OUP user menu

Assessment of reversible myocardial dysfunction in chronic ischaemic heart disease: comparison of contrast-enhanced cardiovascular magnetic resonance and a combined positron emission tomography–single photon emission computed tomography imaging protocol

Harald P. Kühl, Claudia S.A. Lipke, Gabriele A. Krombach, Marcus Katoh, Thomas F. Battenberg, Bernd Nowak, Nicole Heussen, Arno Buecker, Wolfgang M. Schaefer
DOI: http://dx.doi.org/10.1093/eurheartj/ehi747 846-853 First published online: 24 January 2006


Aims The aim of the study was to compare, in patients with chronic ischaemic cardiomyopathy, contrast-enhanced cardiovascular magnetic resonance (ce-CMR) imaging and a combined 18F-fluorodeoxyglucose (18F-FDG) positron emission tomography (PET) and 99mTc-sestamibi single-photon emission computed tomography (SPECT) protocols for the prediction of functional recovery after revascularization, as assessed by cine CMR.

Methods and results Twenty-nine patients with ischaemic cardiomyopathy (ejection fraction 32±10%) were investigated with ce-CMR and PET/SPECT. For the assessment of global and regional functions, cine CMR was performed at baseline and at 6 months follow-up. For ce-CMR, the segmental extent of hyperenhancement (SEH) was quantitated, and for PET/SPECT, different viability categories were defined according to a validated quantitative protocol. Functional improvement was related to the SEH by ce-CMR, as well as to the viability categories by PET/SPECT. Sensitivity and specificity for the prediction of functional recovery at follow-up was 97 and 68% for ce-CMR and 87 and 76% for PET/SPECT. The positive predictive value was identical for both techniques (73%). However, ce-CMR achieved a higher negative predictive value (93 vs. 77%, respectively), indicating that ce-CMR may be superior to PET/SPECT for the identification of segments unlikely to recover function after revascularization. Both methods had a similar yield in the prediction of global functional improvement.

Conclusion ce-CMR is comparable with a PET/SPECT imaging protocol for the prediction of regional and global functional improvement after revascularization. However, ce-CMR may be superior to nuclear imaging for the identification of segments that are unlikely to recover function at follow-up.

  • Magnetic resonance
  • Positron emission tomography
  • Single-photon emission computed tomography
  • Ischaemic cardiomyopathy
  • Hibernation
  • Left ventricular function


The assessment of myocardial viability provides important prognostic information in patients with chronic ischaemic heart disease.1 Metabolic imaging using 18F-fluorodeoxyglucose (18F-FDG) positron emission tomography (PET) in combination with perfusion imaging, using either PET or single-photon emission computed tomography (SPECT) technology, has long been accepted as the reference method for the assessment of myocardial viability.2 This technique allows discrimination of different myocardial viability states based on information of regional metabolism and flow, providing important information on the probability for functional recovery after revascularization. However, nuclear imaging is associated with radiation exposure. In addition, the imaging protocol is relatively time-consuming, especially if a glucose insulin clamp is performed to drive glucose uptake into the myocardial cells.

Contrast-enhanced cardiovascular magnetic resonance (ce-CMR) imaging has been suggested as a new modality for the assessment of myocardial viability. This technique allows direct visualization of scar tissue with high contrast and high spatial resolution.3 In patients with chronic ischaemic heart disease, it has been demonstrated that the transmural extent of scar predicts functional recovery after myocardial revascularization.4,5 The imaging protocol is rather simple and straightforward, allowing for rapid scanning in clinical practice. Thus, ce-CMR has been proposed as an alternative technique to 18F-FDG-PET imaging for the prediction of reversible dysfunction in chronic ischaemic heart disease. However, no direct head-to-head comparison between the methods has been performed so far. Therefore, this study was conducted to prospectively compare, using predefined and validated viability protocols, ce-CMR and a combined nuclear imaging protocol using 18F-FDG-PET for metabolic imaging and 99mTc-sestamibi SPECT for perfusion imaging for the prediction of reversible myocardial dysfunction in patients with chronic ischaemic heart disease.


