In vivo assessment of left ventricular wall and chamber dynamics during transient myocardial ischemia using cine computed tomography

In Vivo Assessment of Left Ventricular Wall and Chamber Dynamics During Transient Myocardial lschemia Using Cine Computed Tomography DONALD FARMER, MD, MARTIN J. LIPTON, MD, CHARLES B. HIGGINS, MD, HANS RINGERTZ, MD, PhD, PETER B. DEAN, MD, RICHARD SIEVERS, BS, and DOUGLAS P. BOYD, PhD

Using a new computed tomographic (CT) scanner design that uses a rapidly moving focused electron beam, 50-ms CT scans were obtained at 2 axial levels simultaneously through the hearts of 6 dogs in order to analyze left ventricular (LV) wall thickness and cross-sectional chamber area after acute occlusion of the left anterior descending coronary artery (LAD). Ten or fifteen 50-ms CT scans (rate of 17 scans/s) through the middle of the left ventricle were performed in 1 second (tine acquisition) during intravenous administration of contrast medium at rest, 60 seconds afler acute occlusion of the LAD, and 60 seconds after release of the occlusion. The percent extent of systolic wall thickening of the potentially ischemic anterior segment was 37 f 15 % (f standard deviation) in the control state and

-5 f 6.5 % during LAD occlusion (p
Effective cardiac diagnosis is dependent on the early detection of myocardial ischemia and subsequent quantitation of the extent of myocardial ischemic damage. Current, noninvasive functional imaging methods require the demonstration of regional contraction abnormalities to detect myocardial ischemia. Prospective13 and retrospective4y5electrocardiographically gated computerized transmission tomography

(CT) not only provides good discrimination of the ischemic myocardium, but also assessesleft ventricular (LV) dimensions, monitors LV wall thickness throughout the cardiac cycle and detects changesin wall thickening dynamics. However, the clinical application of both electrocardiographic (ECG) gating techniques to the assessment of acute myocardial infarction is limited by the ability of patients to lie flat and suspend their respiration for the 45 to 60 seconds necessary for data acquisition using ECG gating. Becauseof the rapid movement of cardiac structures, the only alternative to gated CT imaging of the heart is to use millisecond scan times. These millisecond CT scans have recently been provided by a new CT scanner design that usesa rapidly moving focused electron beam.6T7This study was designed to document the ability of the high-speed tine CT scanner to detect changes in LV wall thickening dynamics throughout the cardiac cycle of the in situ heart by analyzing LV wall thickness and LV cross-

(Am J Cardiol 1965;55:560-565)

From the Department of Radiology, University of California, San Francisco School of Medicine, San Francisco, California. This study was supported in part by Grant #HL-23430 from the National Institute of Health, Imatron, Inc. and by the Radiology Research and Education Foundation. Dr. Peter 6. Dean was supported by the Union Against Cancer; American Cancer Society-Eleanor Roosevelt-International Cancer Fellowship. Manuscript received July 20, 1984; revised manuscript received October 30, 1984, accepted October 31, 1984. Address for reprints: Charles B. Higgins, MD, Professor of Radiology, UCSF School of Medicine, San Francisco, California 94143. 560

February 15,1985

sectional area during 1 complete cardiac cycle before, during and after release of the acute coronary arterial occlusion,

Methods Experimental model: Eight conditioned mongrel dogsthat weighed 26 to 55 kg were prepared. Under general anesthesia with pentobarbital sodium (25 mg/kg), a left thoracotomy was performed and a hydraulic occluder was positioned around the proximal left anterior descending coronary artery (LAD) proximal to the origin of the first diagonal branch and distal to the first septal branch. After complete surgical recovery at 7 days, the experiment was conducted. Before scanning, the dogs were premeditated with morphine sulfate (2 mg/kg intramuscularly), anesthetized with phentobarbital(20 mg/kg intravenously) and ventilated with a Harvard respirator (15 breaths/min, tidal volume 12 to 15 ml/kg). The heart rate was usually 100 to 150 beats/min. Intravenous contrast material (Conray 400) was administered in a bolus fashion, through a femoral venous catheter placed at the inferior vena caval-right atria1 junction. The contrast medium was injected as a bolus at a dose of 0.5 ml/kg of body weight. Ten to 15 axial CT scans were obtained at the midventricular level using the tine acquisition mode of the scanner at a rate of 17 scans/s. This scanning sequencewas performed during control, at the conclusion of a 60-second period of occlusion of the LAD, and 60 seconds after release of the 60second LAD occlusion. Before the tine acquisition, the flow (triggered) mode of operation (scanner described later) was used to determine the exact time for the bolus administration during each tine sequence so that optimal opacification was present in both the left and right ventricles. During the flow acquisition, the bolus of contrast medium could be sequentially imaged by each of 20 scans obtained at end-diastole of every other heartbeat as it progressed from the right to the left-sided cardiac chambers. During the entire experiment, 40 to 50 ml of contrast medium was administered. Scanning was performed at the mid-LV level during suspension of respiration at full inspiration. The scanning x-ray factors were: 125 kVp, 50-ms scan time, 650 mA, with an 8-mm slice thickness. Two 8-mm slices were

