Abstract
1697
Objectives Gated-PET assessment of segmental left ventricular (LV) function is poorly documented in small animal. We aimed to compare FDG-PET to cardiac MR in an animal model of myocardial infarction.
Methods 89 myocardial segments were evaluated in mid-ventricular short-axis view in 16 rats after myocardial infarction (n=15 coronary ligation) or sham-operation (n=1). Cine-MR (cardiac and respiratory gating) was acquired (Flash 2D) 24h after surgery (Biospin 7T, Brucker). Rest ECG-gated PET (Inveon, Siemens) was acquired at H48, 30 min after a rest FDG injection (37 MBq). Segmental LV function was assessed for each segment (i) quantitatively as the percentage of systolic count increase by PET, and as the percentage of wall thickening by MR, and (ii) qualitatively using a 4-grade scale (PET & MR). A series of PET acquisitions was also acquired using a cylindrical phantom of increasing thickness filled with a solution of F-18 and segmented using CT. All PET reconstructions were performed using 3D-OSEM (16 subsets, 4 iterations, 128x128 matrix) and attenuation correction.
Results The relation between F-18 relative counts and phantom thickness was linear until a thickness of 2.6mm. In rats, the agreement of visual analysis between PET and MR was good (Kappa: 0.73) but PET underestimated wall thickening (PET: 24%±11% vs. MR: 32%±21%, P=0.001), the difference being significant in segments with normal FDG uptake (mean difference= -9.1, -3.9 and 0.9 respectively for FDG≥60% [P=0.0005], FDG 50-60% [P=NS] and FDG<50% [P=NS]). Systolic thickness evaluated by MR increased with FDG uptake (<50%: 2.03±0.5 mm, 50-60%: 2.17±0.44 mm, ≥ 60%: 2.64±0.38 mm, P<0.0001). In 30 segments with systolic thickness > 2.6 mm, the difference between MR and PET wall thickening was high (22±18% vs. 0.2±17%, P<0.0001), and 27/30 (90%) had a FDG uptake ≥60% (P=0.0004).
Conclusions In rats, preclinical FDG gated-PET underestimated segmental thickening in normal segments, but provided valuable assessment within infarcted areas