Keywords
cardiac - gadolinium - myocardium - PET-CT
Introduction
Ischemic heart disease is a major problem worldwide, although the mortality attributed
to this condition has gradually declined over the last decades in the west,[1]
[2]
[3] due to the improvement in health care and awareness among the people. However, it
is still responsible for one-third of all deaths in people older than 35 years and
remains the major cause of morbidity worldwide.[1]
In this era of precision medicine, imaging myocardial viability is very essential
for providing adequate care to the patient. Viability of myocardium is assessed by
various methods and techniques, including perfusion analysis, motion analysis, and
anatomical analysis.[4]
[5]
[6]
[7]
[8] Studies have been done comparing the perfusion assessment of SPECT and cardiac magnetic
resonance (CMR), proving the superiority of the latter mainly in terms of sensitivity
in identifying the sub-endocardial infarct and small infarcts in the inferior wall
in chronic as well as in acute infarction.[7]
Ample studies also show CMR is comparable with positron emission tomography (PET)
in viability assessment.[4]
[5]
[6] However, widespread usage of CMR needs much more supportive evidence. The aim of
this study was to evaluate the various CMR tools to assess viability in comparison
to PET-CT perfusion assessment.
Materials and Methods
Study Design
This was a prospective analytical study. A total of 28 consecutive adult (age >18
years) patients referred to PET-CT from the cardiology department for persistent heart
failure symptoms in spite of prior surgical/percutaneous intervention, who showed
matched or mismatched defects, were taken for MR evaluation typically within a week
after the PET-CT completion. This study was done for a duration of one and a half
years.
Patients with contraindications for MRI, with acute cardiac symptoms, with a history
or suspicion of other causes of cardiomyopathy, severe arrhythmia, and structural
heart diseases, were excluded from the study.
Informed consent was obtained from all the included patients. All guidelines as per
the Declaration of Helsinki and good clinical practice guidelines were followed.
Imaging Parameters
Patients with perfusion/metabolism matched or mismatched defects were evaluated with
MR, assessing late gadolinium enhancement (LGE), functional cine images and tagging
sequence to analyze the presence of scar, the presence or absence of residual regional
contraction, and end-diastolic myocardial thickness in the hibernating myocardium.
Image Acquisition
-
PET-CT data are acquired with 13N ammonia and fluorodeoxyglucose (FDG) for calculating
the perfusion and metabolic defects.
-
All CMR examinations were performed on a 3T system (Ingenia; Philips Healthcare, Best,
the Netherlands). MRI sequence is as follows:
-
– Dark blood—axial.
-
– Bright blood balanced turbo field echo—axial.
-
– Cine—horizontal long axis.
-
– Cine—vertical long axis.
-
– Cine—short axis.
-
– Dynamic perfusion.
-
– Early gadolinium enhancement.
-
– LGE.
-
– Tagging—grid.
Image Analysis
PET-CT perfusion data and CMR were analyzed by two nuclear medicine specialists and
two radiologists, respectively. Any discrepancy was sorted by mutual consensus. Radiologists
were blinded to the PET-CT results to improve objectivity. Image reporting was done
after segmentation of the myocardium as per the American Heart Association (AHA) model
([Fig. 1]). The AHA model helped to compare the MRI and PET-CT data objectively.
Fig. 1 Polar plot of myocardial segments of the left ventricle (the American Heart Association
model).
Patients with defects in the 13N ammonia perfusion study and FDG study were deemed
as matched defects in the particular segment ([Fig. 2]). However, no defects in the FDG study were deemed as mismatched defects in the
particular segment. Segment-wise data were collected for all the segments of myocardium,
similar to the MRI data ([Fig. 3]).
Fig. 2 (A–F) A 53-year-old man with chest pain and dyspnea for the past 6 months with an
ejection fraction of 30%. PET NH3 vertical long axis stack images showed perfusion
defect along the anterior wall of the left ventricle; corresponding FDG images show
minimal uptake, suggesting a matched defect–non-viable myocardium. (G-I) Cine short-axis
images of the same patient at the end-diastolic phase show diffuse thinning of
myocardium more pronounced along the septum (end-diastolic wall thickness 3.7 mm)–non-viable
(G). On tagging (H) and cine sequences (not shown); there was grade 4–severe hypokinesia.
(I) Vertical long–axis section shows diffuse myocardial thinning along the anterior
wall of the left ventricle with subtle but near–complete enhancement
(grade 3) s/o non-viable myocardium in concordance with PET result.
Fig. 3 (A–F) A 62-year-old male with chronic LV dysfunction short-axis stacks. PET NH3 (A–C) shows a defect in the basal inferior segment, corresponding FDG images (D–F) show uptake, suggesting a mismatch defect—viable but hibernating myocardium. (G–I) Cine short-axis images of the same patient showing thinning with T2 hypointensity
in the basal inferior segment end-diastolic wall thickness was ∼4.7 mm (G), showing subendocardial late gadolinium enhancement (<25%) (H). On tagging sequence, poor grid crunching in the basal inferior segment is seen
(I).
