SCMR 22nd Annual Scientific Sessions
Evaluation of late gadolinium enhancement imaging (LGE) for stress perfusion cardiac magnetic resonance imaging (SP-CMR) in patients with suspected coronary heart disease (CHD) is necessary for determination of ischemic but viable myocardium that can facilitate guidance of invasive revascularization. Semi-quantitative signal threshold techniques for LGE quantification provide more accurate determination of myocardial viability but are more time consuming in routine clinical practice. We sought to determine the accuracy of routine qualitative LGE quantification compared with semi-quantitative analysis by core-laboratory.
We examined 512 patients referred for SP-CMR based on symptoms suspicious for CHD. All patients were scanned at 3.0 Tesla CMR (Skyra or Prisma, Siemens, Germany) with LGE imaging 10 minutes following intravenous administration of 0.15 mmol/kg of gadolinium (Gadovist, Bayer). Qualitative LGE was scored using standardized 16-segment American Heart Association (AHA) model for the left ventricle, each segment divided into 4 transmural zones (Acuity, Cohesic Inc, Calgary). Signal threshold-based analysis of LGE was conducted by an independent core-laboratory blinded to all clinical data using cut-off of >2 standard deviations (SD) above mean signal of normal myocardium for 16-segment AHA model (Circle CVI, Calgary).
Results: Overall, 496 (97%) patients achieved diagnostic imaging quality for LGE analysis. Mean age of patients was 59.1±13.6 years (35% female) with 72 (15%) reporting prior history of myocardial infarction. Semi-quantitative analysis identified 112 (23%) patients with subendocardial LGE (Table 1) and mean LVEF of 56.4 ± 14.2. Mean global LGE percentage of myocardium was greater by routine qualitative quantification but this difference did not reach statistical significance (3.7 ± 7.2 vs 3.0 ± 7.7, P=0.18). Linear regression analysis between qualitative and semi-quantitative techniques demonstrated high correlation for both the total cohort (r=0.87, P=<0.0001, Figure 1A) and stratified by sex (r=0.89, P<0.0001 males; r=0.69, P<0.0001 females; Figure 1B). Segmental analysis of left ventricular segments for poor myocardial viability by qualitative LGE analysis showed high diagnostic performance for both segments with >50% (AUC 0.94, 95% CI 0.91 to 0.96, Figure 2A) and >75% (AUC 0.96, 95% CI 0.93 to 0.99, Figure 2B) LGE.
In a large cohort of patients referred for stress perfusion CMR, routine qualitative subendocardial LGE quantification showed strong agreement with semi-quantitative signal-threshold based analysis. Routine semi-quantitative LGE quantification for myocardial viability assessment of stress perfusion CMR may not provide additional clinical utility.