乳头状肾细胞癌患者术后2年无复发生存率的个体化预测

Individualized prediction of recurrence-free survival after operation in patients with papillary renal cell carcinoma

目的探讨乳头状肾细胞癌(PRCC)患者术后无复发生存率(RFS)的独立预测因素,并建立诺模图对其术后2年RFS进行个体化预测。方法回顾性分析2009年6月至2018年5月空军军医大学西京医院诊治的71例PRCC患者的临床资料。男51例,女20例;中位年龄54岁(25~83岁);左侧肿瘤36例,右侧肿瘤35例;临床分期Ⅰ、Ⅱ、Ⅲ、Ⅳ期患者分别为47、12、9、3例;Ⅰ型PRCC27例,Ⅱ型PRCC44例;15例(21.1%)发生肿瘤坏死;术前碱性磷酸酶(ALP)中位值89.0U/L(43.0~217.0U/L);术前血红蛋白(132.5±19.9)g/L,其中20例(28.2%)<120g/L;术前中性粒细胞/淋巴细胞比值(NLR)中位值2.40(1.03~6.77)。71例中43例行根治性肾切除术,28例行保留肾单位手术。采用Kaplan-Meier法计算术后1、2、3年RFS,采用Cox单因素及多因素回归分析确定术后RFS的独立预测因素,使用R软件整合所有具有独立预测意义的变量绘制诺模图,并采用Bootstrap法计算C-index、绘制校准图对诺模图的预测性能进行内部验证。结果71例中位随访41个月(5~119个月),16例(22.5%)复发,其中4例为局部复发,12例为远处复发;11例(68.8%)于术后2年内复发。术后1、2、3年RFS分别为88.6%、80.4%和78.7%。临床分期Ⅱ期(HR=3.655,95%CI1.036~12.890,P=0.044)、≥Ⅲ期(HR=3.654,95%CI1.008~13.248,P=0.049)及术前ALP≥90U/L(HR=3.274,95%CI1.044~10.267,P=0.042)、血红蛋白<120g/L(HR=4.771,95%CI1.553~14.660,P=0.006)、NLR≥2.40(HR=4.701,95%CI1.238~17.849,P=0.023)是术后RFS的独立预测因素。经内部验证,整合临床分期、术前ALP、术前血红蛋白和术前NLR4个预测变量的诺模图预测性能良好(C-index=0.829,95%CI0.819~0.839)。结论本研究基于单中心数据确定了临床分期≥Ⅱ期及术前ALP≥90U/L、血红蛋白<120g/L、NLR≥2.40是PRCC患者术后RFS的独立预测因素;建立的术前即可对PRCC患者术后2年RFS进行个体化预测的诺模图,有助于泌尿外科医生制订更为系统的治疗方案。

Objective To explore the independent predictors for recurrence-free survival (RFS) in patients with papillary renal cell carcinoma (PRCC), and establish a nomogram to predict individual 2-year RFS. Methods The data of patients diagnosed in Xijing Hospital of the Air Force Medical University from June 2009 to May 2018 were retrospectively analyzed. A total of 71 patients with primary T1-4N0M0 unilateral tumor were included in this study. The median age was 54 (range from 25 to 83) years. There were 51 males and 20 females. There were 27 cases of typeⅠ, 44 cases of type Ⅱ, 36 cases of left tumor and 35 cases of right tumor. Patients with clinical Ⅰ,Ⅱ,Ⅲ,Ⅳ stage were 47 cases, 12 cases, 9 cases and 3 cases, respectively. Tumor necrosis occurred in 15 cases (21.1%). The average preoperative alkaline phosphatase (ALP) was 89.0(43.0-217.0) U/L, the average preoperative hemoglobin (Hb) was (132.5±19.9) g/L, with 20 cases (28.2%) of Hb<120 g/L before surgery;the average of preoperative Hb was (132.5±19.9) g/L, among which 20 cases (28.2%) had Hb<120 g/L before surgery;the average of preoperative neutrophils/lymphocytes (NLR) was 2.40 (1.03-6.77);radical nephrectomy was performed in 43 patients and nephrectomy was performed in 28 patients. The 1-, 2-, and 3-year RFS were calculated using Kaplan-Meier method and the differences among different subgroups were assessed using Log-rank test. Univariate and multivariate Cox regression analysis were used to identify the independent predictors for RFS, and the nomogram was developed using R software according to the results of multivariate Cox regression analysis. Furthermore, the predictive ability of the nomogram was internally validated using the Bootstrap method by calculating the C-index and drawing the calibration plot. Results After a median follow-up of 41 ( range from 25 to 83) months, 16 (22.5%) recurrence occurred, including 4 local recurrence and 12 distant recurrence, and 11 (68.8%) patients relapsed within 2 years after surgery. The 1, 2, and 3-year RFS were 88.6%, 80.4%, and 78.7%, respectively. Clinical stage Ⅱ(HR=3.655, 95%CI 1.036-12.890, P=0.044) and stage ≥Ⅲ(HR=3.654, 95%CI 1.008-13.248, P=0.049), preoperative ALP≥90U/L(HR=3.274, 95%CI 1.044-10.267, P=0.042), preoperative Hb<120 g/L (HR=4.771, 95%CI 1.553-14.660, P=0.006), and preoperative NLR≥2.40(HR=4.701, 95%CI 1.238-17.849, P=0.023) were identified as independent risk factors for RFS. On internal validation, the nomogram which integrates the four predictors of clinical stage, preoperative ALP, preoperative HB and preoperative NLR, has a good predictive performance (C-index =0.829, 95%CI 0.819-0.839). Conclusions In the present study based on single center data, clinical stage Ⅱ and ≥Ⅲ, preoperative ALP≥90 U/L, preoperative Hb<120 g/L and preoperative NLR ≥2.40 were independent predictors for postoperative RFS in patients with PRCC, and a new preoperative nomogram for predicting individual postoperative 2-year RFS was established, which would be helpful for urologists to develop more systematic treatment plans.