Skip main navigation

Reproducibility of Chronic Infarct Size Measurement by Contrast-Enhanced Magnetic Resonance Imaging

Originally publishedhttps://doi.org/10.1161/01.CIR.0000036368.63317.1CCirculation. 2002;106:2322–2327

Abstract

Background— The reproducibility of contrast-enhanced MRI has not been established. We compared MRI reproducibility for infarct size determination with that of 99mTc-sestamibi (MIBI) single photon emission computed tomography (SPECT).

Methods and Results— Patients with chronic myocardial infarction defined by enzymes (peak creatine kinase-MB 173±119 U/L) were scanned twice by MRI (MRI I and MRI II, n=20) and twice by SPECT (SPECT I and SPECT II, n=15) on the same day. The MRI contrast agent was injected during MRI I but not MRI II to test the effect of imaging time after contrast. Resting Tc-MIBI SPECT images were acquired and infarct size was determined with commercial software. Infarct size in patients scanned by MRI and SPECT was 14±6% of left ventricular mass (%LV) by MRI (range 4%LV to 27%LV) and 14±7%LV by SPECT (range 4%LV to 26%LV). MRI I and II scans were performed 10±2 and 27±3 minutes after contrast, respectively. For MRI, the difference in infarct size between scans I and II (bias) was −0.1%LV, and the coefficient of repeatability was ±2.4%LV. For SPECT, bias was −1.3%LV, and the coefficient of repeatability was ±4.0%LV. Within individual patients, no systematic differences in infarct size were detected when the 2 MRI scans were compared, the 2 SPECT scans were compared, or MRI was compared to SPECT.

Conclusion— The size of healed infarcts measured by contrast-enhanced MRI does not change between 10 and 30 minutes after contrast. The clinical reproducibility of contrast-enhanced MRI for infarct size determination compares favorably with that of routine clinical SPECT.

Contrast-enhanced MRI (ceMRI) is used clinically for the detection and sizing of myocardial infarction at an increasing number of institutions in the United States and Europe. Although there is substantial experimental evidence supporting the use of ceMRI to detect infarction,1–4 the reproducibility of ceMRI in a clinical setting has not been established, nor has the utility of this approach been compared with that of single photon emission computed tomography (SPECT) imaging. In particular, no previous studies have addressed the question of how the reproducibility of ceMRI compares with that of 99mTc-sestamibi (MIBI) SPECT in patients with healed myocardial infarction.

The goal of this study was to establish the clinical reproducibility of ceMRI for the measurement of infarct size. Our study design consisted of 2 consecutive MRI scans (MRI I and II) in patients with documented, healed myocardial infarction. Between MRI scans, the patient was removed from the scanner, and the second MRI scan was performed by a different scanner operator. The MRI contrast agent was injected for MRI I but not for MRI II to address the question of whether infarct size measurement in the setting of chronic myocardial infarction depends on imaging time after contrast injection.5 To allow the reproducibility of ceMRI to be directly compared with the reproducibility of an existing widely used clinical technique, patients were also scanned twice by resting Tc-MIBI SPECT on the same day as the 2 MRI scans.

Methods

Patient Population

This study was prospectively planned and was approved by the Institutional Review Board of Northwestern University. All patients gave informed consent. Twenty consecutive patients (62±7 years old) with myocardial infarcts at least 1 year old (3.9±2.6 years, range 1 to 11 years) who agreed to undergo repeated scans on the same day were enrolled. In all patients, the presence and age of infarction were defined by cardiac enzyme levels (peak creatine kinase [CK] 1934±1103 U/L, range 327 to 3765; peak CK-MB 173±119 μg/mL, range 39 to 366) at the time of the index event.

Imaging Protocols

Figure 1 summarizes the study design. Two MRI scans and 2 SPECT scans were acquired in 15 patients on the same day by different operators after only 1 injection of the MRI contrast agent and 1 injection of MIBI. Five additional patients underwent repeated MRI scans only.

Figure 1. Summary of study design. Two MRI scans and 2 SPECT scans were acquired in most patients on same day by different operators after only 1 injection of MRI contrast agent and 1 injection of 99mTc sestamibi SPECT. See text for details.