A Likert scale was utilized in reporting regional wall motion abnormalities (RWMA)
and LGE. Scores 1 to 5 represent normokinesia to akinesia or dyskinesia. Scores 1
to 4 were assumed as characteristics of viable myocardium. Score 5 is assumed to be
the characteristics of non-viable myocardium. RWMA was analyzed using both cine and
grid sequence (tagging). The grid sequence compartmentalized the myocardial thickness
into three layers from the inner to outer surface. Score 1: if all the layers are
contracting equally, Score 2: if the outer 2 layers are contracting, Score 3: if only
the outer layer is contracting, Score 4: if there is sluggish contraction not enough
to place in any of the grades, and Score 5: if no contraction or dyskinetic contraction.
Grading was given to reduce the subjectivity between the observers, and the grid sequence
was also helpful for the same purpose.
Grades 0 to 4 were given according to the extent of the enhancement from subendocardial
to transmural. In our study, LGE was graded according to the percentage of transmural
involvement into grade 1 (0–25%), grade 2 (26–50%), grade 3 (51–75%), and grade 4
(76–100%). Grades 1 and 2 were considered viable and grades 3 and 4 were considered
non-viable.[9]
Left ventricular ejection fraction (EF), end-diastolic and systolic volumes, and stroke
volumes were calculated for each patient along with the amount of enhancement for
the entire myocardium. Semiautomatic analysis was used for calculating the EF from
the MRI data.
Statistical Analysis
Analysis was done for individual myocardial segments with comparison of various cardiac
MRI parameters with PET-CT data. All the statistical tests were two-sided and were
performed at a significance level of α = 0.05. Concordance and discordance were calculated by the method of the kappa test
of agreement. Sensitivity and specificity of each test were assessed using the area
under the curve generation for both modalities.
Results
A total of 28 patients aged between 31 and 87 years (mean = 51.5 years) were included
in the study. All the patients were symptomatic at the time of study, with 71.4% having
dyspnea, 21.4% having orthopnea, 21.4% having chest pain, and 17.9% having pedal edema
as chief presenting complaints. Out of these, 32.14% (n = 9) had triple vessel disease. The distribution of vessel disease is as follows.
Nearly all the patients had a previous history of acute coronary syndrome and were
treated either with coronary artery bypass graft (CABG) (46.4%; n = 13) or percutaneous coronary intervention (50%; n = 14). All of them were on anti-failure drugs at the time of the study.
Mean EF calculated using echocardiography was 27.54% (27–32%) with SD of 8.47, and
mean EF calculated using cardiac MRI semiautomatic software was 28.96% with SD of
12.05. There is good agreement between EF MRI and EF echo (p < 0.005).
According to the PET results, 6% (n = 27) of the segments are considered not viable with a matched defect. Remaining
segments were considered viable, which were either normal (56.5%; n = 253) or hibernating—having a mismatched defect of 37.5% (n = 168). Out of 168 mismatch segments, 79.8% (n = 134) had normal wall thickness, whereas 55.6% (n = 15) of the matched segments also showed normal wall thickness. End-diastolic wall
thickness (EDWT) results had very good sensitivity but poor specificity as far as
PET results are concerned (p < 0.005).
Distribution of LGE Grading with Pet Results
Out of 168 mismatched segments, 33.3% (n = 56) showed no enhancement; 23.8% (n = 40) showed subendocardial enhancement; 28% (n = 47) showed transmural enhancement but less than half of myocardial thickness; 14.8%
(n = 15) of the mismatched myocardium were considered non-viable by LGE ([Fig. 4]); 55% (n = 15) of the myocardium considered non-viable by the PET was considered viable according
to LGE, a majority of them showing transmural enhancement involving less than half
of the myocardium (37%, n = 10). Out of 253 normal myocardium according to PET, 83% (n = 210) showed no enhancement; 13.1% (n = 33) showed enhancement but considered viable; and 4% (n = 10) were considered non-viable.
Fig. 4 (A–F) A 57-year-old male with LV dysfunction. PET NH3 short-axis images show perfusion
defect in the basal inferolateral segments (A); corresponding FDG images also show
poor uptake (B) s/o matched defect–non-viable myocardium. Cine short-axis image (C)
of the same patient in end-diastolic phase shows focal thinning in the basal inferolateral
segment; however, average end-diastolic wall thickness of 6.8 mm (F) which is suggestive
of a viable myocardial segment. LGE sequence (D) shows grade 2 (less than half of
myocardial thickness) late gadolinium enhancement, suggesting viable myocardium discordant
with the PET results. Tagging sequence (E) of the same patient showed grade II contraction.