Forty-six consecutive patients with chronic ischaemic heart disease, regional wall motion abnormalities, and an ejection fraction (EF) <50% with clinical indication for myocardial viability were initially assessed for inclusion into the study. Five patients were excluded for clinical reasons (severe concomitant disease such as cerebrovascular disease, chronic pulmonary disease, chronic kidney disease, or peripheral vascular disease impeding revascularization). In addition, four patients were excluded because of previous pacemaker or defibrillator implantation and one patient refused to undergo CMR because of claustrophobia. Thus, 36 patients were included into the study. From the 36 patients, 29 subjects completed the follow-up CMR examination. There were three deaths: two patients died from cardiac cause (cardiogenic shock after acute stent thrombosis of the LAD in one patient and low-output failure 10 days after cardiac bypass surgery in the other patient) and the third patient died form severe sepsis after non-cardiac surgery. Moreover, three patients refused to complete the follow-up CMR examination and one patient was lost to follow-up. Thus, 29 patients completed the study protocol. None of the patients had a recent myocardial infarction.

Characteristics of the patient population are demonstrated in Table 1. Cardiovascular risk factors including nicotine abuse, hypertension (use of antihypertensive medication or blood pressure at rest >140/90 mmHg), diabetes mellitus (use of insulin or oral anti-diabetic agents or fasting serum glucose >130 mg/dL), and hypercholesterolaemia (total fasting serum cholesterol >200 mg/dL or use of cholesterol-lowering medication) were assessed from the patient chart. All patients received a standard pharmacological therapy including β-blockers, ACE-inhibitors, and statins. Additionally, 84% of patients received diuretics for symptoms of congestive heart failure. The patients were scheduled for a PET/SPECT viability study for clinical reasons. In addition, a CMR examination was performed within a median of 2 days (inter-quartile range 1–3.5). There were no clinical events or change in medication between the different examinations. Patients were invited to return for a follow-up CMR examination 6 months after revascularization. The study was approved by the local Ethics Committee and all patients gave written informed consent.

View this table:
Table 1

Patient characteristics

Age (years)66±9
Gender (m/f)21/8
Previous infarction (%)24 (83)
Diabetes (%)10 (34)
Hypertension (%)22 (76)
Hypercholesterolaemia (%)19 (66)
Nicotine abuse (%)18 (62)
Elevated serum creatinine (%)7 (24)
End-diastolic volume (mL)200±50
End-systolic volume (mL)137±47
EF (%)32±10
Left ventricular mass (g)116±16

Magnetic resonance imaging

Images were acquired on a 1.5 T clinical scanner (Intera, Philips, Best, The Netherlands) using a five-element-phased array cardiac synergy coil. To speed up image acquisition, the parallel imaging technique SENSE was applied.6 Scout images were acquired in long-axis and short-axis orientations for the planning of the final double-oblique long-axis and short-axis views. ECG-gated cine images were acquired using a segmented steady-state free precession sequence (TE/TR 1.8/3.6 ms, resolution 1.8×1.5×8 mm3, 25 frames per RR interval). Three long-axis views and 10–12 short-axis views (slice thickness 8 mm, 2 mm gap) covering the whole left ventricle were obtained during repeated breath-holds. A gadolinium-based contrast agent (Magnevist, Schering; 0.2 mmol/kg) was then administered intravenously. After 15 min, contrast-enhanced images were acquired in the same orientation as the cine images using a segmented inversion-recovery gradient-echo pulse sequence triggered to end-diastole.7 The inversion time was set to null the signal of normal myocardium after contrast administration (typically 250–300 ms) and was adjusted in the course of the investigation if necessary. Other parameters of the sequence were TR/TE two heart beats/5.0 ms, FA 25°, matrix 256×256, and a typical voxel size of 1.4×1.4×8.0 mm3.

PET/SPECT imaging

Scanning was performed according to a validated protocol8 designed for routine clinical scanning including metabolic imaging with 18F-FDG PET and perfusion imaging with 99 mTc-sestamibi SPECT. This protocol is also practicable in centres with a PET camera in which an on-site cyclotron is not available.

Patient preparation

All 18F-FDG PET and perfusion 99mTc-sestamibi SPECT studies were performed on the same day with the patient supine. At presentation, all patients had fasted and had been taking their regular cardiac medication. Between injection of 99mTc-sestamibi and SPECT acquisition, a meal was served to reduce biliary activity. To reduce myocardial fatty acid metabolism, 250 mg of acipimox was given to all patients approximately 2 h before administration of 18F-FDG.9 About 1 h before injection of 18F-FDG, non-diabetics received an oral glucose load of 50 g. Insulin was administered intravenously to diabetics 5–10 min before administration of 18F-FDG.

