Multivariate Analysis Of Prognostic Factors In Unresectable HCC Retreatment With TACE And Establishment Of Risk Prediction Model

Objective: Analyse independent prognostic factors in unresectable hepatocellular carcinoma(HCC)patients retreated with transcatheter arterial chemoembolization(TACE)in a single center.Establish or confirm a simple prognostic scoring system about retreatment with TACE in HCC patients.Validate the ART and ABCR score,and compared them with the new scoring system in the ability of distinguishing high-risk patients from low-risk ones.Metheds: We retrospectively analyzed the clinical data of patients with HCC treated with TACE at least twice in our single center from April 2007 to December 2013,a total of 146 patients were included.Used Kaplan-Meier and COX regression model to find independent prognostic factors from demographic,liver function,tumor factors,radiologic response and the change of liver function after the intial treatment,and then establish or confirm a simple prognostic scoring system about retreatment with TACE in HCC patients.Calculate the ART and ABCR score,and used C-index to compare the accuracy of prognosis evaluation among the ART,ABCR and our new scoring system.Results:1.COX multi-factor analysis showed the multiple tumors,BCLC stage C,the first TACE treatment without tumor response(SD and PD)and AFP more than 400 ng/m L before the second TACE were independent prognostic factors.2.Using the four indexes formed a new prognostic score(range from 0 to 5 points),with the score increased,the survival time shortened correspondingly,with 0-1,2-3 and 4-5 points can distinguish low risk,medium risk and high risk patients with different prognosis.There were 57,61,28 patients in the three groups respectively.Median overall survival(OS)of the three groups were 27.6 months(95%CI 14.5-44.7),14.2 months(95%CI 11.0-17.3)and 6.7 months(95%CI 3.1-10.3,)respectively.3.The ABCR score was: 73 patients had a score ≤0,55 patients had a score from 1 to 3,and the reminder,18 patients had a score≥4,median OS was 21.9 months(95%CI 14.4-29.4),11.9 months(95%CI 8.5-15.3),7.2 months(95%CI 4.3-10.1)(P<0.001).The ART score was: 108 patients had a score ≤1.5,and the reminder,38 patients had a score≥2.5,median OS was not statistically different(p=0.059): 13.5 months(95%CI 7.9-19.1)in the low-score group and 11.9 months(95%CI 8.9-14.9)in the high-score group.4.Using C-index to evaluate the precision of the scores,the C-index value for our scoring system was 0.624,for ABCR was 0.603,and ART was 0.572.Conclusions: Multiple tumors,BCLC stage C,the first TACE treatment without tumor response(SD and PD),AFP more than 400ng/m L before the second TACE treatment were independent risk factors.Using the four indexes formed a new prognostic scoring system ranging from 0 to 5 points,with 0-1,2-3 and 4-5 points can roughly distinguish three groups of different prognosis patients.The OS of high-risk group(4-5)was short,patients may not benefit from the second TACE,which were needed to be changed with other treatments.The new scoring system required further research to testify it’s evaluation accuracy before apply to clinical use.ABCR scoring system had limited predictive power and predictive ability was lower than our new scoring system.ART score predictive power was relatively weak compared to the other scoring system.