THE AMERICAN JOURNAL OF CARDIOLOGY Volume 55

obtained simultaneously. All images were reconstructed using the same algorithms with a 256 X 256 matrix. The approximate radiation dose for each tine acquisition was 3.7 or 5.5 rads over a 2-cm thickness through the midthorax, depending on the number of scans in the tine sequence (10 or 15, respectively). Current spatial resolution is 1.5 mm. Six of the 8 dogs evaluated provided satisfactory imagesand were analyzed in the study. The 2 not used were rejected either becauseof intermittent detector failures or suboptimal timing on the tine acquisition, leading to poor LV contrast levels. Cine computed tomographic scanner: The prototype Imatron C-100 scanner used for this experiment is a unique type of x-ray CT scanner that uses an electron beam that is magnetically focused and deflected to replace the mechanical motion used in standard CT scanners. The C-100 scanner has been described in detail.s-s Briefly, the electron beam scan tube comprises an electron gun and accelerator which produces a 650-mA beam at 125 kV. This beam is focused and deflected through an angle of 33 to 37°C and swept along 1 of 4 fixed tungsten target rings through an angle rotation of 210°. A simultaneous pair of dual scans is thereby produced from a fan beam of radiation, which is emitted from each tungsten target ring and is converted by a collimator into 2 side-by-side 8.0-mm-thick slices. Two rings of stationary detectors are placed above the subject and receive the transmitted fans of radiation (Fig. 1). Although the 4 target rings can be sequentially scanned in 240 ms, producing 8 adjacent l-cm-thick levels, only 1 of the target rings was used during the current study. The C-100 tine system affords 3 modes of operation, of which the tine and flow (triggered) acquisition modes were used on the current study. The volume mode, which uses all 4 target rings, was not used. During the tine acquisition sequence, up to fifteen 50-ms scans can be obtained during a single cardiac cycle at the rate of 17 scans/s. The initial image of the sequence is triggered in relation to the R wave of the electrocardiogram so as to occur at end-diastole of the imaged heartbeat. The flow (triggered) sequence is used for perfusion or flow analysis. In this mode of operation, each of twenty 50-ms exposures are triggered by the electrocardiogram at the same

DA5 /

FIGURE 1. Diagram of the tine computed tomographic scanner showing the course of the electron beam from which a fan-shaped beam of radiation is produced. The beam passes through the subject and strikes 2 fixed detector rings located above the couch.

561

562

CINE COMPUTED

TOMOGRAPHY

OF LEFT VENTRICLE

FIGURE 2. Nine sequential, 50-ms images triggered to begin at enddiastole of a single cardiac cycle in the control state (top), at the conclusion of a go-second occlusion of the left anterior descending coronary artery, (middle) and at 60 seconds after the release of the occlusion (bottom). The images near the end of each sequence are in diastole of the next cardiac cycle.

