Mgnelic Resonance Printed in the USA.
Imaging, Vol. 9, pp. ISS-lS8, All rights reserved.
1991 Copyright
0730-725x/91 %3.00 + .oo 0 1991 Pergamon Press plc
l Original Contribution
SERIAL MAGNETIC RESONANCE IMAGING IN PATIENTS FOLLOWING ACUTE MYOCARDIAL INFARCTION C. THOMPSON,PETER LIU, THOMAS J. BRADY, ROBERT D. OKADA, AND DONALD L. JOHNSTON Cardiac Unit, Department of Medicine and Department of Radiology, Massachusetts General Hospital and Harvard Medical School, Boston, Massachusetts, 02138, USA RANDALL
The detection of serial changes in magnetic resonance (MR) signal intensity of the heart following acute myocardial infarction may provide a useful method of characterizing tissue healing. Fourteen patients with acute Q-wave infarction underwent T2-weighted, spin-echo cardiac imaging during hospitalization, followedby one or more additional MR studies (total 31) over a 6- to 27-wk period (mean: 3 mo). Visual assessment of the images demonstrated a gradual reduction in signal intensity and localization of the bright signal to the subendocardium of the infarction region over the three-mo study period.A quantitative measurement of signal intensity (infarction/normal myocardium) fell from 1.81 f 0.42 on the initial study to 1.34 f 0.37 (p < 0.05) at a-mean of 14 wk. Two patients had an increase in signal intensity on the follow-up study and both patients had been readmitted with acute coronary syndromes. In summary, characterization of changes in signal intensity may provide a useful method of assessing myocardial healing following acute myocardial infarction. Further studies are indicated to determine the prognostic significance of these parameters. Keywords:
Magnetic resonance imaging; Acute myocardial
INTRODUCTION
infarction.
additional insight into the pathophysiology healing myocardium.
Acute myocardial infarction is readily visualized with standard spin-echo, magnetic resonance (MR) imaging techniques. Studies performed in man have demonstrated increased signal intensity in the infarction region when the images were obtained relatively early (3-30 days) after the event.’ However, there is relatively little information available regarding the evolutionary changes in signal intensity following myocardial infarction. The only previous study was performed by Dilworth et al., in which 9 patients were imaged at 35 days and lo-14 days after myocardial infarction.* This study showed that significant serial changes in infarction signal intensity did not occur over this short observation period. The purpose of the present study was to measure the evolutionary changes in MR signal intensity that occur following acute myocardial infarction in man. Knowledge of these changes may eventually provide
of the
METHODS Patient Selection Fourteen patients (11 males, 3 females) with a recent Q-wave myocardial infarction (11 anterior, 3 inferior) were enrolled in the study. The patients were chosen from a larger group of patients having cardiac MR imaging following acute myocardial infarction. They were selected for good image quality of the initial study and the presence of increased signal in the segment of the left ventricle that corresponded to the infarction location by electrocardiogram (ECG). Other inclusive criteria included the ability to give informed consent and a stable medical status without important arrhythmias. The initial study was obtained at a mean of 8.5 days (range: 4-14 days) after acute infarction, and the follow-up MR study (last study obtained) was
RECENED 7/18/90; ACCEPTED9/17/90.