18F-fluorodeoxyglucose positron emission tomography

Static 18F-FDG PET scanning (ECAT EXACT 922/47; CTI, Knoxville, TN, USA/Siemens Medical Systems, Inc., Hoffman Estates, IL, USA) was performed 60 min after intravenous administration of about 300 MBq of 18F-FDG. The acquisition time was 30 min for emission (two-dimensional mode) and 15 min for transmission (68Ge/68Ga rod sources). Attenuation-corrected data were reconstructed using an iterative algorithm (OSEM, 6 iterations, 16 subsets). The matrix size was 128×128 pixels and the reconstruction zoom was 2.154. The resulting pixel size was 2.39×2.39×3.38 mm3.

99mTc-sestamibi SPECT

Myocardial perfusion SPECT was performed 60 min after injection of about 400 MBq of 99mTc-sestamibi, using a dual-head gamma camera (Solus; ADAC Laboratories, Milpitas, CA, USA) equipped with a low-energy all-purpose collimator. A validated method10 using a triple-energy-window acquisition was used to correct the data sets for attenuation and scatter. Data were reconstructed using a Butterworth filter (cutoff, 0.7 Nyquist; order, 5; 128×128 matrix).


After reorientation according to the left ventricular horizontal and vertical long axes, myocardial 99mTc-sestamibi uptake and 18F-FDG uptake were assessed by using a volumetric vector sampling method on the short-axes slices yielding polar maps of 460 data points. The reference region for both 99mTc-sestamibi uptake and 18F-FDG uptake was the region with the maximum 99mTc-sestamibi uptake. Segmental 99mTc-sestamibi and 18F-FDG uptake values were expressed as percentages of the respective reference region in each patient.

Data analysis

Segmentation model

For all imaging techniques, the 17-segment model as proposed by the American Heart Association was used.11 The procedures for obtaining this segmentation in the CMR and PET/SPECT data sets have been described previously.12 The anterior insertion of the right ventricular free wall into the left ventricular myocardium was used as landmark for the composition of the segmental model.

Magnetic resonance

Using cine CMR, regional wall motion was assessed visually by consensus interpretation of two experienced observers before revascularization and at 6 months follow-up. Each myocardial segment was scored using a five-point scale: 1=normal contractility, 2=mild to moderate hypokinesia, 3=severe hypokinesia, 4=akinesia, and 5=dyskinesia.

Quantitative evaluation of the contrast-enhanced images was performed using the MASS software (Medis, Leiden, The Netherlands), as previously described.12 This analysis was performed separately and independently from the wall motion analysis. Each myocardial segment was evaluated for the presence of hyperenhancement, defined as an area of signal enhancement ≥3 SD of the signal intensity of non-enhanced myocardium. The total myocardial area and the contrast-enhanced area of each segment were traced manually. The segmental extent of hyperenhancement (SEH) was calculated, defined as the percentage contrast-enhanced area of the total myocardial area (Ahyperenhancement/Amyocardium×100, where A denotes area).

Nuclear imaging

Data were analysed blinded to CMR results and patient data. The same 17-segment model was used as with CMR, with the anterior insertion of the right ventricular free wall into the left ventricle serving as landmark. The information on wall motion abnormalities was not available to the investigator.

Definitions for myocardial viability

The definitions for myocardial viability using ce-CMR and PET/SPECT are summarized in Table 2. For ce-CMR, a predefined threshold of 50% SEH was used to discriminate myocardial viability.5,12 For PET/SPECT, four different categories of segments were defined8: viability was assumed to be present in segments with normal perfusion by 99mTC-sestamibi SPECT as well as in segments demonstrating a mismatch pattern (reduced 99mTC-sestamibi uptake, preserved or increased 18F-FDG-uptake). Segments with a mild match pattern (mildly reduced 99mTC-sestamibi uptake and 18F-FDG uptake consistent with a non-transmural scar) or a match pattern (severely reduced 99mTC-sestamibi uptake and 18F-FDG uptake consistent with a transmural scar) were considered non-viable. For recovery of function only segments with severe dysfunction at baseline [wall motion score (WMS) ≥3] were considered in the final analysis. Improvement of segmental myocardial function was assumed to be present when the difference in WMS between baseline and follow-up examination was ≥1. Functional impairment was defined as difference in WMS ≤−1.