MRI

The methodology for acquiring MR images for the measurement of infarct size has been described in detail elsewhere.6,7 In brief, 10 minutes after intravenous injection of an MRI contrast agent (gadoteridol, Bracco Pharmaceuticals, 0.125 mmol/kg), ECG gated short-axis images were acquired during repeated breath holds every 10 mm from base to apex by an inversion-recovery turbo FLASH (fast, low-angle shot) pulse sequence.7 After acquisition of the first complete set of contrast images (MRI I), the patient was removed from the scanner and asked to stand up, then was put back in the scanner for the acquisition of the second set of images (MRI II). A different scanner operator, starting from new scout images, acquired images for MRI II. No additional contrast agent was administered between MRI I and MRI II to test the effects of imaging time after contrast.

SPECT

SPECT imaging was performed in all patients on the same day as the MRI procedure (Figure 1). Resting 99mTc sestamibi SPECT images were acquired in the clinical nuclear cardiology laboratory of Northwestern Memorial Hospital with a dual-detector gamma camera (ADAC Vertex), with 64 projections, each for 25 seconds, in a circular 180° orbit. The first set of SPECT images was acquired ≈2 hours after isotope injection (SPECT I). The patient was then asked to stand up and to lie down again for a second SPECT scan (SPECT II) performed ≈20 minutes after SPECT I.

Determination of Infarct Size

MRI

Infarct size by MRI was determined automatically by computer counting of all hyperenhanced pixels in the myocardium on each of the 6 to 8 short-axis images. Hyperenhanced pixels were defined as those with image intensities >2 SDs above the mean of image intensities in a remote myocardial region in the same image. Infarct size was determined as a percentage of left ventricular mass (%LV), as the sum of hyperenhanced pixels from each of the 6 to 8 short-axis images divided by the total number of pixels within the LV myocardium multiplied by 100%.

SPECT

Infarct size by SPECT was determined with an automated 3D software package written by investigators at Cedars-Sinai Hospital in Los Angeles, Calif.8–10 This program has been tested for reproducibility in 420 patients,10 and the underlying approach has been systematically compared with the results of expert visual scores.8 We determined SPECT infarct size as %LV by running a commercial version of this software (QPS Autoquant, ADAC Laboratories) on each of the 30 3D data sets (15 patients times 2 SPECT scans).

Statistical Analysis

Continuous data are expressed as mean±SD. The reproducibility of MRI and SPECT was analyzed with the repeatability analysis method of Bland-Altman.11 The agreement between MRI and SPECT (MRI I, MRI II and SPECT I, SPECT II) was analyzed by the agreement using repeated measurements method of Bland-Altman.11 The bias and 95% CIs were calculated as described by Bland-Altman.11

Results

Table 1 summarizes the size of infarction in all patients. In the 15 patients who underwent both MRI and SPECT, infarct size was 14±6%LV for MRI I and 14±6%LV for MRI II (14±6%LV overall). For SPECT, infarct size was 14±8%LV for SPECT I and 15±7%LV for SPECT II (14±7%LV overall). Infarct size for the 5 additional patients who only underwent repeated MRI was 25±7%LV for MRI I and 25±8%LV for MRI II.

TABLE 1. Infarct Size Measured by MRI and by SPECT

PatientPeak CK, U/LPeak CK-MB, U/LMRI I, %LVMRI II, %LVSPECT I, %LVSPECT II, %LV
All values are mean±SD.
137512993132
237653422324
312236911101719
4188112513142323
532758171589
634892742626
71582126151555
810596810101918
9213419327262126
109616418191115
115553991147
123447303129
1313359917191818
147235745910
15230835714152220
16513616546
1731012261413
18266136621202626
19267334427261115
202009284141377

Reproducibility of MRI

Figure 2 shows a full set of short-axis views of MRI I and MRI II acquired in patient 10. The presence, location, and size of the hyperenhanced region were similar in both MRI scans. The first MRI scan was 9 minutes after contrast, whereas the second scan was 32 minutes after contrast. Figure 3 shows similar results in 6 additional patients. Table 2 summarizes the timing of the 2 MRI scans and the inversion time (TI) for each scan selected by the scanner operators. On average, MRI I was performed 10±2 minutes after contrast, whereas MRI II was performed 27±3 minutes after contrast. For every patient, the inversion times selected by the scanner operators were longer for the second MRI scan (average increase 69±21 ms). The overall inversion times were 316±20 ms for MRI I and 385±20 ms for MRI II.