West 16B, The Mayo Clinic, 200 First St., Rochester, MN
Address all correspondence to Donald L. Johnston, MD,
55905, USA. 155
156
Magnetic Resonance Imaging l Volume 9, Number 2, 1991
at 12 wk (range: 6-27 wk). Fourteen patients had 2 studies, and 3 patients had 3 studies. performed
Imaging Technique ECG-gated, spin-echo imaging of the heart was performed as previously described. ’ In brief, a superconducting magnet (Technicare, Solon, Ohio) operating at a field strength of 0.6 Tesla was used for all studies. A multislice, short-axis series of images was acquired along the left ventricle from base to apex. Contiguous single-slice, spin-echo (TE = 60 msec) images were then obtained at end-systole or end-diastole (determined from the ECG) along the length of the ventricle. The ECG-gate delay time for acquisition was always the same for initial and follow-up studies. A single-slice, single-echo technique minimized the amount of blood pool signal present in the ventricular chamber and permitted better definition of the endocardium. Data Analysis Qualitative Assessment. All MR studies were analyzed by three reviewers and a visual estimate was made of changes in signal intensity from the initial study to the follow-up study. The images were specifically examined for changes in infarction signal intensity, differences in size, location and distribution of the infarction, and for the presence of myocardial thinning. A change in the size of the infarction region was determined by estimating the circumferential extent of the bright infarction signal on the initial study and comparing it to subsequent studies. These features were assessed for each tomographic slice. Regional left ventricular thinning was determined by comparing the infarction segment with the opposite myocardial wall thickness. Quantitative Analysis. The infarction signal intensity was measured by placing a circular region of interest cursor of defined size over the center of the infarction region. The signal intensity of the opposite myocardial segment was then measured in a similar fashion. Three determinations of each measurement were averaged to generate the infarction/normal myocardium signal intensity ratio. This ratio was obtained in a similar fashion on follow-up studies. Care was taken to perform the measurements on nearly identical tomograms. In 5 patients, signal intensity ratios were not determined due to difficulty in obtaining an accurate and reproducible region of interest on the follow-up studies. Problems were encountered in these instances due to marked wall thinning or regres-
sion of the infarction region to a thin portion of the endocardium. Statistical Analysis Differences in signal intensity ratios were determined by performing a Student t-test. P < 0.05 was considered significant. All values are mean + standard deviation. RESULTS Visual Assessment (Fig. I) Thirty-one studies were performed in 14 patients. In 13 patients, signal intensity remained increased in the same segment on the first follow-up study (Fig. 1). In one patient imaged at 26 wk, infarction signal intensity was normal and the left ventricular wall was thinned. Of the 3 patients with 3 studies each, 2 had increased signal intensity on the second follow-up study and one had normal signal at 27 wk. For the 13 patients with increased infarction signal at first followup, the region was smaller in size in 11 patients and unchanged in size in 2 patients. For the latter 2 patients, both had early follow-up studies (3 and 6 wk). The infarction was noted to have receded to the subendocardium in 9 patients as compared to the initial study. No change in distribution was noted in the remaining four patients. Myocardial thinning was present in 10 of 14 patients on the initial study and was apparent in 3 additional patients on follow-up studies. Wall thinning appeared to progress slightly on follow-up in six patients, three of whom had some initial thinning and three of whom showed no initial thinning. Myocardial Signal Intensity Ratio (Fig. 2) Figure 2 illustrates the signal intensity ratios measured for initial and follow-up MR studies in nine patients. The mean initial ratio of 1.81 + 0.42 decreased to 1.34 + 0.37 (p < 0.05) at a mean of 14 wk. Of interest, two patients had an increase in the signal intensity ratio on follow-up and both patients had been readmitted to the hospital at the time of the follow-up study with subsequent acute coronary events. In one patient, reinfarction occurred while the other patient presented with unstable angina and was found to have a left ventricular aneurysm. DISCUSSION This study is the first to demonstrate in man that evolutionary changes occur in MR signal intensity following acute myocardial infarction. The findings in-
Serial MRI following acute myocardial infarction 0 R.C.
THOMPSON ET AL.
157
09 Fig. 1. Spin-echo (TE = 60 msec) images of a patient 10 days (A) after acute myocardial infarction and at 16 wk (B). Signal intensity was increased in the anterior wall (arrow). At 16 wk, signal intensity was reduced compared to 10 days, and hrfarction size was smaller and more localized to the endocardium.
dicate that tissue healing proceeds from the epicardium to the endocardium over a period of time exceeding three mo. This pattern of healing was most likely due to a relatively poorly formed endocardial coronary collateral circulation that enhanced the severity of tis-
2.6 2.4 2.2 0 ‘i; t!i
2.0
.g
1.6
E
9) E .ia
_& c/l
1.6
1.4 1.2 1.0 0.6
0.6 0.4 0.2 0
I1
0
1
4
I
I,
6
I
12
I
16
I
I
20
I
I,
24
Time (weeks) Fig. 2. Infarction signal intensity ratio compared to number of weeks after myocardial infarction. The average signal intensity ratio for nine patients fell from 1.81 & 0.42 on the initial study to 1.34 + 0.37 at a mean of 14 wk.
