术前血液炎性标志物对胸段食管鳞状细胞癌生存预后的影响

Predictive role of preoperative henmtological inflatmmtory markers for patients with thoracic esophageal squacell carcinoma receiving surgery

目的 探索术前中性粒细胞与淋巴细胞比值（NLR）、血小板与淋巴细胞比值（PLR）和淋巴细胞与单核细胞比值（LMR）等血液炎性指标在胸段食管鳞状细胞癌生存预后方面的价值.方法 选取2010年1月至2012年12月在扬州大学附属泰兴人民医院胸外科接受食管癌根治术的117例患者,根据NLR、PLR和LMR的中位值2.8、127.3、3.8,分别将患者分为NLR≥2.8组和NLR＜ 2.8组、PLR≥127.3组和PLR＜ 127.3组、LMR≥3.8组和LMR＜ 3.8组,比较患者一般生存情况,分析NLR、PLR、LMR对患者中位无进展生存时间（PFS）、中位总生存时间（OS）的影响,COX回归模型进行单因素和多因素分析,受试者工作特征（ROC）曲线评估NLR、PLR和LMR对患者预后的预测价值.结果 117例胸段食管鳞状细胞癌患者中位PFS为17个月,1、3、5年PFS率分别为66.7％、21.4％和17.9％;中位OS为36个月,1、3、5年OS率分别为94.9％、46.2％和28.2％.其中,患者NLR、PLR、LMR的高低均与食管癌术后肿瘤复发相关（均P＜0.05）,在NLR＜ 2.8组与≥2.8组患者中,其中位PFS分别为24个月（95％ CI为19.788～ 28.212）和13个月（95％ CI为10.153 ～ 15.847）,1、3、5年PFS率分别为78.9％、35.1％、31.6％和55.0％、8.3％、5.0％（x2=15.601,P＜0.001）;在PLR＜ 127.3组与≥127.3组患者中,其中位PFS分别为24个月（95％ CI为19.891 ～28.109）和15个月（95％CI为11.832 ～ 18.168）,1、3、5年PFS率分别为78.0％、30.5％、27.1％和55.2％、12.1％、8.6％ （x2=7.621,P=0.006）;而在LMR＜ 3.8组与≥3.8组患者中,其中位PFS为14个月（95％ CI为11.534 ～ 16.466）和21个月（95％ CI为16.783 ～ 25.217）,1、3、5年PFS率分别为57.9％、8.8％、5.3％和75.0％、33.3％、30.0％ （x2=10.201,P=0.001）.另外,在NLR ＜2.8组与≥2.8组患者中,中位OS分别为42个月（95％ CI为29.188～48.282）、27个月（95％ CI为20.358～33.642）,其1、3、5年OS率分别为98.2％、56.1％、47.4％和91.7％、36.7％、10.0％（x2=19.161,P＜0.001）;在PLR＜127.3组与≥127.3组患者中,其中位OS分别为38个月（95％ CI为31.310～44.690）、27个月（95％ CI为19.537 ～ 34.463）,1、3、5年OS率分别为94.9％、54.2％、37.3％和93.1％、37.9％、19.6％ （x2 =7.019,P=0.008）;在LMR ＜3.8与≥3.8的患者中,中位OS为30个月（95％ CI为23.659 ～ 36.341）和38个月（95％ CI为27.878 ～ 48.121）,1、3、5年OS率分别为91.2％、36.8％、12.3％和95.0％、55.3％、43.3％（x2=10.201,P=0.001）.单因素分析结果显示,T分期（HR=1.292,95％CI为1.077 ～2.211,P=0.048）、N分期（HR=1.773,95％ CI为1.186 ～2.651,P=0.005）、TNM分期（HR=1.768,95％CI为1.181 ～2.645,P =0.006）、NLR（HR=2.193,95％ CI为1.450 ～3.316,P＜0.001）、PLR（HR=1.722,95％CI为1.149 ～2.581,P=0.009）和LMR（HR =0.531,95％ CI为0.353 ～0.799,P=0.002）均与食管鳞状细胞癌术后患者的PFS密切相关,T分期（HR=1.982,95％ CI为1.162 ～3.383,P=0.012）、N分期（HR=1.910,95％CI为1.243 ～2.934,P=0.003）、TNM分期（HR =2.115,95％ CI为1.375 ～3.252,P=0.001）、NLR（HR=2.599,95％ CI为1.657 ～4.078,P＜0.001）、PLR（HR=1.764,95％CI为1.145～2.717,P=0.010）和LMR（HR=0.470,95％ CI为0.303～0.728,P=0.001）也均与食管鳞状细胞癌术后患者的OS密切相关.但在多因素COX回归模型中显示,只有TNM分期（HR=1.608,95％CI为1.057～2.445,P=0.026）与NLR（HR=1.886,95％CI为1.133 ～3.138,P=0.015）是PFS的独立预后因素,并且TNM分期（HR=1.867,95％ CI为1.190 ～2.928,P=0.007）与NLR（HR =2.226,95％CI为1.292 ～3.835,P=0.004）也是OS的独立预后因素.ROC曲线分析NLR、PLR和LMR对患者PFS的诊断价值,结果显示NLR、PLR和LMR的曲线下面积（AUC）分别为0.725（95％ CI为0.615～0.835,P=0.001）、0.657（95％ CI为0.533～0.781,P=0.025）和0.290（95％CI为0.178～0.402,P=0.003）;ROC曲线分析NLR、PLR和LMR对患者OS的诊断价值,结果显示NLR、PLR和LMR的AUC分别为0.731（95％CI为0.632 ～0.829,P＜0.001）、0.613（95％CI为0.501 ～0.726,P=0.057）和0.308（95％ CI为0.205 ～0.412,P=0.053）.结论 在预测食管鳞状细胞癌根治术患者的生存预后方面,NLR优于PLR和LMR.NLR在预测胸段食管鳞状细胞癌术后患者生存预后方面具有重要价值.

