新型炎性指标AAR-GPR对肝细胞癌患者术后预后的预测价值

Prognostic prediction values of a novel inflammatory index AAR-GPR for hepatocellular carcinoma patients after hepatectomy

背景与目的:肝细胞癌(HCC)发病隐匿,病死率高。肿瘤标志物对于HCC的早期诊断、预后判断和疗效监测具有重要意义。研究表明炎性指标与HCC的预后密切相关,因此本研究探讨术前炎性复合指标碱性磷酸酶-白蛋白比值(AAR)、谷氨酰转肽酶-血小板比值(GPR)与HCC切除术后预后的关系。方法:回顾性分析中国人民解放军空军军医大学西京医院肝胆外科2014年1月—2017年12月间347例符合标准的行HCC切除术患者资料。使用随机数生成器,将347例患者按照7∶3随机分为训练集(243例)与验证集(104例)。在训练集中,通过X-tile软件,依据生存资料获取AAR和GPR的最佳截断值,分别在训练集和验证集中分析AAR、GPR或两者联合指标(AAR-GPR)连同其他临床病理因素与患者生存的关系。分别比较训练集和验证集中不同AAR或GPR水平以及不同AAR-GPR评分(AAR与GPR均高于截断值为2分,其余为1分)患者生存率的差异。通过R3.2软件分析,比较AAR、GPR和AAR-GPR对预后的预测价值。结果:347例患者的中位随访时间为45个月,共有147例死亡,1、3、5年的累积生存率分别为84.6%、59.4%、52.2%。训练集和验证集之间,除了在AJCC分期上有差异(χ;=6.21,P=0.045),其他变量差异均无统计学意义(均P>0.05)。AAR的最佳截断值为2.61,GPR的最佳截断值为0.39。训练集和验证集的分析均显示,AAR (训练集:HR=1.904,P=0.003;验证集:HR=2.245,P=0.008)、GPR (训练集:HR=2.159,P=0.002;验证集:HR=2.646,P=0.006)、AAR-GPR (训练集:HR=2.872,P<0.001;验证集:HR=4.643,P<0.001)均为影响术后生存期的独立危险因素。AAR-GPR对预后的预测价值(C指数=0.831,似然比=24.36)优于单独使用AAR (C指数=0.765,似然比=12.11)或GPR (C指数=0.772,似然比=13.43)。结论:AAR、GPR均是影响HCC切除术预后生存期的独立危险因素。与单独使用AAR或GPR相比,两者联合能明显提高预测效能,更准确地预测术后生存。由于为单中心回顾性研究,后续仍需要大样本量、多中心的高质量研究进一步验证。

Background and Aims: Hepatocellular carcinoma(HCC) has an insidious onset and high mortality.Tumor makers have great importance in early detection, prognostic estimation and efficacy monitoring.Studies have demonstrated that inflammatory indices are closely associated with the prognosis of HCC.Therefore, this study was conducted to investigate the relations of preoperative integrated inflammatory marker alkaline phosphatase to albumin ratio(AAR) and gamma-glutamyl transpeptidase to platelet ratio(GPR) with the postoperative prognosis of HCC.Methods: The clinical data of 347 eligible HCC patients undergoing hepatectomy in the Department of Hepatobiliary Surgery of Xijing Hospital from January 2014 to December 2017 were retrospectively analyzed. The patients were randomly divided into training set(243 cases) and validation set(104 cases)in a ratio of 7∶3 by a random number generator. The optimal cutoff values of AAR and GPR were obtained by the X-tile software according to the survival data in training set. The associations of AAR,GPR and their combination index(AAR-GPR) as well as other clinicopathologic variables with the survival of patients were analyzed in training set and validation set, respectively. The differences in survival rates among patients with different AAR or GPR values or different AAR-GPR scores(those with both AAR and GPR above the cutoff values were defined as 2, and the remaining conditions were classified as 1) were compared in training set and validation set, respectively. The prognostic prediction values among AAR, GPR and AAR-GPR were compared by analysis of R3.2 software.Results: The median follow-up time of the 347 patients was 45 months. A total of 147 patients died, and the cumulative 1-, 3-and 5-year survival rates were 84.6%, 59.4% and 52.2%, respectively. There were no statistically significant differences(P>0.05) between training set and validation set except for AJCC stage(χ;=6.21,P=0.045). The optimal cutoff value for AAR was 2.61 and for GPR was 0.39. Results of analyses in both training and validation sets showed that AAR(training set: HR=1.904, P=0.003;validation set: HR=2.245, P=0.008), GPR(training set: HR=2.159, P=0.002;validation set: HR=2.646, P=0.006) and AAR-GPR(training set: HR=2.872, P<0.001;validation set: HR=4.643, P<0.001) were independent risk factors for postoperative survival. The prognostic prediction value of AAR-GPR(Cindex: 0.831, likelihood ratio: 24.36) was superior to that of either AAR(C-index: 0.765, likelihood ratio: 12.11) or GPR(C-index: 0.772, likelihood ratio: 13.43) alone.Conclusion: Both AAR and GPR are independent risk factors for survival after HCC resection.Compared with AAR or GPR alone, their combination index can significantly improve the predictive efficiency and predict postoperative survival more accurately. Since this is a single-center study, further validation is still needed by large-sample size and multi-center high-quality studies.