View this table:
Table 2

Definitions for myocardial viability using ce-CMR and PET/SPECT

ModalityViability statusSEH99mTc-sestamibi uptake18F-FDG uptake
PET/SPECTNormal perfusion (viable)>70%Not considered
Mismatch (viable)≤70%>70% and (difference FDG−sestamibi) ≥20%
Mild match (non-viable)≤70%≤70% or >70% and (difference FDG−sestamibi) ≤20%
Match (non-viable)≤50%≤50%

SEH denotes segmental extent of hyperenhancement.


Data are expressed as mean±standard deviation. For the prediction of regional functional improvement, analysis was performed on a per segment basis. A generalized estimating equation (GEE) approach with a Bernoulli variance function and a working correlation matrix with exchangeable correlation assumption were used to estimate sensitivity and specificity.13 This GEE model is identical to a logistic regression model Math (p, probability of a positive result with ce-CMR or PET/SPECT; X1, viable/non-viable according to standard of reference; β0, regression intercept; β1, coefficient for viable/non-viable segment) in which correlation between segments within a patient is taken into account. Therefore, sensitivity may be calculated as Math and specificity as Math Repeated measures analysis of variance was applied to compare mean WMS-baseline values between viable and non-viable segments, considering the assumption that segments within a patient are correlated.

For the prediction of global functional recovery, analysis was performed on a per patient basis. Spearman correlation coefficients were calculated to assess the association between the number of viable segments per patient, with each imaging technique at baseline and change of EF at follow-up. No formal sample size calculation was performed. Eligible patients were included consecutively in the study. All tests were two-sided; because of an exploratory study approach, no alpha-adjustment was performed. P-values less than an alpha level of 0.05 were considered to indicate local statistical significance. For the comparisons of parameters between groups of patients, the Wilcoxon test was applied. In addition, the software SAS (SAS Institute Inc., Cary, NC, USA) was used.


Of the 29 patients enrolled in the study, 15 subjects underwent percutaneous coronary intervention with stent implantation and 14 patients underwent coronary artery bypass grafting. The average number of revascularized vessels was 1.2±0.4 in the stent group and 3.2±0.7 in the bypass group. EF at baseline was not different between the groups of patients undergoing stent implantation or bypass surgery (32±9% vs. 33±10%; P=0.75).

A total of 493 myocardial segments were included in the analysis. At baseline, 187 segments demonstrated a severely reduced function (WMS≥3). After revascularization 96 segments improved function, 87 remained unchanged, and four showed impaired function. Segments with improvement of function after revascularization revealed less scar (P<0.0001) and a higher 18F-FDG uptake (P<0.0001) as well as 99mTc-sestamibi uptake (P=0.0033) when compared with segments without recovery of function (Table 3).

View this table:
Table 3

WMS, SEH, and tracer uptake by nuclear imaging in relation to the functional segmental status after revascularization

Segments with severe dysfunctionWMS at baselineWMS at follow-upSEH (%)FDG-uptake (%)Sestamibi uptake (%)
With functional recovery (n=96)3.5±0.51.7±0.713±1683±1862±19
Without functional recovery (n=91)3.8±0.43.8±0.563±3062±2152±18

WMS, wall motion score; SEH, segmental extent of hyperenhancement.

Improvement of segmental myocardial function was associated with the SEH by ce-CMR. A gradual decrease in the probability of functional recovery with increasing SEH was observed (Figure 1). Segments demonstrating ≤50% SEH improved function in 77%. For segments with >50% SEH, there was a low likelihood for functional recovery (7%, P<0.0001 vs. viable segments). None of the segments with >75% SEH improved function at follow-up.

Figure 1 Prevalence of improvement of segmental myocardial function for increasing quartiles of SEH by ce-CMR (black bars). The percentage of viable (stippled bars) and non-viable (hatched bars) segments by PET/SPECT depending on SEH by ce-CMR is given.

For PET/SPECT, segments considered viable (normal and mismatch patterns) demonstrated a high probability for functional improvement (78%), whereas segments considered non-viable (mild match and match patterns) revealed a low likelihood for functional recovery (16%, P<0.0001 vs. viable segments; Figure 2).

