Use of computerized tomography to assess myocardial infarct size and ventricular function in dogs during acute coronary occlusion and reperfusion

Use of computerized tomography to assess myocardial infarct size and ventricular function in dogs during acute coronary occlusion and reperfusion

Use of Computerized Tomography to Assess Myocardial Infarct Size and Ventricular Function in Dogs During Acute Coronary Occlusion and Reperfusion G. B...

1MB Sizes 0 Downloads 74 Views

Use of Computerized Tomography to Assess Myocardial Infarct Size and Ventricular Function in Dogs During Acute Coronary Occlusion and Reperfusion G. B. JOHN MANCINI, MD, WALLACE

W. PECK, MD, ROBERT A. SLUTKSY,

MD,

JOHN ROSS, Jr., MD, and CHARLES B. HIGGINS, MD

Prospectively ECG-gated and nongated computed tomography (CT) can be used to assess global and regional left ventricular (LV) function and to measure myocardial infarct (Ml) size. In the current study, CT was used to assess the effects of coronary occlusion and reperfusion in 16 dogs. Ten dogs were subjected to permanent occlusion of the proximal left anterior descending coronary artery and 6 dogs were reperfused after a 2-hour period of total coronary occlusion. Gated scans were used to quantitate the extent of wall thickening in the ischemic zone and to assess changes in mid-LV cross-sectional chamber area at end-diastole and end-systole. Nongated scans were used to estimate the size of the initial perfusion defect during contrast injection shortly after coronary occlusion and the size of the Ml as indicated by delayed enhancement of the infarct 10 to 30 minutes after cessation of contrast administration. Neither group showed significant changes in end-diastolic chamber area during acute occlusion or 3 days later. Both groups showed a significant deterioration in percent change in chamber area both early after coronary occlusion and 3 days later; however, in the permanent occlusion group, percent wall thickening in the ischemic zone decreased from 46.2 f 16.5% (mean f standard deviation) to 1.6 f 9.0% during acute occlusion (p
lO.l%, p
Recent implementation of gated computerized tomography (CT) has allowed assessment of global and regional left ventriclar (LV) function in response to

ischemia1p2 and pharmacologic agents3 The ability of CT to detect myocardial infarction (MI) has been evident since the reports of Adams et a1.4Since then, several studies have examined the ability of CT to detect and determine the size of MI in ex vivo and in vivo experimental models, and although specific protocols have differed, findings indicate that CT is a relatively accurate method for measuring MI size.sm1sBecause CT can provide information about both the functional and the morphologic effects of ischemia, it offers the possibility of assessing several effects of interventions designed to reduce ischemic damage.

(Am J Cardiol

From the Departments of Medicine and Radiology, University Hospital, University of California, San Diego, California. This study was supported by SCOR Grant 3 P50HL17582-0981 awarded by the National Heart, Lung, and Blood Institute, Bethesda, Maryland. Dr. Mancini is supported by the Canadian Heart Foundation, Ottawa, Canada. Manuscript received June 16, 1983; revised manuscript received September 9, 1983, accepted September 22, 1983. Address for reprints: G. B. John Mancini, MD, Division of Cardiology (111 A), Veterans Administration Medical Center, 2215 Fuller Road, Ann Arbor, Michigan 48105. 282

1984;53:282-289)

January 15, 1984

Cross-sectional echocardiography has been used to study the functional derangements caused by coronary occlusion and reperfusion, 14-16 but areas at risk must be defined by in vivo Monastral blue injections or microsphere injections. An immediate potential advantage of CT over echocardiography in assessing effects of interventions lies in the 2 distinct phases of ischemial infarct imaging provided by CT. Several studies confirm that imaging during constrast infusion generally produces an area of decreased attenuation distinct from normal myocardium.5,7-10J2J3 This “perfusion defect” has been detected early after coronary occlusion17 and likely represents an area of ischemic jeopardy. At a later time after the administration of contrast media, the infarcted myocardium demonstrates preferential enhancement compared with normal myocardium.5,7-13 This delayed enhancement has been correlated with the accumulation of iodine in cells that have had ischemic damage and disruption of normal membrane functions.4Js Such changes are known to occur early after ischemic injurylgp21 and to correlate with subsequent necrosis.“2