phase of the cardiac cycle, usually at end-diastole for every heartbeat or a multiple of every heartbeat. The time course of the contrast bolus can thus be assessedfrom its appearance in the right side of the heart to peak opacification and washout of the left ventricle. Wall thickness measurement: The 15 frames of the movie sequence obtained for each level during the tine acquisition were sequentially displayed for the control, occlusion and postrelease states (Fig. 2). End-diastolic and end-systolic images for each cardiac cycle were defined by the largest and the smallest measured LV luminal area, respectively. These were then displayed simultaneously, magnified 2 times and wall measurements were taken (Fig. 3). LV wall boundaries were objectively defined by a computer program that displays a half contour to provide an objective criteria for placement of the measuring cursor. This objective analysis method was selected because of considerable interobserver variability using subjective analysis. For this technique, regions of interest were selectively placed in the mid-LV cavity, over the LV myocardium (M) (usually the lateral wall) and in the lung (L) adjacent to the left ventricle away from any major pulmonary vascularity. The mean CT number for each region of interest is displayed. The mean values of the chamber and lung are each subtracted from that obtained in the myocardium, with the respective differences then being halved. The “half-density contour” levels (L-M)/2 and (M-LV/2), thus defined the epicardial and endocardial interfaces respectively. This same technique was used to define the boundaries of the LV cavity for measurement of LV luminal area. Left ventricular measurementi The LV luminal area was calculated using the same magnified images as those for the wall thickness measurement. Using the half-density contour level as a guide, an irregular region of interest was traced along the endocardial margin of the left ventricle and mitral valve plane. Subsequent closure of the region and placement of the cursor within the region of interest, provided the value for luminal area, mean CT number of the region as well as the standard deviation of this mean value (Fig. 4). Data analysis: As described above, the end-diastolic and end-systolic images were chosen as the largest and smallest luminal areas. It is unlikely, however, that the end-diastolic and end-systolic images are at exactly the same sagittal level of the left ventricle, because of shortening of the major axis during systole. The extent of wall thickening at control, at 60 seconds after LAD occlusion, and 60 seconds after release of the LAD occlusion was calculated for the septal, anterior and lateral walls. The percent extent of wall thickening was determined by subtracting the end-diastolic wall thickness from the end-systolic wall thickness and dividing the difference by the end-diastolic thickness X 100. Similarly, subtracting the end-systolic luminal area from the end-diastolic luminal area and dividing by end-systolic area X 100 provided the percent change in LV luminal area. Statistics: Wall thickness measurements for the septal, anterior and lateral walls were compared for each end-systolic and end-diastolic frame obtained at the control state, during occlusion, and after release of occlusion (Fig. 4, Table I). The calculated percent wall thickening and percent change in LV luminal area were also compared at rest, during occlusion and after release. All data are presented as the group mean f standard deviation. Diastolic and systolic values were analyzed for overall significance using the paired t test. Other comparisons were made using the standard t test.

Results Effect of LAD occlusion on extent of wall thickening: The group data for percent extent of wall

February 15, 1985

FIGURE 3. End-diastolic (left) end-systolic (right) images from acquisition in the control state. sites at which the septal, anterior lateral walls are measured shown.

THE AMERICAN

JOURNAL

OF CARDIOLOGY

Volume 55

563

and tine The and are

thickening in the control state, during LAD occlusion, and after release of LAD occlusion are presented in Figure 5. In the control state the extent of wall thickening was 53 f 26.7%, 37.2 f 15%, and 48.1 f 8.6% for the septal, anterior and lateral walls, respectively. During occlusion of the LAD, the ischemic anterior wall showed paradoxical thinning during systole, reflected by the percent extent of wall thickening of -5 f 6.5%, (p
decreasedby 50 f 17%from end-diastole to end-systole. During occlusion of the LAD, the change in luminal area was not significantly different from that during control, 47 f 21%vs 50 f 17%.At 60 seconds after the release of LAD occlusion, the change in LV luminal area was also not significantly different from that in the control state, 44 f 23%. Discussion Wall thickness and extent of wall thickening during the cardiac cycle are sensitive and reliable measures of segmental myocardial function in regional ischemia.e-19 Indeed, wall thickening measurements may be the most sensitive and proximate measure of regional myocardial function.il Ischemia-induced alterations in segmental wall thickening dynamics have been demonstrated using cross-sectional tomographic imaging techniques such as ECG-gated CT.1-5J3-16and sector scan echocardiography. g- I2917Prospectively ECG-gated CT has been used to monitor regional wall thickness throughout the cardiac cycle and has documented either complete

0 Contra/

q

Occlusion @ Release

T

1Ti TT

TT

Diastote

10

(mm1 8t

lyJh...... .--------*-------+

,

Conirol

Anterior

Loterol

FIGURE 4. End-diastolic and end-systolic wall measurements for the control occlusion and release states in the septal, anterior and lateral walls. There is a marked decrease in systolic thickness of the anterior wall during occlusion.

