomputer system. Thirty-six projections, at every 10 degreesof rotation, were acquired in 64 X 64 matrices, with dual camerasrotating around the patients over 180degrees.The data acquisition time was2 minutes (later 1 minute) in each projection. Therefore, actual acquisition time was 2 X 18 = 36 minutes (later 18 minutes). Each cardiac cycle was divided into 12 frames. Full width at half maximum of the emission computed tomography system was 15 mm at a center of the reconstructed image,23 cm from the surfaceof eachcollimator. The transverse imageswere reconstructed by meansof a filtered-back projection algorithm, using Shepp and Logan’s filtering supplied by the manufacturer. The long axis of the heart was indicated manually and was rotated to obtain short-axis, vertical long-axis, and horizontal long-axis images.Two or three slicesof gated images,12 frames/cardiac cycle, were selected near the base of the heart in short-axis images.In long-axis images,the section about midportion of the ventricles was selected. No correction for attenuation wasperformed. Phase analysis. The phaseand amplitude of the fundamental wave of the discrete Fourier transform in each time-activity curve were calculated and mapped as functional images(Fig. 2). The phaseimage was displayed in an isocount display with a gray scalefrom white to black in x-ray film. Therefore, the timing of the contraction was

Volume

109

Number

4

Nuclear

SHORT

AXIS

tomographic

VERTICAL LONG AXIS

phase analysis

in WPW syndrome

811

HORIZONTAL LONG AXIS

,ant.basal

api&

AL

.at.

L

Fig. 2. Schematic representation of ventricular segments in each tomographic plane. In the short-axis image, the atrioventricular ring is divided into eight segments on each side. Note that short-axis images are displayed as observed from the apical direction.. RV = right ventricle; LV = left ventricle; A = anterior; AL = anterolateral; L = lateral; PL = posterolateral; P = posterior; PS = posterior septal or posterior paraseptal; S = septal; AS = anterior septal or anterior paraseptal.

PLANAR PHASE ANALYSIS RAO35O LA035O

TOMOGRAPHIC

SHORT AXIS

PHASE

L.LAT

160

ANALYSIS

VERTICAL LONG AXIS

HORIZONTAL LONG AXIS

80 Fig. 3. Phase images in a patient with a left anterolateral ACP (No. 12). The upper level of the scale is 160 degrees and the lower level is 80 degrees. Planar images in the 35-degree left anterior oblique and 35-degree right anterior oblique projections show the initial phase in the lateral region of the base of the left ventricle (arrows). In the left lateral view, the initial phase is observed also in the base of the heart. Tomographic phase images clearly represent t.he initial contraction in the anterolateral region in each section.

supposed to proceed from white to black regions, as shown in Figs. 3, 4, and 5. The range of the color scale was generally 10 degrees for each color in 8 or 16 steps. However, it was changed arbitrarily to clarify the ditference in timing of the contractions in some cases. The mask of the phase image was determined by thresholding of the amplitude image, at 50% of the maximum amplitude, to eliminate noise around the cardiac chambers. The initial site of the contraction, that is, the abnormal initial phase

in the phase image, was assessed by these phase patterns. The region of pulmonary truncus and outflow tract just before the pulmonic valve usually showed low amplitude. Therefore, we excluded these regions from the interpretation of phase images. When interpreting tomographic phase images, we used only the border of the phase images, that is, outside of the end-systolic perimeters, because time-activity curves interior to the end-systolic perimeters were considered to be meaningless. We

April,

812

Nukujima et al.

American

PLANAR PHASE ANALYSIS RAO35O LA035'

L.LAI

TOMOGRAPHIC PHASE ANALYSIS VERTICAL S:iijRT AXIS LONG AXIS

HORIZONTAL LONG AXIS

Heart

1985 Journal

Fig. 4. Phaseimagesin a patient with a right posterior septal ACP (No. 7). The initial phaseis observed in the right ventricle (arrows). The site of ACP is supposedto be in the right posterior segment;however, it is difficult to determine that the actual site of ACP is on the right posterior septal segment.On the other hand, tomographic phaseimagesrepresent the location of the initial contraction in the posterior septal segment.