FIGURE 1. Ungated tomographic slices from the base to the apex of a single dog. The slices are oriented with the basal slice at the top, progressing to the apical slice at the bottom. Left, the appearance during contrast infusion during the acute occlusion stage. Large areas of decreased attenuation (initial perfusion defect) are readily seen in each tomographic slice. Middle, the analogous slices obtained after 3 days of reperfusion and 10 to 30 minutes after cessation of an infusion of contrast material. Large areas of delayed contrast enhancement can be seen and represent areas of infarction. Right, a reproduction of the tracings obtained of the ventricular slices at the time of autopsy and after staining with 2,3,5 triphenyltetrazolium chloride. The areas of gross infarction are shaded. These slices were obtained ex vivo and are analogous, but not identical, to the tomographic slices shown in the lefl and middle panels. The diagrams and scans are oriented such that the anteroapical wall of the left ventricle is at the top, the right ventricle is to the left and the lateral wall of the left ventricle is on the right. The dog is viewed in a caudocranial direction.

THE AMERICAN JOURNAL OF CARDIOLOGY

Volume 53

283

Quantitation of MI size is important in assessing clinical prognosis after MI and in evaluating the effects of interventions designed to limit myocardial damage during &hernia. 23-27 Although there has been much interest in the ability of early reperfusion to limit MI size, relatively few studies have used CT to assess this, and these studies have primarily been directed toward the detection of myocardial edema.28T2g Therefore, this study was designed to assess the use of prospectively ECG-gated and nongated CT scanning for the detection of ischemia and the size of MI as well as to follow the temporal changes in global and regional LV function attending total coronary occlusion and subsequent reperfusion. Methods Protocol: The experiments were conducted in 16 conditioned mongrel dogs that weighed 22 to 36 kg. Each dog was anesthetized with i.v. sodium pentobarbital (25 mg/kg) before left thoracotomy. A hydraulic occluder was placed around the proximal left anterior descending coronary artery (LAD). The catheter was burrowed subcutaneously and externalized. The wound was aseptically closed and the dogs were allowed to recover for 4 to 7 days. Before each scan, the dogs were premeditated with subcutaneous morphine sulfate (2.5 to 3.0 m/kg), anesthetized with i.v. sodium pentobarbital (25 mg/kg) and ventilated with a Harvard respirator (14 to 16 breaths/min, tidal volume 12 to 15 ml/kg). Control CT scans were obtained 3 to 7 days after recovery from surgery and 3 to 4 days before coronary occlusion. Ten dogs were subjected to a permanent LAD occlusion. The hydraulic occluder was inflated with the dog in the scanner, and scans were performed immediately after occlusion and then at 3 to 4 days. Six other dogs subjected to a 2-hour occlusion were scanned immediately after occlusion, at the end of the occlusion period, after 1 hour of reperfusion and then again 3 days later. After the final scan, the reperfused dogs were killed and postmortem examinations performed. Antemortem angiography confirmed LAD patency in this group. Imaging technique: Ungated scans: All studies were performed on a Technicare 2020 whole body scanner with an individual scan time of 2 seconds. On the day of infarction, 10 minutes after inflation of the hydraulic occluder, a contrast infusion of sodium meglumine diatrizoate (Renografin-76@) was administered over 10 minutes through a leg vein at a rate of 6.1 ml/min. At the end of this time ungated scans were obtained from the apex to the base of the heart during held inspiration (6 to 7 slices, taking 60 to 80 seconds) (Fig. 11, followed by gated scans at the mid-LV level (discussed subsequently). Thereafter, the infusion of contrast was stopped. At 10 to 30 minutes after cessation of contrast infusion, a series of ungated scans was obtained from the apex to the base of the left ventricle. The scans taken during infusion of contrast media were used to quantify the perfusion defect, whereas the scans performed 10 to 30 minutes after infusion were used to quantify the region of contrast enhancement around and within areas of necrosis (size of the MI itself) .I;3 In the 6 dogs undergoing reperfusion, scans to assess enhancement of MI were also obtained at the end of the 2-hour occlusion period and again after 1 hour of reperfusion. Each set of ungated images (apex to base) was then printed on transparent film along with spatial calibration markers. In each dog, 2 different regions were planimetered for every ungated slice acquired on each scan day. One region was de-