FIGURE 5. Extent of wall thickening in the septal, anterior and lateral walls during the control state, left anterior descending coronary artery occlusion and after release of occlusion. There is significant thinning of the anterior wall during occlusion.

564

CINE COMPUTED

TABLE I

TOMOGRAPHY

OF LEFT VENTRICLE

Wall Thickness Measurements for the Septal, Anterior and Lateral Walls in the Control State, During Occlusion and After Release of Occlusion in 6 Dogs

C

Lateral

Anterior

Septum 0

R

C

R

0

C

R

0

Dog 1 13.99

D S

8.56

13.5 10.8

149::

5.76

15.6 12.1

10.00 15.3

13.4 12.1

13-i:

12 12

9.4 13.9

12.2 16.8

10 16.1

10.8 14.4

7.16

128:;

14.8 10.6

1:::

12.2 10

10.4 10.6

15.5 11

1:::

t:;

10.1 8.6

2:

1:::

11 16.5

14?

11.6 14.4

8::

8.7 11.9

1;:

11 15.2

11.5 15.2

Dog 2 1;::

D s

10 14.9

7.8 13.5

11.7 14 Dog 3

:

12.9 9.1

9.6 9.1

11.49

284

:

10.1 13

11.5 12.2

13.3 11

15!:

D S

10.1 13.8

1::::

9.7 11.6

1:::

Dog 4

Dog5

Dog 6 6.5 9.4

D s Mean x Diastole Systole

8k 1” 13.1 f 2

1::; 9.2 f 2 12f2

* p = 0.01; t p = 0.05. C = control state; 0 * occlusion;

8.6 12.2

192:;

9.3 f 1” 12.6 f 1

10.1 f it 13.7 f 1

9.6 f 2’ 13.5 f 2

10.4 f 2” 15.9 f 2

llfl 13f2

R = release.

loss of wall thickening or actual wall thinning of the jeopardized region after acute coronary occlusion.3 The results of the present study are similar to those obtained using ECG-gated CT,3 with paradoxical thinning in the ischemic segment after acute coronary occlusion.3J*Je As with all previous angiographic digital fluoroscopic and CT studies devoted to the evaluation of LV function, contrast medium has an influence on cardiovascular

80 r

Control

8.6 f 2 8.5 f 2

Occlusion

hemodynamics. Consequently, like many previous studies published over the past 2 decades, the current results must be interpreted in this light. Although ECG-gated CT can accurately assess myocardial wall dynamics during the cardiac cycle, such studies Deereachieved under carefully controlled experimental conditions that are not directly applicable to the clinical situation. The handicaps of gating have been overcome by the development of a CT scanner with exposure times of 50 ms or less.*p20 The C-100 tine CT scanner used in the current study requires only 50 ms per image, enabling multiple, complete images to be obtained during a single cardiac cycle. The potential to image acute myocardial events now seems realistic because this can be accomplished with minimal amounts of intravenous contrast and a minimal period of breath holding. The present study indicates that tine CT can be used for detecting and quantitating acute myocardial ischemia. However, the ability of tine CT to detect wall thickening abnormalities that occur with nonocclusive stenosis that may result in only mild reduction in coronary blood flow is uncertain. Nonetheless, the study of LV contraction abnormalities in patients is a current ongoing project, the results of which are the subject of another report.

Release

FIGURE 6. Changes in segmental left ventricular luminal area during control, occlusion and after release of occlusion. There is no significant difference in the percent change in luminal area among the 3 states.

Acknowledgment: We thank Lauranne Cox, RT, Peter Martin PhD, and Reggie Pike for technical expertise and dedication in operating the Cine-CT prototype scanner used

February 151985

in this study. We also thank Pedi Alcala for editing and typing this manuscript.