Figure 2 Prevalence of improvement of segmental myocardial function depending on the different viability categories by PET/SPECT (black bars). The SEH by ce-CMR (open bars) for each viability category by PET/SPECT is indicated.

Table 4 depicts the segmental viability status by ce-CMR and PET/SPECT, depending on the functional outcome after revascularization. Table 5 summarizes the sensitivity, specificity, and positive and negative predictive value of each imaging technique for the prediction of regional functional recovery. The number of false-positive segments was comparable between the two imaging techniques (27 and 24 segments for ce-CMR and PET/SPECT, respectively). However, ce-CMR demonstrated a lower number of false-negative segments when compared with PET/SPEC (two vs. 13 segments, respectively), which resulted in a higher negative predictive value for ce-CMR, demonstrating that a negative test result by ce-CMR indicated a higher probability for a lack of segmental functional improvement when compared with a negative result by PET/SPECT.

View this table:
Table 4

Segmental viability status by ce-CMR and PET/SPECT depending on the functional outcome after revascularization

Severely dysfunctional segments (n=187)ce-CMRPET/SPECT
With functional improvement (n=96)94 28313
Without functional improvement (n=91)27642467
View this table:
Table 5

Sensitivity, specificity, and predictive values of ce-CMR and PET/SPECT for the prediction of recovery of regional myocardial function

Sensitivity (%)97 (91;99)87 (78;93)
Specificity (%)68 (11;97)76 (37;94)
Positive predictive value (%)73 (59;83)73 (58;84)
Negative predictive value (%)93 (72;99)77 (40;94)

Numbers in brackets give 95% confidence intervals.

Recovery of function was highly likely when both modalities demonstrated preserved viability (87%, 83/95 segments; Figure 3) and was negligible when both techniques indicated non-viability (4%, 2/54 segments; Figure 4). Segments showing preserved viability by ce-CMR and non-viability by PET/SPECT demonstrated an intermediate likelihood for functional improvement (42%, 11/26 segments; Figure 5). Conversely, none of the 12 segments assessed as non-viable by ce-CMR and viable by PET/SPECT improved function at follow-up. Segments considered viable by PET/SPECT, which did not improve function at follow-up, demonstrated a larger amount of SEH by ce-CMR when compared with viable segments by PET/SPECT with functional improvement (52±35% vs. 12±15%; P<0.001).

Figure 3 Example of a patient with concordant positive results between PET/SPECT and ce-CMR. Cine CMR (see Supplementary material online, Video S1) shows a severe wall motion abnormality of the anterior wall [arrows in the right upper panel (Cine pre), stop frame image at end-systole]. Global EF was quantified as 25%. 99mTc-sestamibi-SPECT imaging (left upper image) revealed a severe perfusion deficit of the anterior wall. On 18F-FDG-PET imaging (left lower image), a preserved glucose metabolism is recognized in the same area indicating a mismatch pattern. The corresponding ce-CMR image (middle panel) demonstrates no hyperenhancement in the anterior wall. Cine CMR (see Supplementary material online, Video S2), 6 months after revascularization of the LAD [right lower panel (Cine post), stop frame image at end-systole], reveals an almost normalized motion of the anterior wall (arrows) and global functional improvement (EF at follow-up, 45%).

Figure 4 Example of a patient with concordant negative results between PET/SPECT and ce-CMR. Cine-CMR (see Supplementary material online, Video S3) reveals akinesia of the inferior wall [arrows in the right upper panel (Cine pre), stop frame image at end-systole]. Global EF was assessed as 41%. 99mTc-sestamibi-SPECT (left upper panel) and 18F-FDG-PET imaging reveal a concordant deficit in the inferior wall, indicating non-viability. At ce-CMR (middle panel), a nearly transmural enhancement pattern is recognized. At follow-up (see Supplementary material online, Video S4) after recanalization of the right coronary artery, no functional improvement is observed in the inferior wall [(Cine post), stop frame image at end-systole]. Global EF remained unchanged (40%). Notice that the subendocardial scar in the anterolateral wall (white arrow in the middle panel) is not perceived in the PET and SPECT images.