204

COMPUTERIZED

TOMOGRAPHIC

ASSESSMENT

OF REPERFUSION

fined as the “initial perfusion defect,” representing an area of ischemic jeopardy. This area of the myocardium has lower attenuation than the remaining myocardium on images obtained during contrast infusion. A second region represented the area of myocardium demarcated by the outer margin of differential contrast enhancement at the periphery of the MI zone on delayed scans after cessation of contrast infusion. These delayed scans (ungated) were obtained at 10 to 30 minutes after the infusion was stopped. This area represents irreversibly infarcted tissue. 4~18Planimetry was accomplished using a Hewlett-Packard computer/digitizer system (9825/ 98/4A) by manually tracing the MI region from the scan, and the computer then calculated the volume contained within the closed loop. The LV slices were each assumed to be cylindrical. The volume of involved myocardium was obtained by summing the volumes for each individual slice from apex to the base of the heart (each slice was 1 cm thick). The volume of LV mass was also calculated from the ungated images. This was accomplished by planimetering the total LV myocardium from the apex to the level of the LV outflow tract. All volume measurements were converted to mass assuming a density of 1.05 g/cm”. Immediately after the final scans, each dog was killed and the heart removed and sectioned at l-cm intervals from apex to base along the major axis of the left ventricle. The slices were incubated in a 1% solution of 2,3,5 triphenyltetrazolium chloride (TTC) (Sigma Chemicals) in a phosphate buffer for 10 minutes at 37°C. The endocardial and epicardial borders of the LV wall and the area of grossly visible MI were traced onto clear film overlays (Fig. 1). LV mass and MI size were measured as described. Gated scans: Imaging was performed using a 2020 Technicare scanner equipped with a prototype prospective gating system.lm:’ Gated CT scans were obtained as previously describedlm:j by acquiring x-ray data relative to a preselected fraction of the electrocardiographic RR interval called the biologic window. The biologic window was set at 10% of the RR interval in this study. The prospective gating technique programmed sequential scans, one 2-second scan each 7 seconds, to obtain all 360’ of data for each 10% fraction of the cardiac cycle. Six to 8 scans were required to obtain a full complement of gated angular data for the preselected biologic window. Because each scan actually collected data during the entire 360’ rotation of the gantry, multiple images could be reconstructed (10 per RR interval in this study), all having the same preselected interval width depicting the cardiac cycle from end-diastole to end-diastole. Unlike retrospective gating, prospective gating minimizes the time and number of scans necessary to image the preselected fraction of the RR interval width with small or no gaps in the angular data. Because each

FIGURE 2. End-diastolic (left) and end-systolic (right) frames from a gated series. Endocardial outlines for measurement of midventricular areas and percent change are shown. The ischemic bed produced by occlusion of the left anterior coronary artery is in the anteroapical area of these figures (top), where wall thickness was measured.

biologic window was 10% of the RR interval and the dogs maintained a heart rate of 80 to 150 beats/min during scanning, the end-diastolic and end-systolic gating intervals were 40 to 75 ms in duration. Gated scans were performed over the mid-LV level during held inspiration. The scans were acquired during a steady state of contrast infusion while the heart rate remained stable. The approximate radiation dose for a gated series of scans (usually 8 scans) was 1 to 2 rads in a precisely collimated x-ray beam 1 cm thick. The 10 frames at the mid-LV level representing 10% portions of the average heart beat were displayed along with reference distance markers, and the end-diastolic and endsystolic frames were selected. The end-diastolic figure was the initial frame that corresponded to the 10% interval encompassing the QRS complex, whereas the end-systolic frame was identified as the one with the smallest luminal area. Enddiastolic and end-systolic areas were planimetered (Fig. 2). The change in mid-LV cross-sectional area was calculated as the difference between end-diastolic and end-systolic chamber area divided by end-diastolic area and expressed as a percentage. LV wall thickness was measured by placing a movable cursor across the LV wall. The measurement was given automatically by the display program. The anterior wall was measured at the apex of the LV chamber. The same position was measured on the end-diastolic and end-systolic images. Thickening was calculated as the difference between endsystolic and end-diastolic thickness, divided by end-diastolic thickness and expressed as a percentage. Statistics: Variables within each group were analyzed by repeated measures of analysis of variance. Differences between groups were assessed by the unpaired Student t test. Differences were considered significant at p <0.05. Correlations were obtained by linear regression analysis. Reproducibility: Inter- and intraobserver variability of measurement of MI size, total LV mass, wall thickness and mid-LV luminal areas were assessed in 6 randomly selected

PERMANENT OCCLUSION GROUP

REPERFUSION GROUP

60 -

* p< 05

vs ACUTE PERFUSION DEFECT

ACUTE PERFUSION

ENHANCEMENT AT

ACUTE PERFUSION

DEFECT

THREE OAYS

DEFECT

ENHANCEMENT AT THREE DAYS

FIGURE 3. Relation between initial perfusion defect and infarct size (enhancement) at 3 days.