sue injury and slowed formation of mature scar tissue compared to the epicardium. Our present findings are consistent with previously reported animal studies. Pflugfelder et al. demonstrated serial changes in infarction signal intensity in the dog following permanent ligation of the left anterior descending coronary artery.3 Signal intensity of the infarction increased from day 1 to days 4-6, then gradually decreased, but was still elevated in most dogs at 20 days. These changes were attributed to infarction healing. Wisenberg et al. measured signal intensity in dogs imaged after 1-3 hr of coronary artery occlusion, followed by 21 days of reperfusion.4 They found that Tl and T2 relaxation times (measured from the images) increased abruptly in the first hour, while T2returned to baseline by day 5. Johnston et al. reported an abrupt increase in T,relaxation time at three days in the infarcted rabbit heart.’ This change corresponded to the appearance of granulation tissue in the region of the infarction. T2returned to normal with the development of mature scar tissue at three mo. Since infarction healing proceeds more rapidly in the experimental animal than in man, it was not surprising that normalization of T2 in this study came before a return to normal in signal intensity in our study. In the present study, infarction signal intensity on the follow-up study was not found to be less than the
158
Magnetic Resonance Imaging 0 Volume 9, Number 2, 1991
surrounding normal myocardium. This was consistent with previous experimental data reported by Johnston et al., in which T, and T2relaxation times of mature scar were similar or only slightly lower than surrounding normal myocardium.5 However, McNamara and Higgins reported a decrease in infarction signal intensity and relaxation times in patients imaged 9 mo to 16 yr after acute myocardial infarction.6 It is possible that additional scar tissue maturation occurs after three mo that results in a further reduction in relaxation times. Further cardiac studies are indicated to assess the prognostic significance of evolutionary changes in MR signal intensity of myocardial infarction. For example, persistent bright signal 2-3 mo after acute myocardial infarction is consistent with residual necrosis or delayed healing, which might result in development of ventricular tachyarrhythmias and aneurysm formation. Alternatively, rapid normalization of infarction signal intensity is suggestive of early healing and might correlate with successful revascularization following acute myocardial infarction. In summary, our study demonstrates that serial changes occur in MR signal intensity over a period of at least three mo following acute myocardial infarction. A gradual reduction in the intensity of the infarction signal, localization of the bright signal to the endocardium, and progressive wall thinning in the area of the infarction are consistent with healing. MR imaging provides a new method of assessing infarc-
tion healing and may be useful for determining the effects of acute interventions following myocardial infarction. REFERENCES 1. Johnston, D.L.; Thompson, R.C.; Liu, P.; Dinsmore, R.E.; Wismer, G.L.; Saini, S.; Kaul, S.; Rosen, B.R.; Brady, T.J.; Okada, R.D. Magnetic resonance imaging during acute myocardial infarction. Am. J. Cardiol. 51: 1059-1065; 1986. 2. Dilworth, R.L.; Aisen, A.M.; Mancini, G.B.J.; Buda, A. J. Serial nuclear magnetic resonance imaging in acute myocardial infarction. Am. J. Cardiol. 59:1203-1205; 1987. 3. Pflugfelder, P.W.; Wisenberg, G.; Prato, ES.; Turner, K.L.; Carroll, S.E. Serial imaging of canine myocardial infarction by in vivo nuclear magnetic resonance. J. Am. CON. Cardiol. 7:843-849; 1986. 4. Wisenberg, G.; Prato, ES.; Carroll, S.E.; Turner, K.L.; Marshall, T. Serial nuclear magnetic resonance imaging of acute myocardial infraction with and without reperfusion. Am. Heart J. 115:510-518; 1988. 5. Johnston, D.L.; Homma, S.; Liu, P.; Weilbaecher, D.G.; Rokey, R.; Brady, T.J.; Okada, R.D. Serial changes in nuclear magnetic resonance relaxation times after myocardial infarction in the rabbit: Relationship to water content, severity of ischemia, and histopathology over a six-month period. Magn. Reson. Med. 8:363-379; 1988. 6. McNamara, M.T.; Higgins, C.B. Magnetic resonance imaging of chronic myocardial infarcts in man. AJR 146: 315-320; 1986.
