Purpose: Modified Response Evaluation Criteria In Solid Tumors (mRECIST),developed by the American Association for the Study of Liver Diseases (AASLD) criteria measure changes in arterialized hepatocellular carcinoma (HCC) and aim at providing a common framework for the design of clinical trials.It still isn’t determined whether mRECIST can be applied in routine clinical practice and whether mRECIST could estimate viable tumor correctly.Methods: We retrospectively analyzed data from patients subjected to transcatheter arterial chemoembolization (TACE) as initial treatment for advanced HCC in our institution.Not suitable for using mRECIST standard cases and the agreement in response between RECIST and mRECIST were assessed.Then we selected HCC patients who achieved complete response (CR) according to mRECIST,following PET-CT examinations.We also compared arterial enhanced computed tomography (CT) or magnetic resonance imaging (MRI) with positron emission tomography (PET)-CT examination and analyzed their correlation.Results: Out of 143 HCC patients,mRECIST evaluation appeared to be applicable for 128 (89.51%) assessable patients.In these 128 assessable patients,the objective response (OR) rates (complete/CR+partial response/PR) according to RECIST and mRECIST were 64.06% (82 of 128 patients) and 78.13% (100 of 128; p<0.001),respectively.Discordance in the response evaluations between the two methods was observed in 46 patients (35.94%) and was statistically significant (Kappa=0.491; p<0.001).The overall survival (OS) of patients who achieved an OR as assessed by mRECIST or by RECIST was significantly better than the survival of non-responding patients (stable disease/SD,or progressive disease/PD).Conclusions: Although mRECIST criteria show a good correlation with prognosis,they demand strict requirements for patient selection and couldn’t be useful as a tool for routine clinical practice.Furthermore,merely by means of contrast-enhanced CT or MRI,mRECIST couldn’t estimate viable tumor sufficiently.