January 15, 1984

by 1 observer on 2 occasions separated by greater than 3 weeks and by a second independent observer. cases

Results Analysis of ischemic zone and infarct size: The control dogs subjected to permanent occlusion had an initial perfusion defect (zone of ischemic jeopardy) of 18.6 f 4.0 g (23.0 f 3.4% of LV mass) (mean f standard deviation). After 3 to 4 days an area of delayed contrast enhancement was present in each animal, and was larger than the initial perfusion defect in 9 of 10 dogs (mean of 30.1 f 9.9 g, p <0.05). In the dogs subjected to reperfusion after a 2-hour occlusion, the initial perfusion defect was 7.7 f 4.9 g (9.0 f 7% of LV mass), which was significantly smaller than the initial perfusion defect of the other group (p
THE AMERICAN

JOURNAL OF CARDIOLOGY

Volume 53

205

2 hours after occlusion and in 5 of the 6 at 1 hour after reperfusion (3 hours after occlusion). In view of the subsequent detection of MI at 3 days after occlusion in all but 1 dog, assessment of necrosis during the acute stages (within 3 hours of acute coronary occlusion) was considered unreliable. The correlation between the MI size measured at necropsy and the MI measured from the CT scan just before autopsy was 0.95 (p
Assessment of global and regional contractile function: Figure 6 and Table I show the sequential

changes in end-diastolic and end-systolic mid-LV luminal areas and the percent changes in mid-LV area. End-diastolic area showed no significant change over the course of acute occlusion and reperfusion or at 3 days after reperfusion. The end-systolic mid-LV chamber area increased immediately after coronary occlusion and was significantly larger than control by the end of the 2-hour period of ischemia (12.8 f 7.1 vs 8.5 f 3.4 cm2, p <0.05). At the end of 1 hour of reperfusion, the end-systolic area declined partially toward control values, but by 3 days it remained significantly enlarged (12.3 f 5.2 cm2, p <0.05 vs control). Similar findings were noted in the permanently occluded group, that is, between acute occlusion and the 3-day follow-up, end-diastolic mid-LV chamber area showed no significant change (17.7 f 4.3 cm2 at control, 14.8 f 2.6 cm2 at 3 days), whereas the end-systolic chamber area increased significantly during acute occlusion (11.5 f 4.0 cm2 at control vs 13.4 f 2.9 cm2, p <0.05). The percent change in mid-LV chamber area generally mirrored the changes in end-systolic area. The percent change decreased acutely in the reperfusion group, from 44.5 f 7.0 to 28.3 f 12.6% (p <0.05). Just before the end of the occlusion period this value re-

l

0

I 5

I

1

I

10

15

20

I

25

I

30

MYOCAROIALINFARCTSIZE (autopsy) (gm) FIGURE 4. Left, a small initial perfusion defect during acute occlusion in 2 tomographic slices (arrows). Right, after 3 days, the postcontrast infusion scan demonstrated no enhancement and necropsy revealed no evidence of infarction.

FIGURE 5. Correlation between infarct size as measured morphometrically and from the contrast enhanced areas seen on the tomograms 3 days after reperfusion and just before autopsy. The regression line is y = 0.89x - 1.4, r = 0.95, p
266

COMPUTERIZED

TOMOGRAPHIC

ASSESSMENT

OF REPERFUSION

mained low (26.8f 11.7%, p <0.05). It increased slightly after 1 hour of reperfusion (31.7 f l&7%), but was still significantly lower than the control value (p <0.05). After 3 days of reperfusion, the percent mid-LV area change remained lower than control (30.2 f ll.O%, p cO.05). The control group demonstrated a similar directional decrease over the 3 days of total coronary occlusion: 35.9 f 8.8% at control vs 27.1 f 7.3% after acute occlusion (p <0.05) and 27.7 f 10.7% at 3 days (p <0.05 vs control). Figure 7 and Table II show the changes in wall thickness and extent of thickening in the reperfusion group. The end-diastolic wall thickness did not change significantly over the time course of the protocol, but the end-systolic wall thickness was significantly thinner immediately after occlusion (9.2 f 2.1 vs 15.7 f 2.3 mm at control, p
2 hours of ischemia (-1.0 f 14.6%, p <0.05) and 1 hour of reperfusion (-4.5 f 28.5%, p <0.05). At 3 days, all but 1 dog showed persistent impairment of wall thickening. Although the mean thickening at this point was low (17.3 f 24.7%), this was not significantly different from the control value. In contrast, dogs with permanent total occlusion showed an acute decrease in extent of wall thickening (46.2 f 16.5% at control vs 1.6 f 9.0% after acute occlusion, p
Discussion The present study demonstrates that the temporal alterations in global and regional LV function resulting from ischemia and reperfusion can be assessed with CT. CT also defines the relation between the initial perfusion defect and eventual MI size, as shown by the volume of delayed contrast enhancement. Consequently, this may be useful in assessing the effects of intervention on the natural history of MI. The ability to measure both the area at risk and the infarcted region by this technique circumvents many of the problems in attempting to assess the effects of reperfusion using other