Figure 2. Full set of short-axis views of MRI I and MRI II acquired in patient 10. Presence, location, and size of hyperenhanced region were similar in both MRI scans. First MRI scan was 9 minutes after contrast, whereas second MR scan was 32 minutes after contrast.

Figure 3. Results of MRI I and MRI II in 6 additional patients. As in Figure 2, presence, location, and size of hyperenhanced regions in Figure 3 were similar in both MRI scans.

TABLE 2. Timing of MRI and SPECT

PatientMRI IMRI IITime Between MRI II and SPECT I, minSPECT I, Avg Time After MIBI, minSPECT II, Avg Time After MIBI, min
Avg Time After Gd, minTI, msAvg Time After Gd, minTI, ms
Avg indicates average; Gd, Gadoteridol.All values are mean±SD.
11032025380
2932025370
3123103043045117137
493002537056120140
5103002637092158178
6932026360
763002138048105125
883002838051117137
982902638071135155
1093003240053124144
1183203038078146166
121133028370
13113502639066133153
14103302540064129149
15103102435059123143
16113002938049119139
171234028390
18112903039051122142
19113603143089161181
20103302740044111131
All10±2316±2027±3385±2061±15128±16148±16

Figure 4A shows the results of Bland-Altman repeatability analysis of the MRI data of all 20 patients. The average difference in infarct size between scans I and II (bias) was −0.1%LV, and the coefficient of repeatability was ±2.4%LV. The 95% CIs for infarct size for comparison of MRI I to MRI II were 2.3%LV and −2.5%LV, ie, there were no systematic differences in infarct size between the 2 MRI scans.

Figure 4. Results of Bland-Altman repeatability analyses of MRI (A) and SPECT (B) data. COR indicates coefficient of repeatability. See text for details.

Reproducibility of SPECT

Figure 4B shows the results of Bland-Altman repeatability analysis of the SPECT data. The average difference in infarct size between scans I and II (bias) was −1.3%LV, and the coefficient of repeatability was ±4.0%LV. The 95% CIs for infarct size for comparison of MRI I to MRI II were 2.7%LV and −5.3%LV, ie, there were no systematic differences in infarct size between the 2 SPECT scans.

Comparison of MRI and SPECT

Bland-Altman analyses of the agreement using repeated measurements of the data (MRI I, MRI II, SPECT I, and SPECT II) revealed that the average difference in infarct size between MRI and SPECT (bias) in the 15 patients who underwent both studies was −0.5%LV (MRI infarct size smaller than SPECT), and the limit of agreement was ±19.2%LV. The 95% CIs for infarct size for comparison of MRI to SPECT were 18.7%LV and −19.7%LV, ie, there were no systematic differences in infarct size between MRI and SPECT.

Discussion

The main findings of this study were that the size of healed infarcts as determined by ceMRI does not change between 10 and 30 minutes after contrast injection and that the reproducibility of the MRI measurement compares favorably with that of MIBI SPECT.

Reproducibility of MRI

We found no differences in infarct size measurement by ceMRI between the MRI scans (eg, Figures 2 through 4). This finding contradicts the recent findings of Oshinski et al,5 who reported that infarct size measurement by MRI depends on the timing of imaging after contrast. Specifically, Oshinski et al5 reported that the spatial extent of hyperenhancement by ceMRI decreased from ≈60%LV at 3 minutes after con- trast to 30%LV by 40 minutes after contrast. This decrease in the spatial extent of hyperenhancement is ≈100-fold greater than the differences in sizes we observed (Figure 4, bias=0.3%LV) for images acquired between 10 and 27 minutes after contrast (Table 2). The discrepancy between the present study and the previous report5 could be due to differences in imaging times after contrast (3 to 40 minutes compared with 10 and 27 minutes), species (rat versus human), or the time elapsed after infarction (2 days versus >1 year). As briefly discussed in a recent correspondence,12 these differences could also relate to details of the MRI technique itself and the pharmacokinetics of the MRI contrast agent. These issues are described in more detail here.