 Objective To investigate the prognostic values of systemic inflammatory markers,including preoperative neutrophil-to-lymphocyte ratio （NLR）,platelet-to-lymphocyte ratio （PLR） and the lymphocyte-to-monocyte ratio （LMR）,in patients with esophageal squamous cell carcinoma （ESCC） by curative esophagectomy.Methods A total of 117 patients with ESCC from January 2010 to December 2012 in Affiliated Taixing People＇s Hospital of Yangzhou University were retrospectively analyzed.They were treated with standard curative esophagectomy.These patients were divided into NLR≥2.8 group and NLR 〈2.8 group,PLR≥127.3 group and PLR 〈127.3 group,LMR≥3.8 group and LMR 〈3.8 group for comparing the patients＇ general survival conditions and analyzing the influence on the progression-free survival （PFS） and overall survival （OS） rates according to the median values 2.8,127.3,3.8 of NLR,PLR and LMR,respectively.The COX proportional hazards models of NLR,PLR and LMR were used to carry out univariate and multivariate analyses for PFS and OS.The evaluation of prognostic values of NLR,PLR and LMR were carried by receiver operating characteristic （ROC） curve.Results For 117 patients,the median PFS time was 17 months,and the PFS rates at the 1-,3-and 5-year period were 66.7%,21.4% and 17.9%,respectively;the median OS time was 36 months,and the OS rates at the 1-,3-and 5-year time were 94.9%,46.2% and 28.2%,separately.In addition,a close relationship was identified between high NLR,high PLR,low LMR and tumor relapse （all P 〈0.05）.Furthermore,in the NLR 〈2.8 group,the median PFS time was 24 months （95% CI：19.788-28.212）,and the 1-,3-,5-year PFS rates were 78.9%,35.1% and 31.6% separately,while in the NLR≥2.8 group,the median PFS time was 13 months （95%CI：10.153-15.847）,and the 1-,3-,5-year PFS rates were 55.0%,8.3% and 5.0%,respectively （x2 =15.601,P 〈 0.001）.In the PLR 〈 127.3 group,the median PFS time was 24 months （95% CI：19.891-28.109）,and the 1-,3-,5-year PFS rates were 78.0%,30.5% and 27.1%.In the PLR≥ 127.3 group,the median PFS time was 15 months （95%CI：11.832-18.168）,and the 1-,3-,5-year PFS rates were 55.2%,12.1% and 8.6% （x2 =7.621,P =0.006）.In the LMR 〈3.8 group,the median PFS time was 14 months （95% CI：11.534-16.466）,and the 1-,3-,5-year PFS rates were 57.9%,8.8% and 5.3%,whilein the LMR≥3.8 group,the median PFS time was 21 months （95% CI：16.783-25.217）,and the 1-,3-,5-year PFS rates were 75.0%,33.3% and 30.0% （x2 =10.201,P =0.001）.Correspondingly,the median OS time was 42 months （95% CI：29.188-48.282） and the 1-,3-,5-year OS rates were 98.2%,56.1% and 47.4% in