20r END-DIASTOLE

* pc.05 ** p-c.01 *

END-SYSTOLE

5

I 50r

;

50-

* pc.05

vs CONTROL

2 $

40-

% a g 30it @z

*

* “:-.--; CONTROL

vs CONTROL vs CONTROL

g

40-

Z 5 z

3020-

k

lo-

g

o-

:

-lO-

T

* I

I I I I 1 HOUR 3 OAYS PRIORTO ACUTE AFTER AFTER OCCLUSION RELEASE REPERFUSIONREPERFUSIOb OF OCCLUSION

FIGURE 6. Temporal changes in global function after occlusion and reperfusion.

I



L

CONTROL

ACUTE OCCLUSION

I

1

I

3 DAYS 1 HOUR PRIOR TO AFTER AFTER RELEASE REPERFIJSIONREPERFUSION OF OCCLUSION

FIGURE 7. Sequential changes in wall thickness and extent of wall thickening after coronary occlusion and reperfusion.

January 15. 1984

imaging techniques and enzyme release methods for estimating MI size. This study is analogous to a recent study that used cross-sectional echocardiography to define the functional impairment caused by a 2-hour period of coronary occlusion followed by reperfusion.15 In that study, early and profound deterioration in systolic wall thickening after coronary occlusion showed very little improvement after 4 hours of reperfusion. After 4 days of reperfusion, near-maximal functional recovery had occurred and only slight improvement ensued thereafter. The preocclusion level of contractile function was not attained even after 14 days of reperfusion. The results of the present investigation show comparable changes in wall thickening over a S-day period of reperfusion after a similar 2-hour interval of coronary occlusion. An assessment of area at risk of infarction in the study by Ellis et all5 could not be provided solely by 2-dimensional echocardiography. Direct left atrial injection of Monastral blue just before killing and subsequent staining with TTC was required to define this measurement. Determination of the area at risk in the present study could be attained solely by the imaging technique during intravenous contrast infusion in the early phase after coronary occlusion. The imaging protocol outlined in this report dictates that the initial perfusion defect is quantitated 20 to 30 minutes after abrupt termination of coronary flow. Although the recruitment of collaterals in the canine model is known to be variable over time and might therefore significantly alter the area of ischemic jeopardy,14*30 the greatest quantitative increase in collateral myocardial blood flow as assessed by radiolabeled microspheres after abrupt coronary occlusion has been shown to occur within several minutes and shows little change thereafter.31,32 In addition, previous studies show good correlations between the zone of low x-ray attenuation detected on CT scans after coronary occlusion with similar zones measured by independent techniques that reflect regional differences in coronary flow.lsJ7 Thus, the initial perfusion defect as determined in this study at 20 to 30 minutes after acute coronary occlusion appears to provide a reasonable estimate of jeopardized myocardium serving as a base from which to assess the effects of intervention on subsequent MI size. Later after coronary occlusion the perfusion defect could not

THE AMERICAN JOURNAL OF CARDIOLOGY Volume 53

be considered an accurate estimate of jeopardized area due to the interfering effects of the contrast material previously administered: Late diffusion of contrast material into the jeopardized region could be expected to obscure a perfusion defect on subsequent injections of contrast media. That the delayed, postcontrast infusion, myocardial enhancement areas noted on the CT scans represent ischemically damaged tissue has been demonstrated in previous reports.4Jss33 Scanning electron microscopy and energy dispersive x-ray microanalysis of fresh frozen cryosections of myocardium have shown that iodine was located in ischemically damaged but was excluded from normal myocardial cells.r8 These iodine-laden cells also showed a lower K/Nathan normal cells, implying a disruption of membrane function or integrity. Another possible mechanism for preferential iodine accumulation is the increased leakage of blood constituents including contrast material into infarcted areas secondary to local microvascular damage.34,35 In addition, the net amount of iodine in a particular myocardial area is partially a result of an ongoing flux of contrast material in and out of the zone. Thus, iodine concentration in the myocardium is partially dependent on a residual level of coronary flow provided by either reperfusion, collateral flow or both.5J2 These rates of flux are obviously slower in the infarcted areas and this fact may further contribute to the preferential, delayed contrast enhancement of areas of cell damage. As previously shown, there is a good correlation between the size of enhancement areas on CT scans and morphometric MI size measured at necropsy.7JoJ3*36 The present study provides additional evidence that the area of delayed contrast enhancement is a good index of MI size even after reperfusion. Therefore, the relation between the initial perfusion defect and the late enhancement image may be important in assessing both the natural and the modified progression of infarction. This study shows that the enhancement area after 3 days was larger than the initial perfusion defect in the group of animals with permanent occlusion but not in animals that were reperfused. However, even in the reperfusion group, dogs with resultant large infarctions also showed an increase in the enhancement area at 3 days compared with the initial perfusion defect. These