Figure 5 Example of a patient with discrepant finding between nuclear imaging and ce-CMR. Cine CMR (see Supplementary material online, Video S5) shows akinesia to dyskinesia of two-thirds of the anterior wall, apex, and apical inferior wall [arrows in the right upper panel (Cine pre), stop frame image at end-systole]. Additionally, marked thinning of the anterior wall is present. Global EF was quantified as 21%. 99mTc-sestamibi-SPECT (left upper image) and 18F-FDG-PET imaging (left lower image) reveal a severe defect of the anterior wall and apex classified as match (transmural scar). The corresponding ce-CMR image (middle panel) demonstrates a small subendocardial rim of contrast enhancement, whereas large proportions of the anterior wall remain dark and thus viable. Cine CMR (see Supplementary material online, Video S6), 6 months after revascularization of a high grade stenosis of the LAD (right lower panel (Cine post), stop frame image at end-systole), reveals a marked improvement in regional (arrows) and global functions (EF at follow-up, 39%), along with regain in wall thickness.

Seven of the nine patients who underwent repeat angiography 6 months after stent implantation demonstrated patent vessels at follow-up. In these patients, a total of 29 severely dysfunctional segments at baseline were identified, which were successfully revascularized. With ce-CMR, 13/16 viable segments improved function, whereas none of the 13 non-viable segments recovered at follow-up. With PET/SPECT 9/10, viable segments demonstrated functional improvement, whereas 4/19 segments considered non-viable improved function after revascularization.

Global systolic function improved by 5.1±8% after revascularization (P<0.001). There was no difference in the extent of global functional improvement between the revascularization strategies (4.6±6.4% for bypass surgery and 5.6±9.4 for stent implantation; P=0.78). A similar correlation between the number of viable segments per patient detected by PET/SPECT (r=0.61, P<0.001) as well as by ce-CMR (r=0.62, P<0.0001) and the change of EF at follow-up was observed (Figure 6).

Figure 6 Relation of the number of severely dysfunctional, viable segments by ce-CMR (solid dots and solid regression line) and PET/SPECT (open triangles and broken regression line) and change of EF at follow-up.


The results of this study demonstrate that in patients with chronic ischaemic heart disease, ce-CMR and a combined 18F-FDG-PET/99mTc-sestamibi-SPECT hybrid protocol are comparable for the prediction of regional and global improvement of LV function after revascularization. However, for segments classified as non-viable, ce-CMR was superior to PET/SPECT in predicting lack of functional recovery. This finding may be clinically important, indicating that ce-CMR may be especially useful in identifying patients who may not need a coronary revascularization.

The findings of this study are in close agreement with the study of Kim et al.,4 who were the first to describe the relationship between the transmural extent of scar and the recovery of segmental function at follow-up. The predefined threshold of 50% for SEH proved to be a good discriminator for functional outcome after revascularization. Overall, 77% of severely dysfunctional segments with ≤50% SEH improved function after revascularization with a decreasing probability for increasing SEH. In contrast, segments with >50% SEH demonstrated a low likelihood for functional improvement at follow-up, which is in line with previous reports.4,5 These results confirm and extent the results by Schvartzman et al.,5 who suggested that the 50% cutoff for scar transmurality may be useful clinically to guide therapeutic decisions on myocardial revascularization.

For PET/SPECT, the findings of the present study are in good agreement with the previously published data.8,14 As expected, segments considered viable yielded a high probability for functional improvement (78%), whereas non-viable segments by PET/SPECT revealed a low likelihood (16%) for functional recovery. Considering the results of both imaging techniques, functional improvement was highly likely when both modalities indicated preserved viability (87%) and negligible when there was no evidence of viability by ce-CMR and PET/SPECT (4%). In the presence of discrepant findings between the modalities, ce-CMR was superior to PET/SPECT for predicting lack of recovery of segmental myocardial function after revascularization. None of the 12 segments classified as non-viable by ce-CMR and viable by PET/SPECT revealed functional recovery at follow-up. Conversely, 11 of 26 segments (42%) assessed as non-viable by PET/SPECT but viable by ce-CMR showed functional improvement. Viable segments by PET/SPECT, which did not improve function at follow-up, demonstrated an increased amount of hyperenhancement (52±35%) believed to be prohibitive for functional recovery after revascularization. This finding may be explained by the insensitivity of 18F-FDG PET imaging in detecting subendocardial scar owing to its limited spatial resolution (Figure 4). In the study of Klein et al.,15 more than half of the segments with subendocardial scar at ce-CMR were not detected by 18F-FDG PET. Another explanation is that a relatively small volume of dysfunctional viable tissue may show increased 18F-FDG uptake, with PET indicating viability, whereas the coexisting amount of scar impedes functional recovery. In accordance with these results, Knuesel et al.16 demonstrated that viable segments by PET revealing a thin viable myocardial rim at ce-CMR (and thus a large amount of transmural scar) showed a relatively low probability for recovery of function at follow-up.