References 1. Doherty PW, Lipton MJ, Berninger WH, Skioldebrand CG, Carlsson E, Redington RW. Detection and quantitation of myocardial ischemia in vivo ;;i;g transmission computed tomography. Circulation 1981;63:5972. Slutsky RA, Mattrey RF, Long SA, Higgins CB. In vivo estimation of myocardial infarction size and left ventricular function by prospectively ECG Circulation transmission tomography. computerized gated 1983;67:759-765. 3. Mattrey RF, Slutsky RA, Long SA, Higgins CB. In vivo assessment of left ventricular wall and chamber dynamics during transient myocardial ischemia using prospectively ECG-gated computerized transmission tomography. Circulation 1983;67:1245-1257. 4. Berninger W, Redington RW, Leue W, Axe1 L, Normand D, Brundage 6, Car&on E, Herfkens RJ, Lipton MJ. Technical aspects and clinical apolications of CT/X, a dvnamic CT scanner. J Comput Tomogr 1981;5: 206-215. 5. Berninger W, Redington RW, Doherty P, Lipton MJ, Carlsson E. Gated cardiac scanning: canine studies. J Comput Tomogr 1979;3:155-163. 6. Boyd DP, Lipton MJ. Cardiac computed tomography. Proc IEEE 1983;71: 29&3of. 7. Lipton MJ, Higgins CB, Farmer D, Boyd DP. Cardiac imaging with a high;$;ed tine-CT scanner: preliminary results. Radiology 1984;152:579-

---.

8. Boyd DP, Gould RG, Ouinn JR, Stanley J, Hermansfeldt W. A proposed dynamic cardiac 3-D densitometer for early detection and evaluation of heart disease. IEEE Trans Sci 1979:26:2724-2727. 9. Kerber RE, Marcus ML, Ehrhardt J, Wilson R, Abboud FM. Correlation between echocardiographically demonstrated segmental dyskinesis and

THE AMERICAN

JOURNAL

OF CARDIOLOGY

Volume 55

565

regional myocardial perfusion. Circulation 1975;52:1097-1104. 10. Gassch WH, Bernhardl SA. The effect of acute changes in coronary blood flow on left ventricular end-diastolic wall thickness. An echocardiographic study. Circulation 1977;56:593-598, 11. Sasayama S, Franklin D, Ross J, Kemkper WS, McKown D. Dynamic changes in left ventricular wall thickness and their use in analyzing cardiac function in conscious dogs. Am J Cardiol 1976;38:870-879. 12. Kerber RE, Abboud FM. Echocardiographic detection of regional myocardial infarction: an experimental study. Circulation 1973;47:997-1005. 13. Mattrey RF, Higgins CB. Detection of regional myocardial dysfunction during ischemia with computerized tomography: documentation and physiological basis. Invest Radio1 1982;17:329-335. 14. Lipton MJ, Higgins CB. Evaluation of ischemic heart disease by computerized transmission tomography. Radio1 Clin North Am 1980;18:557576. 15. Skioldebrand CG, Overfors CO, Mavroudls C, Lipton MJ. Assessment of ventricular wall thickness in vivo by computed transmission tomography. Circulation 1980;61:960-965. 18. Lackner K, Thurn P. Computed tomography of the heart: ECG-gated and continuous scans. Radiology 1981;140:413-416. 17. Lieberman AN, Weiss JL, Judgutt BI, Becker LC, Bulkley BH, Garrison JG, Hutchins GM, Dallman CA, Weisfeldt ML. Two-dimensional echocardiography and infarct size: relationship of regional wall motion and thickening to the extent of myocardial infarction in the dog. Circulation 1981;63: 739-746. 16. Slutsky RA, Curtis G, Battler A, Froelicher V, Ross J Jr, Gordon D, Ashburn W, Karliner J. Effect of sublinaual nitroalvcerin on left ventricular function at rest and during spontaneous angina pectoris: assessment with radionuelide aooroach. Am J Cardiol 1979:44:1365-1370. 19. Sharm‘a’ B, Hodges M, Asinger RW, Goodwin JF, Francis GS. Left ventricular function during spontaneous angina1 pectoris: effect of sublingual nitroglycerin Am J Cardiol 1980;46:34-41. 20. Ritman EL, Robb RA, Johnson SA, Chevalier PA, Gilbert BK, Greenleaf JF, Sturn RE, Wood EH. Quantitative imaging of the structure and function of the heart, lungs and circulation, Mayo Clinic Proc 1978:53:3-l 1.