Imaging, Vol. 9, pp. ISS-lS8, All rights reserved.
1991 Copyright
0730-725x/91 %3.00 + .oo 0 1991 Pergamon Press plc
l Original Contribution
SERIAL MAGNETIC RESONANCE IMAGING IN PATIENTS FOLLOWING ACUTE MYOCARDIAL INFARCTION C. THOMPSON,PETER LIU, THOMAS J. BRADY, ROBERT D. OKADA, AND DONALD L. JOHNSTON Cardiac Unit, Department of Medicine and Department of Radiology, Massachusetts General Hospital and Harvard Medical School, Boston, Massachusetts, 02138, USA RANDALL
The detection of serial changes in magnetic resonance (MR) signal intensity of the heart following acute myocardial infarction may provide a useful method of characterizing tissue healing. Fourteen patients with acute Q-wave infarction underwent T2-weighted, spin-echo cardiac imaging during hospitalization, followedby one or more additional MR studies (total 31) over a 6- to 27-wk period (mean: 3 mo). Visual assessment of the images demonstrated a gradual reduction in signal intensity and localization of the bright signal to the subendocardium of the infarction region over the three-mo study period.A quantitative measurement of signal intensity (infarction/normal myocardium) fell from 1.81 f 0.42 on the initial study to 1.34 f 0.37 (p < 0.05) at a-mean of 14 wk. Two patients had an increase in signal intensity on the follow-up study and both patients had been readmitted with acute coronary syndromes. In summary, characterization of changes in signal intensity may provide a useful method of assessing myocardial healing following acute myocardial infarction. Further studies are indicated to determine the prognostic significance of these parameters. Keywords:
Magnetic resonance imaging; Acute myocardial
INTRODUCTION
infarction.
additional insight into the pathophysiology healing myocardium.
Acute myocardial infarction is readily visualized with standard spin-echo, magnetic resonance (MR) imaging techniques. Studies performed in man have demonstrated increased signal intensity in the infarction region when the images were obtained relatively early (3-30 days) after the event.’ However, there is relatively little information available regarding the evolutionary changes in signal intensity following myocardial infarction. The only previous study was performed by Dilworth et al., in which 9 patients were imaged at 35 days and lo-14 days after myocardial infarction.* This study showed that significant serial changes in infarction signal intensity did not occur over this short observation period. The purpose of the present study was to measure the evolutionary changes in MR signal intensity that occur following acute myocardial infarction in man. Knowledge of these changes may eventually provide
of the
METHODS Patient Selection Fourteen patients (11 males, 3 females) with a recent Q-wave myocardial infarction (11 anterior, 3 inferior) were enrolled in the study. The patients were chosen from a larger group of patients having cardiac MR imaging following acute myocardial infarction. They were selected for good image quality of the initial study and the presence of increased signal in the segment of the left ventricle that corresponded to the infarction location by electrocardiogram (ECG). Other inclusive criteria included the ability to give informed consent and a stable medical status without important arrhythmias. The initial study was obtained at a mean of 8.5 days (range: 4-14 days) after acute infarction, and the follow-up MR study (last study obtained) was
RECENED 7/18/90; ACCEPTED9/17/90.