assumedthat the initial phase was visualized in the periphery of the cardiac chambers. Fig. 2 showsthe segmentsof tomographic images.The atrioventricular rings in the tomographic short-axis imageswere divided into eight segmentson eachside. The PS segmentincluded the posterior septum and posterior paraseptalsegments,becauseit wasdifficult to differentiate these two regionson blood pool images.Actually, the right posterior paraseptalbypasstract wasnot included in our study group. The site of the ACP determined electrophysiologically and the abnormal initial phasewere compared. If the initial phase corresponded to the actual segment in which the ACP was present, the correct diagnosiswas given. However, if the region of the initial phase was adjacent to the confirmed site of the ACP, it was also consideredto be correct, becauseit was not easy to localize an initial region in one of the eight segmentsin the short-axial section. RESULTS

Table I shows the comparison of the surgically confirmed ACP and the phase analysis, by means of planar gated blood pool study in multiple projections and gated emission computed tomography. Right type. A correct. diagnosis was made in six of eight patients with right cardiac types (including

right posterior septal and anterior septal types) by both planar and tomographic phase analyses. In a patient with a right. posterior ACP (No. 3) no identification of an abnormal initial phase was possible in planar phase images, whereas tomographic phase images showed two regions of initial phase in the right posterior (vertical long-axial section) and left lateral (short-axial section) regions, and the earliest phase was not determined. Left type. Two of six ACP sites in patients with left cardiac types were detected by planar phase analysis, while all ACPs (six of six) were identified correctly by tomographic phase analysis. Planar phase images in a patient with a left lateral ACP (No. 10) showed the earliest phase in the base of the left ventricle; however, more precise localization of

the segment was difficult. In this patient, tomographic phase analysis was useful for localizing the initial segment,. In two patients with left lateral and left posterolateral ACPs (Nos. 9 and ll), there were no significant differences in timing of contractions between both ventricles. However, in these patients, tomographic phase analysis represented the initial phase in the segments of actual location of the ACPs.

Volume

109

Number

4

Nuclear

SPECT $3

tomographic

phase analysis

in WPW syndrome

813

30 35

F

IM:f3l459GElll.IM

KClNFIZflWFI

UNIV.

HOSP.

Fig. 5. Tomographic short-axis phaseimagesin a patient with an anterior septal ACP (No. 8). The left upper imageis near the baseof the ventricle, and the right lower imageis near the apex. Slice thicknessis 6 mm. The earliest phase in the peripheral region is observed in the right anterior septal segment(arrow). However, an earlier phase(arrowhead) is indicated inside of the phaseimage,implying the location of the anterior septal bypasstract.

Absent delta waves. Six patients without delta waves had contraction patterns, which began near the apex or anterior wall and proceeded to the base of the heart in vertical long-axis images. However, in the short-axis images, the contraction initiated in the anterior, lateral, or posterior segments, and there was no specific segment of initial contraction. On the other hand, 11 of 14 patients with WPW syndrome had the initial phase at the base of the heart and the contraction propagated to the apical direction. Overall accuracy. Overall diagnostic accuracy in detecting the site of the ACP was 57% (8 of 14) for planar phase imaging and 86% (12 of 14) for tomographic phase imaging. Figs. 3, 4, and 5 show examples of planar and tomographic phase analysis with left anterolateral, right posterior septal, and anterior septal ACPs.

Application of phase analysis to conduction anomalies has been attempted, and there has been good agreement between the propagation of excitation and abnormalities in phase images3-” These

excitation has close correlation with mechanical movement of the chambers. We have performed planar phase analysis in 50 patients with WPW syndrome, who underwent surgical division of the ACP. However, we suppose that planar phase analysis is unsatisfactory for detection of the exact site of the ACP as a method of preoperative study. Our previous study of 21 patients with WPW syndrome showed that the side of preexcitation was correctly identified in 80% of cases, 100% for right cardiac type, and 60% for left cardiac type; however, regarding the precise localization of the ACP confirmed by epicardial mapping and surgery, only 50% were identified correctly in the left anterior oblique view.‘O In the planar phase images with multiple projections, as shown in this study, the diagnostic accuracy was slightly better (57%). Even if we use multiple projections, overlap of the blood pool cannot be avoided and an abnormal phase may be calculated. To overcome this limitation, we speculated that tomographic images of the gated blood pool might have the advantage of avoiding superimposition of the chambers. We have tried to evaluate the usefulness of gated blood pool tomography in

observations have indicated that electrophysiologic

detecting the site of the ACP more precisely.‘2

DISCUSSION

April.

