基于SEER数据库第8版肝癌AJCC分期的验证及改良

Validation and modification of AJCC 8th edition staging system for hepatocellular carcinoma--analysis based on SEER database

目的评估美国癌症联合委员会(AJCC)第8版肝癌分期的准确性并提出合理的改良。方法从SEER数据库(2004—2015年)中提取接受外科手术治疗的原发性肝癌患者,分析患者的总体生存及病种别生存情况。结果共纳入7911例患者的资料,其中男5794例、女2117例;60岁以上者4050例。肿瘤直径为24~65 mm,单叶(80.8%)、单发(62.8%)的病灶多见。其中接受放疗和化疗的患者分别为230例(2.9%)和2052例(25.9%)。中位随访时间42个月。第8版AJCC分期生存曲线发现,ⅣA期与ⅢA期的总体生存及病种别生存曲线均存在交点,差异均无统计学意义(均P>0.05)。ⅣA期亚组患者数据分析显示,T1N1M0/T2N0M0、T2N1M0/T3N0M0、T3N1M0/T4N0~1M0及T3N1M0/T1~4N0~1M1差异均无统计学意义(均P>0.05)。因此,在保留第8版AJCC分期的T/N/M定义的基础上,提出第8版改良AJCC分期:保留ⅠA期和ⅠB期,将ⅣA期拆分,T1N1M0纳入Ⅱ期,T2N1M0纳入Ⅲ期,Ⅳ期则包括T3N1M0或T4N0~1M0或T1~4N0~1M1。第8版改良AJCC分期的Cox比例风险回归分析显示,在总体生存中,ⅠB期/ⅠA期(HR=1.462,95%CI:1.294~1.651)、Ⅱ期/ⅠB期(HR=1.091,95%CI:1.003~1.186)、Ⅲ期/Ⅱ期(HR=2.034,95%CI:1.793~2.307)、Ⅲ期/Ⅳ期(HR=1.374,95%CI:1.192~1.583)差异均有统计学意义(均P<0.05);在病种别生存分析中,ⅠB期/ⅠA期(HR=2.007,95%CI:1.671~2.411)、Ⅱ期/ⅠB期(HR=1.140,95%CI:1.023~1.271)、Ⅲ期/Ⅱ期(HR=2.344,95%CI:2.018~2.724)、Ⅲ期/Ⅳ期(HR=1.391,95%CI:1.180~1.639)差异均有统计学意义(均P<0.05)。结论第8版改良AJCC分期可更准确地预测肝癌患者的预后。

Objective To assess the accuracy of the American Joint Committee on Cancer(AJCC)8th edition staging system for hepatocellular carcinoma(HCC)and to make an appropriate modification.Methods Data of patients diagnosed with HCC who underwent surgery were extracted from 2004 to 2015 within the SEER database.Overall survival(OS)and disease-specific survival(DSS)of patients were analyzed.Results A total of 7911 patients were included and there were 2117 females and 5794 males.The male-to-female ratio was 1.00:0.36.There were 4050 patients older than 60 years old.Tumor size ranged from 24 to 65 mm.Tumors with single lobes(80.8%)or single lesions(62.8%)were more common.There were 230 cases and 2052 cases received radiotherapy and chemotherapy,accounting for 2.9%and 25.9%,respectively.The median follow-up was 42 months.Analysis of the 8th edition of AJCC staging system showed that the survival curves ofⅣA stage andⅢA stage intersected in both OS and DSS,and the differences were not statistically significant between them(both P>0.05).Analysis of patients in subgroup ofⅣA stage showed that there was no statistically significant difference in the four groups of T1N1M0/T2N0M0,T2N1M0/T3N0M0,T3N1M0/T4N0-1M0 and T3N1M0/T1-4N0-1M1(all P>0.05).Therefore,the modified 8th edition of the AJCC staging system was proposed after retaining the definition of T/N/M in the old edition:ⅠA andⅠB stages were retained;ⅣA stage was split:T1N1M0 was included inⅡstage,T2N1M0 inⅢstage,andⅣstage included T3N1M0,T4N0-1M0 and T1-4N0-1M1.Cox proportional risk regression analysis of the modified 8th edition of the AJCC staging showed that significant differences were observed among the four groups,withⅠB/ⅠA(HR=1.462,95%CI:1.294-1.651),Ⅱ/ⅠB(HR=1.091,95%CI:1.003-1.186),Ⅲ/Ⅱ(HR=2.034,95%CI:1.793-2.307)andⅢ/Ⅳ(HR=1.374,95%CI:1.192-1.583)for OS,respectively.The similar findings were seen in DSS,withⅠB/ⅠA(HR=2.007,95%CI:1.671-2.411),Ⅱ/ⅠB(HR=1.140,95%CI:1.023-1.271),Ⅲ/Ⅱ(HR=2.344,95%CI:2.018-2.724)andⅢ/Ⅳ(HR=1.391,95%CI:1.180-1.639),respectively.Conclusion The modified AJCC 8th edition staging system could predict the survival outcome of HCC more accurately.