We observed segments with a low 18F-FDG uptake and 99mTc-sestamibi uptake, classified as ‘match’ (transmural scar) by nuclear imaging, which nevertheless improved function at follow-up. These segments demonstrated a thin wall with a low extent of scar at ce-CMR. The reason for this finding may be explained by differences in the way the two techniques assess myocardial viability17: the PET technique determines the viability state of a given segment in relation to a reference segment expressing FDG uptake relative to the FDG uptake of the segment with maximal sestamibi uptake. Thus, a thick reference segment (12 mm) will influence a thin segment (5 mm) with a small rim of subendocardial scar (1 mm), which will show less FDG uptake, potentially decreasing the relative percentage FDG uptake to below the threshold-value considered for viability despite the presence of a viable rim (4 mm). In contrast, ce-CMR allows determining directly the amount of viable myocardium (4 mm) within the thin segment (5 mm). Figure 5 demonstrates such an example of a patient with a thin anterior wall scored non-viable by PET/SPECT. At ce-CMR, a small subendocardial scar and a substantial rim of non-infarcted tissue were detected. After revascularization, reverse remodelling with regain in wall thickness and recovery of regional as well as global function was observed. A similar finding has been reported recently.18 This finding is clinically important because revascularization may be withheld in patients with a negative PET/SPECT result and a thin myocardial wall despite the presence of substantial amount of viability. Thus, ce-CMR may be especially important in identifying myocardial viability in patients with ischaemic cardiomyopathy and a thin myocardial wall.

ce-CMR may have additional benefits when compared with PET/SPECT for viability testing. The time required to perform a viability study varies considerably between the two modalities. With CMR, a complete functional and viability scan can be accomplished in less than 1 h, whereas approximately 3 h is required for the combined PET and SPECT study. This difference may have important implications for patient acceptance and throughput, as well as for economic concerns.

There are limitations to this study. Recovery of regional myocardial function may be an imperfect surrogate for outcome after revascularization. It has been shown that patients with ischaemic cardiomyopathy may benefit from restoration of blood flow to jeopardized myocardium independently of functional recovery, which has been attributed to the prevention of adverse remodelling and electrical stabilization of the ischaemic myocardium.19 The prognostic significance of ce-CMR in patients with ischaemic cardiomyopathy needs to be addressed in future studies.

As in all comparative analyses, the possibility of misregistration of segments between CMR and PET/SPECT cannot be excluded. This may have introduced a certain amount of bias in favour of CMR results owing to the easier registration of images for wall motion and viability diagnostics within the same technique. A technical limitation is that we did not use a PET flow tracer for the assessment of myocardial perfusion, which might have been desirable to overcome the limitations of the SPECT technique relative to the PET technique. In addition, a glucose-insulin clamp was not performed because it is not part of the routine clinical scanning protocol in our institution. As many PET centres do not dispose of a cyclotron for the generation of PET flow tracers and the insulin-clamp is relatively time consuming, it was our aim to compare the techniques in a routine clinical scenario. However, because we have previously validated our protocol for the assessment of reversible dysfunction in a similar patient population,8 we do not believe that these drawbacks might have significantly affected the results of the study. Not all patients included in the study completed the follow-up examination. These subjects had to be excluded from the final analysis, which could have been a source of selection bias. However, as the baseline characteristics of these patients were comparable with the study group, we do not believe that excluding these patients might have altered the message of the manuscript.

In conclusion, ce-CMR is comparable with a PET/SPECT hybrid protocol for the prediction of regional and global functional recovery after revascularization. In the present study, ce-CMR was superior to PET/SPECT imaging for predicting the lack of segmental functional recovery. This finding may be clinically important because ce-CMR may be especially useful to select patients who may not need coronary revascularization.

Supplementary material

Supplementary material is available at European Heart Journal online.


This research project is being supported by the Interdisciplinary Center for Clinical Research on Biomaterials and Tissue-Material-Interaction in Implants (within the Faculty of Medicine at the RWTH Aachen University).

Conflict of interest: none declared.


View Abstract