814

Nakajima

Comparison analyses. The

et al.

of

tomographic

American

and

planar

phase

results of this study indicate that tomographic imaging has a higher sensitivity for detecting the site of the ACP than does planar phase image analysis. Particularly, the detectability of an abnormal initial phase was improved in the left cardiac types. The phase pattern and site of the initial phase were more easily recognized in tomography. When the initial phase adjacent to the actual segment of the ACP was also judged correctly, 12 of 14 patients were diagnosed correctly. We reasoned that the lack of firm agreement in this study was due to the following factors. (1) It was sometimes difficult to pinpoint the initial phase on the phase image. For example, in Fig. 3, the initial phase is observed from anterior to lateral segments. We judged that the initial phase was in the anterolateral segment from contraction patterns; however, it is difficult to localize the segment in one of the segments. (2) Correspondence between the cardiac surface during epicardial mapping and blood pool images was not strict. (3) The incision across the ACP was usually more than one fourth of the atrioventricular ring. It is particularly noteworthy that ACPs not detected in routine phase analysis could be identified in tomographic phase analysis. Limitations of tomographic phase analysis. The phase analysis based on gated emission computed tomography also has limitations. The spatial resolution, 15 mm in full width at half maximum, was apparently poor compared to the actual movement of the ventricular wall induced by preexcitation. The edges of the blood pool in tomography are not sharp and may be distorted because of the processing of filtering and reconstruction of the tomographic plane. Theoretically, in tomographic blood pool images, time-activity curves interior to the endsystolic perimeter must be flat and the phase in the region is supposed to be meaningless. Although the time-activity curves inside of the end-systolic perimeter are not flat, because of a blurring effect during processing, they are considered to be unreliable. Therefore, we used only the border of the phase images, where initial inward movements are represented. The frame rate per cardiac cycle was also restricted in our computer. Data sampling of 12 frames/cardiac cycle may not be sufficient; however, it will be permissible to characterize the shape of the volume curve in each pixel. Our results in this study showed that phase analysis was valuable for detecting the site of ACPs. Although the edge of the ventricle was blurred in blood pool images, the blurred region was also moving and an almost correct phase might have been calculated.

Heart

1995 Journal

The accuracy of phase analysis in detecting preexcitation is also affected by the degree of preexcitation. Swiryn” demonstrated that comparison of the relative ipsilateral vs contralateral mean and early phase correlated with QRS duration. Therefore, some interventional studies, which use atria1 or esophageal pacing7,y.11 or drug administration8 can be effective in patients with a narrow QRS complex, although we did not perform such studies. Detection of septal and paraseptal ACPs. Differentiation of septal and paraseptal bypass tracts is expected in the tomographic phase images. However, as described in the Methods section, we used only the peripheral regions of the phase images, since the region inside of the end-systolic blood pool was considered to be meaningless and confusing. Additionally, the septum is adjacent to the low-amplitude region, and septal and paraseptal segments are not easily differentiated in some patients. Nevertheless, it is interesting to note that Fig. 5 indicates the earliest phase in the anterior septal region not in the peripheral region of the phase images. We have also experienced the earliest phase on the septal region, which implies normal atrioventricular conduction. But this finding is not always observed. To confirm the significance of these findings, further studies are required comparing phase images in patients with septal and paraseptal ACPs as well as normal atrioventricular conduction. Phase analysis in concealed-type WPW syndrome and normal subjects. Among patients without delta

waves, two were intermittent type and four were concealed type with only retrograde conduction. Because there was no evidence of antegrade conduction of the ACP during electrophysiologic mapping. the conduction patterns in these concealed-type patients were considered to be normal atrioventricular conductions. In planar phase images, multiple foci of the earliest contraction were demonstrated other than in the septal region, including inferoapical and basal regions in the right ventricle and multiple regions in the left ventricle.‘,” In tomographic phase analysis of the short-axial section, no specific segments of the initial contraction were noted in the left ventricle, whereas in the right ventricle, the earliest phase was generally observed in the apical or anterior wall. In normal contractions, through the atrioventricular node, the difference in the timing of the contraction is rather small and the phase is also influenced by the direction and magnitude of the contraction. In our opinion, the purpose of the phase analysis in WPW syndrome is not the diagnosis of the presence of WPW syndrome. The correlation of electrical excitation and