the NLR 〈2.8 group.While the median OS time was 27 months （95% CI：20.358-33.642） and the 1-,3-,5-year OS rates were 91.7%,36.7% and 10.0% in the NLR ≥2.8 group （x2 =19.161,P 〈 0.001）.Themedian OS time was 38 months （95% CI：31.310-44.690） and the 1-,3-,5-year OS rates were 94.9%,54.2% and 37.3 % in the PLR 〈 127.3 group and the median OS time was 27 months （95 % CI：19.537-34.463） and the 1-,3-,5-year OS rates were 93.1%,37.9% and 19.6% in the PLR≥127.3 group （x2 =7.019,P =0.008）.The median OS time was 30 months （95% CI：23.659-36.341） and the 1-,3-,5-year OS rates were 91.2%,36.8% and 12.3% in the LMR 〈 3.8 group.While the median OS time was 38 months （95% CI：27.878-48.121） and the 1-,3-,5-year OS rates were 95.0%,55.3% and 43.3% in the LMR≥3.8 group （x2 =10.201,P=0.001）.In univariate analysis,the following factors were significantly associated with poor PFS：T stage （HR =1.292,95% CI：1.077-2.211,P =0.048）,N stage（HR =1.773,95% CI：1.186-2.651,P =0.005）,TNM stage （HR =1.768,95 % CI：1.181-2.645,P =0.006）,NLR （HR =2.193,95 % CI：1.450-3.316,P〈0.001）,PLR（HR =1.722,95%CI：1.149-2.581,P =0.009） and LMR （HR =0.531,95%CI：0.353-0.799,P =0.002）.The univariate analysis further revealed that T stage （HR =1.982,95% CI：1.162-3.383,P=0.012）,N stage （HR =1.910,95% CI：1.243-2.934,P =0.003）,TNM stage （HR =2.115,95% CI：1.375-3.252,P =0.001）,NLR （HR =2.599,95% CI：1.657-4.078,P 〈 0.001）,PLR （HR =1.764,95%CI：1.145-2.717,P =0.010） and LMR （HR =0.470,95% CI：0.303-0.728,P =0.001） were also significantly associated with poor OS.Furthermore,multivariate COX regression analysis showed that TNM stage （HR=1.608,95%CI：1.057-2.445,P =0.026） and NLR （HR =1.886,95%CI：1.133-3.138,P=0.015） were independent prognostic factors for PFS in patients with ESCC after surgery.Correspondingly,TNM stage （HR =1.867,95 % CI：1.190-2.928,P =0.007） and NLR （HR =2.226,95 % CI：1.292-3.835,P =0.004） were also independent prognostic factors for OS in ESCC patients following surgery.Finally,ROC curves of NLR,PLR and LMR for PFS predictive values were as follows：the area under the curve （AUC） for NLR,PLR and LMR were 0.725 （95% CI：0.615-0.835,P =0.001）,0.657 （95% CI：0.533-0.781,P =0.025） and 0.290 （95% CI：0.178-0.402,P =0.003）,respectively.ROC curve analysis of NLR,PLR and LMR in diagnostic value of OS indicated that the AUC was 0.731 （95% CI：0.632-0.829,P 〈 0.001） for NLR,0.613 （95% CI：0.501-0.726,P =0.057） for PLR and 0.308 （95% CI：0.205-0.412,P =0.053） for LMR.Conclusion NLR is superior to PLR.and LMR in predicting the survival outcome of patients with ESCC,and NLR is of great value in predicting the survival and prognosis of patients with thoracic ESCC after operation.