Background: The standardized uptake value (SUV) is the nearly exclusive means for quantitative evaluation of clinical [18F-]fluorodeoxyglucose (18F-FDG) positron emission tomography (PET) whole body investigations. However, the SUV methodology has well-known shortcomings. In this context, it has been recognized that at least part of the problems can be eliminated if tumor SUV is normalized to the SUV of a reference region in the liver (tumor-to-liver [TLR] ratio). In recent publications, we have systematically investigated the tumor-to-blood SUV ratio (SUR) for normalization of tumor SUVs which in our view offers principal advantages in comparison to TLR. The aim of this study was a comprehensive comparison of TLR and SUR in terms of quantification of tumor lesions. Methods: 18F-FDG PET/CT was performed in 424 patients (557 scans) with different tumor entities prior to radio(chemo)therapy. In the PET images, SUV_max of the primary tumor was determined. SUV_liver was calculated in the inferior right lobe of the liver. SUV_blood was determined by manually delineating the aorta in the low-dose CT. TLR and SUR were computed and scan time corrected to 60 min p.i. (TLR_tc and SUR_tc). Correlation analysis was performed for SUV_liver vs. SUV_blood, TLR vs. SUR, SUV_liver/SUV_blood vs. SUV_blood,SUR_tc/TLR vs. SUR_tc, and SUR_tc/TLR_tc vs. SUR_tc. Variability of the respective ratios was assessed via histogram analysis. The prognostic value of TLR and TLR_tc for distant metastases-free survival (DM) was investigated with univariate Cox regression in a homogeneous subgroup (N = 130) and compared to previously published results for SUV and SUR_tc. Results: Correlation analysis revealed a linear correlation of SUV_liver vs. SUV_blood (R ²=0.83) and of TLR vs. SUR_tc (R²=0.92). The SUV_liver/SUV_blood ratio (mean ± s.d.) was 1.47 ± 0.18. For the SUR_tc/TLR ratio, we obtained 1.14 ± 0.21 and for the SUR_tc/TLR_tc ratio 1.38 ± 0.17. Survival analysis revealed TLR and TLR_tc as significant prognostic factors for DM (hazard ratio [HR] = 3.3 and HR = 3, respectively). Both hazard ratios are lower than that of SUR_tc (HR = 4.1) although this reduction does not reach statistical significance for the given limited group size. HRs of TLR and SUR_tc are both significantly higher than HR of SUV (HR = 2.2). Conclusions: Suitability of the liver as surrogate of arterial tracer supply for SUV normalization via TLR computation is limited. Further studies in sufficiently large patient groups are required to better characterize the relative performance of SUV, TLR, and SUR in different settings.