Volume

109

Number

4

Nuclear

the sequence of wall motion is an interesting subject, and if the correlation is good, phase analysis can be applied to analyze the contraction patterns in WPW syndrome. Role of phase analysis.Preoperative estimation of the ACP has been performed by ECG, vectorcardiography, echocardiography, body surface mapping, and intracavitary recording, with the use of multiple electrode catheters. Although, at present, epicardial mapping during surgery is the most reliable and indispensable method, l3 it is useful if noninvasive and accurate methods are available before surgery. Phase analysis is not enough for the localization of the ACP; however, we have demonstrated that it can be a good adjunctive method as a preoperative method. Phase analysis will also be useful when a patient with suspected WPW syndrome has indeterminate ECG patterns, because the side of preexcitation and the location of the ACP are predicted easily and noninvasively. Conclusions. Tomographic phase analysis can correctly identify the site of the initial contraction, that is, the site of the ACP, in 86% (12 of 14) of patients with WPW syndrome. Good correlation was observed between the site of the ACP and the abnormal initial phase in tomographic phase images. The detectability of the ACP was higher than that in conventional planar phase images. The phase analysis can be a good adjunctive method for the evaluation of patients before surgical division of the ACP. REFERENCES

1. Adam WE, Equilibrium vast Radio1

Tarcowska A, Bitter (gated) radionuclide 2:161, 1979.

F, Strauch M, ventriculography.

Geffer H: Cardio-

tomographic

phase analysis

in WPW syndrome

815

2. Links JM, Douglass KH, Wagner HN Jr: Patterns of ventricular emptying by Fourier analysis of gated blood pool studies. J Nucl Med 21:978, 1980. 3. Swiryn S, Pave1 DG, Byrom E, Witham D, Meyer-Pave1 C, Wyndham CRC, Handler B, Rosen KM: Sequential regional phase mapping of radionuclide gated biventriculogram in natients with left bundle branch block. AM HEART d iO2:1000, 1981. 4. Turner DA, von Behren PL, Ruggie NT, Hauser RG, Dennes P., Ali A. -, Messer JV. Fordham EW. Groch MW: Non-invasive identification of initial site of abnormal ventricular activation by least square phase analysis of radionuclide cineangiogram. Circulation 65:1511, 1982. 5. Botvinick EH, Frais MA, Shosa DW, O’Connel JW, PachecoAlvarez JA. Scheinman M. Hattner RS. Moradv F. Faulkner DW: Accurate means of dktecting and charact&zing abnormal patterns of ventricular activation by phase image analysis. Am J Cardiol 50:289, 1982. 6. Mena I, French M, Labs M: Diagnosis and pictorial presentation of Wolff-Parkinson-white syndrome by phase image analysis. J Nucl Med 23:~ 55, 1982. 7. Chan WWC, Kalff V, Dick M II, Rabinovitch MA, Jenkins J, Thrall JH, Pitt B: Topography of pre-emptying ventricular segments in patients with Wolff-Parkinson-White syndrome using scintigraphic phase mapping and esophageal pacing. Circulation 67:1139, 1983. 8. Racovec P, Kranjec I, Fettich J, Fidler V, Pungercar D, Janezic A, Porenta M, Varl B: Localization of accessory conduction pathways in Wolff-Parkinson-White syndrome by phase imaging. Cardiology 70:138, 1983. 9. Swiryn S: Nuclear electrophysiology. PACE 6:1171, 1983. K, Bunko H, Tada A, Taki J, Tonami N, Hisada K, 10. Nakajima Misaki T, Iwa T: Phase analysis in the Wolff-ParkinsonWhite syndrome with surgically proven accessory conduction pathways. J Nucl Med 25:7, 1984. E, Frais M, O’Connell W, Faulkner D, Scheinman 11. Botvinick M, Morady F, Sung R, Shosa D, Dae M: Phase image evaluation of patients with ventricular pre-excitation syndromes. J Am Co11 Cardiol 3:799, 1984. 12. Nakajima K, Bunko H, Tada A, Tonami N, Taki J, Nanbu I, Hisada K, Misaki T, Iwa T: Tomographic phase analysis to detect the site of accessory conduction pathway in WolffParkinson-White syndrome. J Nucl Med 25:~ 87, 1984. M, Misaki T, Iwase T, Magara T: Localiza13. Iwa T, Kawasuji tion and interruption of accessory conduction pathway in the Wolff-Parkinson-White syndrome. J Thorac Surg 80:271, 1980.