TABLE II TABLE I

Changes in Mid-Lefl Ventricular Chamber Area from Gated Scans Permanent Occlusion

Enddiastolic chamber area (cm? End-systolic chamber area (cm21 Percent change in chamber area

Control Acute occlusion 3 days Control Acute occlusion 3 days Control Acute occlusion 3 days

17.7 18.3 14.8 11.5 13.4 10.6 35.9 27.1 27.7

Values are mean & standard deviation. p
f f f f f f f f f

4.3 2.7 2.0 4.0 2.9’ 2.1 8.8 7.3+ 10.7+

Changes in Wall Thickness and Extent of Wall Thickening in the lschemic Zone Assessed from Gated Scans Permanent Occlusion

Reperfusion 15.3 16.3 17.7 8.5 12.0 12.3 44.5 28.3 30.2

f f f f f f f f f

5.4 5.5 5.9 3.4 5.5 5.2+ 7.0 12.6+ 1 l.O+

207

End-diastolic wall thickness (mm) End-systolic wall thickness (mm) Extent of wall thickening (%)

Control Acute occlusion 3 days Control Acute occlusion 3 days Control Acute occlusion 3 days

6.8 6.5 9.6 8.3 6.6 9.9 46.2 1.6 2.4

Values are mean f standard deviation. p
f f f f f f f f f

1.2 1.8 1.7 2.3 2.1’ 2.2 16.5 9.0” 10.1’

Reperfusion 11.0 9.3 11.7 15.7 9.2 13.5 43.3 -0.2 17.3

f f f f f f f f f

1.8 2.7 1.5 2.3 2.1’ 2.3 14.4 10.6+ 24.7

288

COMPUTERIZED TOMOGRAPHIC

ASSESSMENT

OF REPERFUSION

results likely reflect the combined effects of myocardial edema, hemorrhage and inflammation associated with infarction, which could be expected to expand the mass of the infarcted tissue at 3 to 4 days. This and subsequent resorption, organization and scar contraction also account for the naturally occurring change in the anatomic evolution of MI size over time.35,37 This factor must be considered when attempting to ascribe changes in MI size to interventions and also when using functional parameters, such as the extent of wall thickening, as a basis for comparing interventions because such indexes are also affected by wall edema, hemorrhage and inflammation. These are general limitations common to all such studies. A particular limitation of this study is the generally smaller initial perfusion defects in the reperfusion group during coronary occlusion. This is likely due to the known heterogeneity of MI size in the canine model even after similar periods of coronary occlusion. However, despite even this limitation the occurrence of a spectrum of MI from transmural to patchy necrosis, to subendocardial infarction and no MI was seen only in the reperfusion group, which contrasted with the transmural MIS seen uniformly in the group subjected to permanent coronary occlusion. Furthermore, the identical degree of acute functional impairment of the 2 groups which persisted in the permanent occlusion group but not in the reperfused group is also evidence of a salutary effect of reperfusion despite initial differences in the size of the perfusion defect. These functional variables were measured from the center of the initial perfusion defect, which was transmural in all dogs, even in the dog with the smallest defect and no final MI (Fig. 4). Although a perfusion defect can be detected relatively early after coronary occlusion by CT, a systematic evaluation of the ability to consistently detect necrosis at an early stage after coronary occlusion has not been performed. The earliest detection of necrosis by delayed contrast enhancement CT has been reported at 3.5 hours after coronary occlusion.s6 In the present study, contrast enhancement was examined after 2 hours of coronary occlusion and after 1 hour of reperfusion (3 hours after coronary occlusion), but enhancement was either absent or equivocal. However, because contrast infusions were relatively closely spaced (at acute occlusion as well as at 2 and 3 hours later), the disparate washin and washout rates in normal and ischemic tissue may have been obscured by the repeated infusions. Studies using a single infusion of contrast agent at variable intervals after coronary occlusion will be important in further elucidating this apparent limitation of CT for early MI detection. This feature of CT imaging of infarctions will also affect clinical applications. A final limitation of this technique in assessing myocardial salvage is dependent on spatial resolution. For example, 1 dog in the reperfusion group had a moderate area of patchy necrosis, which appeared as an area of diffuse and faint contrast enhancement. The use of the outer perimeter of this area to demarcate the final MI size was likely to include infarcted cells intermingled with normal cells that were salvaged by reperfusion but

