Assignment of liver allocation priority for hepatocellular carcinoma is predicated on accurate imaging staging. We analyzed radiographically defined stage (radiologic stage [RS]) at listing and most recent extension and pathologic stage (PS) data from 789 liver transplant recipients for whom no pretransplant ablative treatment was given. There were no predetermined imaging or pathological protocols in this retrospective analysis of wait list data. Seventy-two (9.1%), 690 (87.5%), and 27 (3.4%) were listed as stage 1, 2 and >2, respectively. Computed tomography (CT) scan alone (46.4%), magnetic resonance image scan alone (37.1%), ultrasound alone (1.3%), and multiple imaging studies (15.2%) were used with no difference in time to transplant for listing or most recent scan among the recipient groups. Overall accuracy (RS = PS) was 44.1% and was not different if original listing RS or most recent RS was used for comparison with PS. No one type of imaging technique had superior accuracy (P = 0.13); however, CT scan used alone or in combination compared to not being used at all, had higher odds of being accurate (odds ratio [OR] 1.38 [1.03-1.84], P = 0.031). In addition, imaging done less than 90 days before transplant had higher odds of being accurate (OR 1.49 [1.06-2.08], P = 0.019) as did RS = 2 or 3 (OR 5.56 [2.70-11.11], P < 0.0001). We observed considerable variation in RS accuracy among the United Network for Organ Sharing and Organ Procurement and Transplantation Network regions that is unexplained. In conclusion, current imaging requirements for RS prior to liver transplantation are unacceptably inaccurate. Future policy should require more accurate modalities or combinations of techniques. Liver Transpl 12:1504-1511, 2006. © 2006 AASLD.