DTC术后患者^(131)I治疗与男性预后关系的倾向得分匹配研究

Propensity score matching study on the relationship between ^(131)I treatment and male prognosis in postoperative differentiated thyroid cancer patients

目的:通过倾向得分匹配(PSM)研究男性是否是分化型甲状腺癌(DTC)术后患者^(131)I治疗预后的危险因素。方法:回顾性分析2016年4月至2021年1月间于天津医科大学总医院接受全甲状腺切除术并进行^(131)I治疗的1677例DTC患者[年龄11~84(43.9±12.5)岁],分为男性组(546例)和女性组(1131例),采用PSM对所有数据进行处理,以减少数据偏差及混杂变量的影响。将患者评估结果分为疗效满意(ER)、疗效不确切(IDR)、生化疗效不佳(BIR)及结构性疗效不佳(SIR),其中ER、IDR被分为预后良好组,BIR、SIR被分为预后欠佳组。2组间差异的比较采用χ^(2)检验,采用多因素logistic回归分析影响DTC术后患者131I治疗预后的危险因素,采用ROC曲线分析刺激性甲状腺球蛋白(sTg)水平与预后欠佳的关系。结果:PSM前,男性患者预后欠佳的比例明显高于女性患者[21.2%(116/546)和14.0%(158/1131);χ^(2)=17.53,P=0.001];PSM后,2组间预后欠佳比例的差异无统计学意义[19.9%(107/537)和15.6%(84/537);χ^(2)=5.43,P=0.143]。多因素logistic回归分析显示,PSM前,男性[比值比(OR)=1.439(95%CI:1.016~2.038),P=0.040]、高T期(T3+T4期)[OR=1.816(95%CI:1.273~2.590),P=0.001]、N1b期[OR=1.766(95%CI:1.233~2.530),P=0.002]、M1期[OR=9.833(95%CI:3.190~30.309),P<0.001]及sTg水平[OR=1.035(95%CI:1.029~1.042),P<0.001]是预后欠佳的危险因素;PSM后,高T期(T3+T4期)[OR=1.870(95%CI:1.212~2.886),P=0.005]、M1期[OR=8.993(95%CI:2.434~33.225),P=0.001]、sTg水平[OR=1.040(95%CI:1.030~1.049),P<0.001]依旧是预后欠佳的危险因素,而N1b期[OR=1.459(95%CI:0.938~2.270),P=0.094]和男性[OR=1.383(95%CI:0.912~2.096),P=0.127]不再是预后欠佳的危险因素。ROC曲线分析示,sTg最佳界值为10.25μg/L,预测灵敏度和特异性分别为81.0%(222/274)和84.2%(1181/1403)。结论:经过PSM减少选择偏差后,男性不再是DTC术后131I治疗预后的危险因素。高T期(T3+T4期)、M1期及sTg≥10.25μg/L是DTC术后131I治疗预后欠佳的危险因素。

Objective To study whether male was the risk factor for prognosis of patients with differentiated thyroid cancer(DTC)after ^(131)I treatment based on propensity score matching(PSM)method.Methods From April 2016 to January 2021,1677 patients(age:11-84(43.9±12.5)years)with DTC who underwent total thyroidectomy and received 131I treatment in Tianjin Medical University General Hospital were retrospectively enrolled and patients were divided into male group(n=546)and female group(n=1131).The evaluation results of patients were divided into excellent response(ER),indeterminate response(IDR),biochemical incomplete response(BIR)and structural incomplete response(SIR).Among them,ER and IDR were divided into good prognosis group,and BIR and SIR were divided into poor prognosis group.The PSM method was adopted to process all data to reduce the influence of data bias and confounding variables.χ^(2) test was used for data analysis.Multivariate logistic regression was used to analyze the risk factors affecting prognosis,and ROC curve was used to analyze the relationship between stimulated thyroglobulin(sTg)level and poor prognosis.Results Before PSM,the proportion of male patients with poor prognosis was significantly higher than that of female patients(21.2%(116/546)vs 14.0%(158/1131);χ^(2)=17.53,P=0.001).After PSM,there was no difference in the proportion of poor prognosis between male and female groups(19.9%(107/537)vs 15.6%(84/537);χ^(2)=5.43,P=0.143).Multivariate logistic regression analysis showed that male(odds radio(OR)=1.439(95%CI:1.016-2.038),P=0.040),high T stage(T3+T4 stage)(OR=1.816(95%CI:1.273-2.590),P=0.001),N1b stage(OR=1.766(95%CI:1.233-2.530),P=0.002),M1 stage(OR=9.833(95%CI:3.190-30.309),P<0.001)and sTg level(OR=1.035(95%CI:1.029-1.042),P<0.001)were risk factors for poor prognosis before PSM,while high T stage(T3+T4 stage)(OR=1.870(95%CI:1.212-2.886),P=0.005),M1 stage(OR=8.993(95%CI:2.434-33.225),P=0.001),sTg level(OR=1.040(95%CI:1.030-1.049),P<0.001)were still risk factors,and N1b stage(OR=1.459(95%CI:0.938-2.270),P=0.094),male(OR=1.383(95%CI:0.912-2.096),P=0.127)were no longer risk factors for poor prognosis after PSM.ROC curve analysis showed that the cut-off value of sTg was 10.25μg/L,with the sensitivity of 81.0%(222/274)and the specificity of 84.2%(1181/1403).Conclusions After reduction of selection bias by PSM,male is no longer a risk factor for prognosis after 131I treatment of DTC.In addition,high T stage(T3+T4 stage),M1 stage and sTg≥10.25μg/L were risk factors for poor prognosis.