which could not be excluded in the measurement because of limits of spatial resolution. Thus, the degree of salvage in this infarction pattern may be underestimated by CT. In summary, prospectively ECG-gated and nongated CT can be used to allow minimally invasive assessment of evolutionary changes in the infarct size and concomitant functional alterations as well as to assess the effects of interventions, including reperfusion, on these variables. Future investigations and technical improvements in CT imaging will elucidate the ultimate applicability of these methods to clinical practice.

References 1. Mattrey RF, Higgins CB. Detection of regional myocardial dysfunction during ischemia with computerized tomography: documentation and physiologrc basis. Invest Radio1 1982;17:329-335. 2. Mattrey RF, Slutsky RA, Long SA, Higgins CB. In viva assessment of left ventricular wall and chamber dynamics during transient myocardial ischemia using prospectively ECGgated computerized transmission tomography. Circulation 1983;67:1245-1251. 3. hlancini GBJ, Peck WW, Slutsky RA, Mattrey RF, Higgins CB. Pharmacologically induced changes in wall thickening dynamics and midventricular volumes in dogs assessed by prospectively gated computerized tomography. Am .I Cardiol 1983:51:1739-1743. 4. Adams DF, HessefSJ, Judy PF, Stein JA, Abrams HL. Differeing attenuation coefficients of normal and infarcted myocardium. Science 1976; 192:467-468. 5. Higgins CB, Sovak M, Schmidt W, Siemers PT. Differential accumulation of radiopaque contrast material in acute myocardial infarction. Am J Cardiol 1979;43:47-51. 6. Higgins CB, Sovak hl, Schmidt W, Siemers PT. Uptake of contrast materials by experimental acute myocardial infarctions: a preliminary report. Invest Radio1 1978;13:337-339. 7. Higgins CB, Siemers PT, Schmidt W, Newell JD. Evaluation of myocardial ischemic damage of various ages by computerized transmission tomography. Time dependent effects of contrast materials. Circulation 1979; 60:284-29 1. a. Newell JD, Higgins CB, Abraham JL, Kelley MJ, Schmidt WS. Haialer F. Computerizediomographic appearance of evolvingmyocardiaf infarctions. Invest Radio1 1980:15:207-214. 9. Higglns CB, Siemers PT, Newell JD, Schmidt W. Role of iodinated contrast material in the evaluation of myocardial infarction by computerized transmission tomography. Invest Radio1 1980;15:Sl76-S182. 10. Doherty PW, Lipton MJ, Beringer WH, Skioldebrand CG, Carlsson E, Redlngton RW. Detection and quantitation of myocardial infarction in viva using transmission computed tomography. Circulation 1981:63:597606. 11. Newell JD, Higgins CB, Abraham JL. Uptake of iodinated contrast material ;);_thh, rschemrcally damaged myocardial cell. Invest Radio1 1962; 17: 12. Higgins CB, Hagan PL, Newell JD, Schmidt WS, Halgler FH. Contrast enhancement of myocardial infarction: dependence on necrosis and residual blood flow and the relationship to distribution of scintigraphic imaging agents. Circulation 1982;65:739-746. 13. Slutsky RA, Mattrey RF, Long SA, Higgins CB. In viva estimation of myocardial infarct size and left ventricular function by prospectively gated computerized transmission tomography. Circulation 1983;67:759-765. 14. Kerber RE, Marcus ML, Ehrhardt J, Wilson R, Abboud FM. Correlation between echocardiographically demonstrated segmental dyskinesis and regional myocardial perfusion. Circulation 1975;52:1097-1104. 15. Ellis SG, Henschke Cl, Sandor T, Wynne J, Braunwald E, Kloner RA. Time course of functional and biochemical recovery of myocardium salvaged by reperfusion. J Am Coll Cardiol 1983;1:1047-1055. 16. Murphy ML, Peng CF, Kane JJ, Straub KD. Ventricular performance and biochemical alteration of regional ischemic myocardium after reperfusion in the pi Am J Cardiol 1982;50:621-828. 17. Hessel 0 J, Adams DF, Judy PF, Fishbein MC, Abrams HL. Detection of myocardial ischemia in vitro by computed tomography. Radiology 1978; 127:413-418. ia. Abraham JL, Higgins CB, Newell JD. Uptake of ixtinated contrast material in ischemic myocardium as an indicator of loss of cellular membrane integrity. Am J Pathol 1980;101:319-327. 19. Whalen DA, Hamilton DG, Ganote CE, Jennings RB. Effect of a transient period of ischemia on myocardial cells. I. Effects on cell volume regulation. Am J Pathol 1974;74:381-397. 20. Kloner RA, Ganote CE, Whalen DA, Jennings RB. Effect of a transient period of ischemia on myocardial cells. II. Fine structure during the first few minutes of reflow. Am J Pathol 1974;74:399-413. 21. Willerson JT, Scales F, Mukherjee A, Platt M, Templeton GH, Fink GS, Buja LM. Abnormal myocardial fluid retention as an early manifestation of ischemic injury. Am J Pathol 1977;87:159-181. 22. DiBona DR, Powell WJ. Quantitative correlation between cell swelling and necrosis in myocardial ischemia in dogs. Circ Res 1980;47:653-665. 23. Sobel B, Bresnehan G, Shell W, Yoder R. Estimation of infarct size in man

