The exact size of hepatocellular carcinoma (HCC) lesions plays an important role in treatment algorithms of this disease^1-5. Therefore, the objective of our study was to compare size measurements of HCC lesions between different MRI sequences and to evaluate their accuracy using histopathology as the standard of reference.We included 53 patients with known HCC who underwent contrast-enhanced MRI of the liver prior to liver transplantation or tumor resection within 90 days. Pathologically assessed tumor-size was available for all patients. The MRI protocol contained axial T2-weighted sequences as well as T1-weighted sequences before and after application of Gd-EOB-DTPA. Tumor diameters were measured on all available MRI sequences. Furthermore the sharpness of lesions and the presence of artifacts were evaluated visually. MRI measurements and pathologically assessed dimensions were correlated using pearson’s correlation coefficient.Mean time between MRI and resection was 21±19 days. Artifacts were present in 9% of T2w turbo-spin-echo (tse) images, 6% of portal-venous-, and 4% of arterial-phase datasets, while there were no artifacts present in hepatobiliary-phase images. HCCs frequently showed up with fuzzy edges in arterial- and late-venous-phase datasets (49%, 45% respectively), while they presented sharply delineated in 85% of all hepatobiliary-phase images. On average there was a deviation from the pathologically measured tumor diameter of 0.93cm±0.89cm (R=0.93) for T2w tse images, 0.86cm±0.97cm (R=0.93) for the arterial-, 0.83cm±0.93cm (R=0.94) for the late-venous-, and 0.79cm±0.82cm (R=0.95) for the hepatobiliary-phase.The hepatobiliary-phase delineates HCC lesions sharply and seems to be most accurate along commonly used MRI sequences to measure HCC tumor-size.