Abstract
1604
Objectives We assessed that whether presences of steatohepatitis (SH) could be estimated from FDG-PET images and influence on the detectability of liver metastases (LMs) in FDG-PET+plain CT (PET+CT).
Methods i) 930 PET cancer screening examinees without known history of viral, alcoholic or other hepatitis, or malignant lesions in the liver who underwent PET+CT and serum blood test were enrolled. We evaluated the correlations between Max, Mean, and Standard Deviation (SD) of SUVs (mesured by placing ϕ6cm VOI) and CT values of normal livers, and presences of SH using these cases. ii) 380 cases with clinically-diagnosed LMs who underwent PET+CT and contrast-enhanced CT were enrolled. We assessed the correlation between presence of 'SH patterns' (increased SDs of SUVs and decreased CT values of livers) and detectability of LMs (whether all 8mm or larger LMs were depicted or not) in PET+CT.
Results i) By multivariate analyses, significant positive correlations were seen between each SUV values and serum AST, ALT, γGTP level, and between SUVmean and CT value. But only SD of SUV had significant correlation with presence of SH. When ‘SH cases’ were defined as the cases with CT value < 60HU and elevation of either serum AST, ALT, or γGTP, 157 cases out of 930 examinees of PET cancer screening would be the cases with SH. In 129 cases (82.2%) out of these 157 'SH cases', SDs of liver SUVs were 0.35 or more, and in 726 cases (93.9%) out of 773 'non-SH cases', SDs were less than 0.35. ii) When 'SH patterns' of PET+CT were defined as 'SD of liver SUV≧0.35 and CT value of liver < 60HU', 93 cases out of 380 examples would be the case with SH patterns. In 267 cases (90.6%) out of 287 cases with 'non-SH patterns', and only 41 cases (44.1%) out of 93 case with 'SH patterns', PET+CT could depict all 8mm or larger LMs.(p<0.00001).
Conclusions Our study indicated that FDG-PET images could estimate presence of SH. SH might make normal liver uptake non-equalized in FDG-PET, and decrease detectability of LM in PET+CT