The primary effect of the MRI contrast agent is to shorten myocardial longitudinal relaxation time, T1, and the underlying physiology results in a situation in which T1 is shortened more in infarcted regions than in normal myocardium. Although these regions of shortened T1 (infarcts) can be visualized with traditional T1-weighted MRI techniques, regional differences in image intensities are greatest when an inversion pulse is used.6,7 For correct implementation, however, the inversion time (delay between inversion pulse and data collection) must be manually selected to null signal from normal myocardial regions. The inversion time needed to null signal from normal myocardium varies from patient to patient because of differences in dose and varies with time after contrast administration because of contrast agent pharmokinetics.6

Weinmann et al13 studied the pharmacokinetics of Gd-DTPA in humans for doses of 0.1 and 0.25 mmol/kg. The solid line of Figure 5 shows a monoexponential fit to their data interpolated to a dose of 0.125 mmol/kg (the dose used in the present study). The plasma concentration of the MRI contrast agent decreased by a factor of ≈2.4 between 3 and 40 minutes after contrast. This decrease in contrast agent concentration will increase myocardial T1 and will require a corresponding increase in the MRI inversion time to appropriately null normal myocardium. Because interstitial concentrations of Gd-DTPA in the myocardium depend primarily on plasma concentrations, the correct MRI inversion time can be estimated from basic physical principles [eg, at 1 minute of 0.125 mmol/kg dose, Δ1/T1=(0.96 mmol/L)×(4.5/s/mmol/L)×(0.30{extracellular space})=1.29/s; T1=1/(1.29+1/0.8{precontrast T1}) =392 ms; inversion time to null normal myocardium=(392)(ln2)=272 ms]. The dashed line in Figure 5 depicts the correct inversion times that are needed for the MRI technique to account for the pharmacokinetics of the MRI contrast agent. The filled circles of Figure 5 show the present data taken from Table 2. The changes in inversion times selected by the scanner operators of the present study (filled circles) were similar to those expected based on the pharmacokinetics of the contrast agent (dotted line).

Figure 5. Solid line: monoexponential fit to serum contrast agent concentration (left-hand y-axis) as function of time after administration calculated (0.125 mmol/kg) based on data of Weinmann et al.13 Dotted line: MRI inversion time (right-hand y-axis) calculated based on data of solid line. Filled circles indicate present data taken from Table 2. See text for details.

MRI Compared With SPECT

The reproducibility of an imaging test has practical implications for the design of clinical trials that use the results of imaging techniques as an end point. Gibbons et al14 described the influence of measurement reproducibility using SPECT on the number of patients necessary to detect the effects of a candidate intervention as a function of the t statistic. These authors showed that the number of patients needed for a clinical trial is proportional to the square of the SD of the imaging end point.14 In the present study, the SD of infarct size is one half the coefficient of repeatability, and the ratio of the numbers of patients needed for clinical trials based on MRI compared with SPECT is 0.42 [(2.6/4.0)2=0.42]. Accordingly, if no other factors play a role, MRI reduces the number of patients needed for a clinical trial to 42% of that needed if infarct size is determined by SPECT. However, this estimate is based on a limited number of patients and does not address the important issue of accuracy, because the true infarct size is not known.

Conclusion

In summary, we found that the size of healed infarcts measured by ceMRI does not change between 10 and 30 minutes after contrast and that the reproducibility of ceMRI compares favorably with that of 99mTc sestamibi SPECT.

This work was supported by a grant from the Robert Bosch Foundation (Dr Mahrholdt), NIH-NHLBI R01-HL63268 and K02-HL04394 (Dr Judd), and R01-HL64726 (Dr Kim). The authors thank Dr M. Honold, Dr C. Kupfahl, and Professor U. Sechtem of the Robert Bosch Medical Center, Stuttgart, Germany, for their contribution to this work.