West 16B, The Mayo Clinic, 200 First St., Rochester, MN
Address all correspondence to Donald L. Johnston, MD,
55905, USA. 155
156
Magnetic Resonance Imaging l Volume 9, Number 2, 1991
at 12 wk (range: 6-27 wk). Fourteen patients had 2 studies, and 3 patients had 3 studies. performed
Imaging Technique ECG-gated, spin-echo imaging of the heart was performed as previously described. ’ In brief, a superconducting magnet (Technicare, Solon, Ohio) operating at a field strength of 0.6 Tesla was used for all studies. A multislice, short-axis series of images was acquired along the left ventricle from base to apex. Contiguous single-slice, spin-echo (TE = 60 msec) images were then obtained at end-systole or end-diastole (determined from the ECG) along the length of the ventricle. The ECG-gate delay time for acquisition was always the same for initial and follow-up studies. A single-slice, single-echo technique minimized the amount of blood pool signal present in the ventricular chamber and permitted better definition of the endocardium. Data Analysis Qualitative Assessment. All MR studies were analyzed by three reviewers and a visual estimate was made of changes in signal intensity from the initial study to the follow-up study. The images were specifically examined for changes in infarction signal intensity, differences in size, location and distribution of the infarction, and for the presence of myocardial thinning. A change in the size of the infarction region was determined by estimating the circumferential extent of the bright infarction signal on the initial study and comparing it to subsequent studies. These features were assessed for each tomographic slice. Regional left ventricular thinning was determined by comparing the infarction segment with the opposite myocardial wall thickness. Quantitative Analysis. The infarction signal intensity was measured by placing a circular region of interest cursor of defined size over the center of the infarction region. The signal intensity of the opposite myocardial segment was then measured in a similar fashion. Three determinations of each measurement were averaged to generate the infarction/normal myocardium signal intensity ratio. This ratio was obtained in a similar fashion on follow-up studies. Care was taken to perform the measurements on nearly identical tomograms. In 5 patients, signal intensity ratios were not determined due to difficulty in obtaining an accurate and reproducible region of interest on the follow-up studies. Problems were encountered in these instances due to marked wall thinning or regres-
sion of the infarction region to a thin portion of the endocardium. Statistical Analysis Differences in signal intensity ratios were determined by performing a Student t-test. P < 0.05 was considered significant. All values are mean + standard deviation. RESULTS Visual Assessment (Fig. I) Thirty-one studies were performed in 14 patients. In 13 patients, signal intensity remained increased in the same segment on the first follow-up study (Fig. 1). In one patient imaged at 26 wk, infarction signal intensity was normal and the left ventricular wall was thinned. Of the 3 patients with 3 studies each, 2 had increased signal intensity on the second follow-up study and one had normal signal at 27 wk. For the 13 patients with increased infarction signal at first followup, the region was smaller in size in 11 patients and unchanged in size in 2 patients. For the latter 2 patients, both had early follow-up studies (3 and 6 wk). The infarction was noted to have receded to the subendocardium in 9 patients as compared to the initial study. No change in distribution was noted in the remaining four patients. Myocardial thinning was present in 10 of 14 patients on the initial study and was apparent in 3 additional patients on follow-up studies. Wall thinning appeared to progress slightly on follow-up in six patients, three of whom had some initial thinning and three of whom showed no initial thinning. Myocardial Signal Intensity Ratio (Fig. 2) Figure 2 illustrates the signal intensity ratios measured for initial and follow-up MR studies in nine patients. The mean initial ratio of 1.81 + 0.42 decreased to 1.34 + 0.37 (p < 0.05) at a mean of 14 wk. Of interest, two patients had an increase in the signal intensity ratio on follow-up and both patients had been readmitted to the hospital at the time of the follow-up study with subsequent acute coronary events. In one patient, reinfarction occurred while the other patient presented with unstable angina and was found to have a left ventricular aneurysm. DISCUSSION This study is the first to demonstrate in man that evolutionary changes occur in MR signal intensity following acute myocardial infarction. The findings in-
Serial MRI following acute myocardial infarction 0 R.C.
THOMPSON ET AL.
157
09 Fig. 1. Spin-echo (TE = 60 msec) images of a patient 10 days (A) after acute myocardial infarction and at 16 wk (B). Signal intensity was increased in the anterior wall (arrow). At 16 wk, signal intensity was reduced compared to 10 days, and hrfarction size was smaller and more localized to the endocardium.
dicate that tissue healing proceeds from the epicardium to the endocardium over a period of time exceeding three mo. This pattern of healing was most likely due to a relatively poorly formed endocardial coronary collateral circulation that enhanced the severity of tis-
2.6 2.4 2.2 0 ‘i; t!i
2.0
.g
1.6
E
9) E .ia
_& c/l
1.6
1.4 1.2 1.0 0.6
0.6 0.4 0.2 0
I1
0
1
4
I
I,
6
I
12
I
16
I
I
20
I
I,
24
Time (weeks) Fig. 2. Infarction signal intensity ratio compared to number of weeks after myocardial infarction. The average signal intensity ratio for nine patients fell from 1.81 & 0.42 on the initial study to 1.34 + 0.37 at a mean of 14 wk.