January 15, 1984

and its relation to prognosis. Circulation 1972;46:640-648. 24. Maroko P, Libby P, Cove8 J, Sobel 8, Ross J, Braunwald E. Precordial ST segment evaluation mapping: an atraumatic method for assessing afterations in the extent of myocardial ischemic injury. Am J Cardiol 1972;29:223230. 25. Shell W, Sobel B. Protection of jeopardized ischemic myocardium by reduction of ventricular afterload. N Engl J Med 1974;291:481-486. 26. Braunwald E, Maroko P. The reduction of infarct size. An idea whose time (for testing) has come. Circulation 1974;50:206-209. 27. Braunwald E,, Maroko P, Libby P. Reduction of infarct size following coronary occlusron. Circ Res 1974;35:suppl lll:lll-192-111-201. 28. Powell WJ, Wittenberg J, Maturi RA, Dinsmore RE, Miller SW. Detection of edema associated with mvocardial ischemia bv comouterized tomooraohv in isolated arrested canine-hearts. Circulation i977;55:99-108. * ’ 29. Wlttenberg J, Powell WJ, Dinsmore RE, Miller SW, Maturi RA. Computerized tomography of ischemic mvocardium: auantitation of extent and severity of edema in an in vitro canine modei. Invest Radio1 1977;12: 215-223. 30. Marcus ML, Kerber RE, Ehrhardt J, Abboud FM. Effects of time on volume and distribution of coronary collateral flow. Am J Physiol 1976;230:279285.

THE AMERICAN JOURNAL OF CARDIOLOGY

Volume 53

289

31. Jugdutt BI, Becker LC, Hutchins GM. Early changes in collateral blood flow during myocardial infarction in conscious dogs. Am J Physiol 1979;237: H371-H380. 32. Patterson RE, Jones-Collins BA, Aamodt R, Ro YM. Differences in collateral myocardial blood flow following gradual vs abrupt coronary occlusion. Cardiovasc Res 1983;17:207-213. 33. Newell JD, Hlgglns CB, Mayer WW, Haigler FH, Werner FG, Gerber KH. Computerized tomographic appearance of the myocardium after reversible and irreversible ischemic injury (abstr). Invest Radio1 1981;16:398. 34. Armiger LC, Gavls JB. Changes in microvasculature of ischemic and infarcted myocardium. Lab Invest 1975;33:51-56. 35. Fishbein MC, Maclean D, Maroko PR. The histopathologic evolution of myocardial infarction. Chest 1978;73:843-849. 36. Huber GJ, Lapray JF, Hessel SJ. In viva evaluation of experimental myocar;ral Infarcts by ungated computed tomography. AJR 1981;136:46937. Reimer KA, Jennings RB. The changing anatomic reference base of evolving myocardial infarction. Underestimation of myocardial collateral blood flow and overestimation of experimental anatomic infarct size due to tissue edema, hemorrhage, and acute inflammation. Circulation 1971; 80:866-878.