Footnotes

Correspondence to Robert M. Judd, PhD, Duke Cardiovascular Magnetic Resonance Center, Duke University Health System, PO Box 3934, Durham, NC 27710. E-mail

References

  • 1 Klein C, Nekolla SG, Bengel FM, et al. Assessment of myocardial viability with contrast-enhanced magnetic resonance imaging: comparison with positron emission tomography. Circulation. 2002; 105: 162–167.CrossrefMedlineGoogle Scholar
  • 2 Fieno DS, Kim RJ, Chen EL, et al. Contrast-enhanced magnetic resonance imaging of myocardium at risk: distinction between reversible and irreversible injury throughout infarct healing. J Am Coll Cardiol. 2000; 36: 1985–1991.CrossrefMedlineGoogle Scholar
  • 3 Wu E, Judd RM, Vargas JD, et al. Visualisation of presence, location, and transmural extent of healed Q-wave and non-Q-wave myocardial infarction. Lancet. 2001; 357: 21–28.CrossrefMedlineGoogle Scholar
  • 4 Kim RJ, Wu E, Rafael A, et al. The use of contrast-enhanced magnetic resonance imaging to identify reversible myocardial dysfunction. N Engl J Med. 2000; 343: 1445–1453.CrossrefMedlineGoogle Scholar
  • 5 Oshinski JN, Yang Z, Jones JR, et al. Imaging time after Gd-DTPA injection is critical in using delayed enhancement to determine infarct size accurately with magnetic resonance imaging. Circulation. 2001; 104: 2838–2842.CrossrefMedlineGoogle Scholar
  • 6 Mahrholdt H, Wagner A, Judd RM, et al. Assessment of myocardial viability by cardiovascular magnetic resonance imaging. Eur Heart J. 2002; 23: 602–619.CrossrefMedlineGoogle Scholar
  • 7 Simonetti OP, Kim RJ, Fieno DS, et al. An improved MR imaging technique for the visualization of myocardial infarction. Radiology. 2001; 218: 215–223.CrossrefMedlineGoogle Scholar
  • 8 Kang X, Berman DS, Van Train KF, et al. Clinical validation of automatic quantitative defect size in rest technetium-99 m-sestamibi myocardial perfusion SPECT. J Nucl Med. 1997; 38: 1441–1446.MedlineGoogle Scholar
  • 9 Germano G, Kavanagh PB, Berman DS. An automatic approach to the analysis, quantitation and review of perfusion and function from myocardial perfusion SPECT images. Int J Card Imaging. 1997; 13: 337–346.CrossrefMedlineGoogle Scholar
  • 10 Germano G, Kavanagh PB, Waechter P, et al. A new algorithm for the quantitation of myocardial perfusion SPECT, I: technical principles and reproducibility. J Nucl Med. 2000; 41: 712–719.MedlineGoogle Scholar
  • 11 Bland JM, Altman DG. Comparing methods of measurement: why plotting difference against standard method is misleading. Lancet. 1995 21; 346: 1085–1087.Google Scholar
  • 12 Judd RM, Kim RJ. Imaging time after Gd-DTPA injection is critical in using delayed enhancement to determine infarct size accurately with magnetic resonance imaging. Circulation. 2002; 106: e6.Letter.LinkGoogle Scholar
  • 13 Weinmann HJ, Laniado M, Mützel W, et al. Pharmacokinetics of GdDTPA/dimeglumine after intravenous injection into healthy volunteers. Physiol Chem Phys Med NMR. 1984; 16: 167–172.MedlineGoogle Scholar
  • 14 Gibbons RJ, Christian TF, Hopfenspirger M, et al. Myocardium at risk and infarct size after thrombolytic therapy for acute myocardial infarction: implications for the design of randomized trials of acute intervention. J Am Coll Cardiol. 1994; 24: 616–623.CrossrefMedlineGoogle Scholar

eLetters(0)

eLetters should relate to an article recently published in the journal and are not a forum for providing unpublished data. Comments are reviewed for appropriate use of tone and language. Comments are not peer-reviewed. Acceptable comments are posted to the journal website only. Comments are not published in an issue and are not indexed in PubMed. Comments should be no longer than 500 words and will only be posted online. References are limited to 10. Authors of the article cited in the comment will be invited to reply, as appropriate.

Comments and feedback on AHA/ASA Scientific Statements and Guidelines should be directed to the AHA/ASA Manuscript Oversight Committee via its Correspondence page.