Out of 168 mismatched segments, 68.5% (n = 115) showed some wall motion, whereas 31.5% (n = 53) did not show any movement or dyskinetic movement; 29.6% (n = 8) also showed some movement. Using the tagging sequence, this percentage was reduced,
thereby increasing the specificity, as only 11.1% of the matched myocardium showed
some movement. But tagging did not show any significant difference in mismatched myocardium,
as 51.2% of mismatched myocardium did not show any movement, thereby drastically reducing
its sensitivity for the identification of mismatched myocardium identified by PET.
Out of 168 mismatch segments, 79.8% (n = 134) had normal wall thickness; 55.6% (n = 15) of the matched segments also showed normal wall thickness and 33.3% (n = 56) of them showed no enhancement; 14.8% (n = 15) of the mismatched myocardium were considered non-viable by LGE ([Fig. 5]).
Fig. 5 Distribution of late gadolinium enhancement grading with PET results.
LGE versus PET in Variable LV Function
Patients were divided into two groups based on left ventricular (LV) EF echo <30%
and >30%. Increased scar detection by cardiac MRI is seen in reduced LV function <30%.
Overall, MRI detects a larger number of scars than PET. There is also increased accuracy,
detecting scars in the lower LV functions (EF < 30%). There is a mild increase in
specificity in higher LV function but is not significant (p > 0.005).
RWMA versus PET
Out of 168 mismatched segments, 68.5% (n = 115) showed some wall motion, whereas 31.5% (n = 53) did not show any movement or dyskinetic movement; 29.6% (n = 8) also showed some movement. Using tagging sequence, this percentage was reduced,
thereby increasing the specificity, as only 11.1% of the matched myocardium showed
some movement.
RWMA versus LGE
There is good agreement of motion abnormalities with LGE. The more the extent of transmural
scar, the more severe is the hypokinesia of the involved segment ([Fig. 6]).
Fig. 6 Distribution of cine regional wall motion abnormalities with late gadolinium enhancement
grading with PET results.
Discussion
Among patients with CAD, fraction of the population showed chronic LV dysfunction,
which is potentially reversible if successfully managed by either by CABG or percutaneous
coronary intervention. These patients are shown to have ischemic but viable myocardium,
otherwise known as hibernating myocardium.[10]
[11] Hibernating myocardium is one whose contraction is chronically depressed because
of chronic ischemia as an adaptive response due to chronic ischemia and the scope
of regaining function on revascularization.[12] Stunned myocardium, which have different pathophysiology, is the prolonged dysfunction
after revascularization and caused by multiple mechanisms as discussed in the review
article by Conti[13] who has also described various ways of differentiating it.
PET with FDG uses glucose utilization as a marker of viability by reducing the plasma
FFA with glucose loading or by fasting, causing the ischemic myocardium to preferentially
use the FDG, the former being performed more commonly.[14]
As discussed previously, nuclear imaging techniques were considered the gold standard
for viability assessment due to their high sensitivity and negative predictive value
(e.g., FDG-PET has 92 and 87% sensitivity and specificity, respectively).[15] Large multicenter trials conducted by Beanland et al[15]—PARR1 and PARR2 that compared (FDG -PET) directed revascularization with standard
care—showed that patients considered viable by FDG-PET on revascularization had improvement
in myocardial function.[16] On comparison with SPECT, FDG-PET had better sensitivity and a good agreement of
82% with SPECT.[17] Deciding upon the optimal management protocol for these patients is a complex, multifactorial
process that has to consider not only viability but also processes such as ischemia
and remodeling.[18]
In a study by Selvanayagam et al, there was a strong correlation between the extent
of LGE and the recovery of RWMA at 6 months (p < 0.001). The study stated that LGE MRI is a powerful predictor of viable myocardium
and improvement of function after surgery, suggesting an important role for this technique
as a viability assessment tool.[19]
Krittayaphong et al in their study showed EDWT as an independent predictor of functional
recovery of hibernating myocardium but to a lesser extent compared with LGE. Their
study demonstrated a cut-off of 5.5 mm.[20]
Limitation
-
• Viability assessment in the study was purely assessed and compared between LGE and
PET findings. Myocardial wall thickness and myocardial edema were not included in
the analysis for this research, which also constitute important predictors of myocardial
viability and may be considered for future research.
Conclusion
Although LGE did not give a significant advantage over the PET-CT in identifying hibernating
myocardium. Cardiac MRI, as a versatile tool, can provide us with various information
about the patient's physiology and guide us more precisely in the patient management.
Improved objectivity and precision were identified in EF analysis and RWMA, especially
with the tagging sequence. This study, although with some limitations, proves that
there is an important and definite place for cardiac MRI in the evaluation of patients
with persistent symptoms of heart failure. Further research into other imaging features
depicting cardiac viability is recommended for improving the precision of cardiac
MRI in these clinical settings.