sue injury and slowed formation of mature scar tissue compared to the epicardium. Our present findings are consistent with previously reported animal studies. Pflugfelder et al. demonstrated serial changes in infarction signal intensity in the dog following permanent ligation of the left anterior descending coronary artery.3 Signal intensity of the infarction increased from day 1 to days 4-6, then gradually decreased, but was still elevated in most dogs at 20 days. These changes were attributed to infarction healing. Wisenberg et al. measured signal intensity in dogs imaged after 1-3 hr of coronary artery occlusion, followed by 21 days of reperfusion.4 They found that Tl and T2 relaxation times (measured from the images) increased abruptly in the first hour, while T2returned to baseline by day 5. Johnston et al. reported an abrupt increase in T,relaxation time at three days in the infarcted rabbit heart.’ This change corresponded to the appearance of granulation tissue in the region of the infarction. T2returned to normal with the development of mature scar tissue at three mo. Since infarction healing proceeds more rapidly in the experimental animal than in man, it was not surprising that normalization of T2 in this study came before a return to normal in signal intensity in our study. In the present study, infarction signal intensity on the follow-up study was not found to be less than the
158
Magnetic Resonance Imaging 0 Volume 9, Number 2, 1991
surrounding normal myocardium. This was consistent with previous experimental data reported by Johnston et al., in which T, and T2relaxation times of mature scar were similar or only slightly lower than surrounding normal myocardium.5 However, McNamara and Higgins reported a decrease in infarction signal intensity and relaxation times in patients imaged 9 mo to 16 yr after acute myocardial infarction.6 It is possible that additional scar tissue maturation occurs after three mo that results in a further reduction in relaxation times. Further cardiac studies are indicated to assess the prognostic significance of evolutionary changes in MR signal intensity of myocardial infarction. For example, persistent bright signal 2-3 mo after acute myocardial infarction is consistent with residual necrosis or delayed healing, which might result in development of ventricular tachyarrhythmias and aneurysm formation. Alternatively, rapid normalization of infarction signal intensity is suggestive of early healing and might correlate with successful revascularization following acute myocardial infarction. In summary, our study demonstrates that serial changes occur in MR signal intensity over a period of at least three mo following acute myocardial infarction. A gradual reduction in the intensity of the infarction signal, localization of the bright signal to the endocardium, and progressive wall thinning in the area of the infarction are consistent with healing. MR imaging provides a new method of assessing infarc-
tion healing and may be useful for determining the effects of acute interventions following myocardial infarction. REFERENCES 1. Johnston, D.L.; Thompson, R.C.; Liu, P.; Dinsmore, R.E.; Wismer, G.L.; Saini, S.; Kaul, S.; Rosen, B.R.; Brady, T.J.; Okada, R.D. Magnetic resonance imaging during acute myocardial infarction. Am. J. Cardiol. 51: 1059-1065; 1986. 2. Dilworth, R.L.; Aisen, A.M.; Mancini, G.B.J.; Buda, A. J. Serial nuclear magnetic resonance imaging in acute myocardial infarction. Am. J. Cardiol. 59:1203-1205; 1987. 3. Pflugfelder, P.W.; Wisenberg, G.; Prato, ES.; Turner, K.L.; Carroll, S.E. Serial imaging of canine myocardial infarction by in vivo nuclear magnetic resonance. J. Am. CON. Cardiol. 7:843-849; 1986. 4. Wisenberg, G.; Prato, ES.; Carroll, S.E.; Turner, K.L.; Marshall, T. Serial nuclear magnetic resonance imaging of acute myocardial infraction with and without reperfusion. Am. Heart J. 115:510-518; 1988. 5. Johnston, D.L.; Homma, S.; Liu, P.; Weilbaecher, D.G.; Rokey, R.; Brady, T.J.; Okada, R.D. Serial changes in nuclear magnetic resonance relaxation times after myocardial infarction in the rabbit: Relationship to water content, severity of ischemia, and histopathology over a six-month period. Magn. Reson. Med. 8:363-379; 1988. 6. McNamara, M.T.; Higgins, C.B. Magnetic resonance imaging of chronic myocardial infarcts in man. AJR 146: 315-320; 1986.
