胃癌患者预后相关影响因素的列线图模型构建及验证
作者:
作者单位:

中国科学技术大学附属第一医院/安徽省立医院 腹部外科,安徽 合肥 230036

作者简介:

李吴寒,中国科学技术大学附属第一医院/安徽省立医院住院医师,主要从事胃肠道肿瘤方面的研究。

通信作者:

王高生,Email: wgs2018@126.com

基金项目:


Construction and validation of a nomogram for prognostic value of NLR and PLR in patients with gastric cancer
Author:
Affiliation:

Department of Abdomen Surgery, the First Affiliated Hospital of University of Science and Technology of China/Anhui Provincial Hospital, Hefei 230036, China

Fund Project:

  • 摘要
  • |
  • 图/表
  • |
  • 访问统计
  • |
  • 参考文献
  • |
  • 相似文献
  • |
  • 引证文献
  • |
  • 资源附件
  • |
  • 音频文件
  • |
  • 视频文件
    摘要:

    背景与目的 中国胃癌疾病负担较重且预后影响因素较多,有关量化和综合评估预后风险的研究较少。因此,本研究基于列线图探究炎症指标中性粒细胞/淋巴细胞比率(NLR)和血小板/淋巴细胞比率(PLR)对胃癌患者预后生存的意义,并将其纳入列线图与传统TNM分期进行预后评估效能比较。方法 回顾性纳入2013年6月—2018年6月在中国科学技术大学第一附属医院胃肠外科接受胃癌根治切除术的胃癌患者作为训练组(n=300),同时从胃肠外科另一病区纳入接受相同手术处理的胃癌患者作为验证组(n=100)。通过医院电子病历系统采集患者的年龄、性别、肿瘤类型、肿瘤部位、侵袭深度和淋巴结转移(LNM)等信息;术前3 d收集外周静脉血数据,并计算NLR和PLR,通过ROC曲线确定NLR(1.98)和PLR(134.87)的最佳临界点。术后2年内每3个月随访1次,2年后每6个月随访1次。采用Cox比例风险模型计算暴露与结局指标的关联,根据多因素分析结果识别影响胃癌预后的独立风险因素,纳入列线图后通过C-指数在训练组和验证组评估列线图的稳定性。最后,基于ROC曲线下面积(AUC)比较列线图和传统TNM分期的预测效能。结果 训练组男性患者220例(73.3%),验证组男性患者69例(69.0%),训练组平均年龄(62.52±10.61)岁,验证组平均年龄(63.67±10.21)岁。两组除肿瘤类型、分化程度和侵袭深度外,其他基线特征差异无统计学意义;训练组中位生存时间(OS)为28个月,1、3、5年OS率分别为63.5%、43.0%和35.1%;验证组中位OS为32个月,1、3、5年OS率分别为58.9%、41.6%和31.7%。单因素Cox回归分析显示,年龄、病理分型、肿瘤分化程度、侵袭深度、存在LNM、NLR、PLR和CEA水平均与OS有关(均P<0.05)。经过多因素调整后,存在LNM、术前NLR>1.98、PLR>134.87和癌胚抗原(CEA)≥5 μg/L的患者OS显著缩短(均P<0.01)。校准曲线结果显示列线图模型在训练组(C-指数=0.81)和验证组(C-指数=0.75)的拟合度良好。此外,列线图模型预测训练组1、3、5年OS率的AUC值(0.865,0.855,0.827)高于TNM分期(0.677,0.690,0.683);验证组1、3、5年OS率的AUC值(0.856,0.788,0.725)高于TNM分期(0.781,0.691,0.605)。结论 NLR和PLR是预测胃癌患者术后生存的独立风险因素,基于两者构建的列线图可以较为准确地预测行胃切除术胃癌患者的1、3、5年OS率,为临床医师提供更精确的治疗、护理决策证据。

    Abstract:

    Background and Aims The disease burden of gastric cancer in China is high and there are many prognostic factors. There are few studies on the quantitative and comprehensive assessment of prognostic risk. Therefore, this study explored the significance of inflammatory indicators neutrophil/lymphocyte ratio (NLR) and platelet/lymphocyte ratio (PLR) on the prognosis and survival of gastric cancer patients based on nomogram and included them in nomogram and traditional TNM staging to compare the prognostic evaluation efficacy.Methods A retrospective study was conducted in the Department of Gastrointestinal Surgery of the First Affiliated Hospital of University of Science and Technology of China from June 2013 to June 2018. Gastric cancer patients who underwent radical gastrectomy were included in the training group (n=300). Patients with the same diagnosis who experienced the same surgical treatment from another ward were included as the validation group (n=100). The patient's age, gender, tumor type, tumor site, invasion depth, and lymph node metastasis (LNM) were collected through the hospital's electronic medical record system. Peripheral venous blood data were collected 3 days before the operation, and NLR and PLR were calculated. The ROC curve determined the optimal critical points of NLR (1.98) and PLR (134.87). The patients were followed up every 3 months within 2 years and every 6 months after 2 years. Cox proportional hazards model was used to calculate the association between exposure and outcome indicators, and the independent risk factors affecting the prognosis of gastric cancer were identified according to the results of multivariate analysis. The stability of the nomogram was evaluated by C-index in the training group and the validation group after inclusion in the nomogram. Finally, the prediction performance of nomogram and traditional TNM staging was compared based on the area under the ROC curve (AUC).Results There were 220 male patients (73.3%) in the training group and 69 male patients (69.0%) in the validation group. The average age of the training and validation groups was (62.52±10.61) years and (63.67±10.21) years, respectively. There was no significant difference in other baseline characteristics between the two groups except tumor type, differentiation degree and invasion depth. The training group's median overall survival (OS) was 28 months, and the 1-year, 3-year and 5-year OS rates were 63.9%, 43.1% and 35.1%, respectively. The median OS in the validation group was 32 months, and the 1-year, 3-year and 5-year OS rates were 58.9%, 41.6% and 31.7%, respectively. Univariate Cox regression analysis showed that age, pathological type, degree of tumor differentiation, depth of invasion, LNM, NLR, PLR and CEA levels were all associated with OS (all P<0.05). After multivariate adjustment, patients with LNM, preoperative NLR>1.98, PLR >134.87 and carcinoembryonic antigen (CEA) ≥5 μg/L had significantly shorter OS (all P<0.01). The calibration curve results showed that the nomogram model fits well in the training group (C-index=0.81) and the validation group (C-index=0.75). In addition, the AUC values of the nomogram model in predicting the 1-year, 3-year, and 5-year OS rates of the training group (0.865, 0.855, 0.827) were higher than those of the TNM stage (0.677, 0.690, 0.683). The AUC values of 1-year, 3-year, and 5-year OS rates in the training group (0.856, 0.788, 0.725) were higher than those of the TNM stage (0.781, 0.691, 0.605).Conclusion NLR and PLR are independent risk factors for predicting the survival of patients with gastric cancer. The constructed nomogram could more accurately predict the 1-, 3-, and 5-year OS rates of gastric cancer patients undergoing gastrectomy and provide clinicians with more accurate treatment and nursing decision-making evidence.

    表 2 训练组单因素及多因素Cox回归分析Table 2 Univariate and multivariate Cox regression analysis in training group
    表 3 列线图模型与TNM分期的ROC预测效果(AUC)比较Table 3 Comparison of AUC between nomogram model and TNM staging
    图1 基于训练组Cox分析的列线图预测模型Fig.1 Predicted nomogram based on multivariate Cox regression of training group
    图2 训练组与验证组预测1、3、5年OS率的校准曲线 A-C: 训练组;D-F: 验证组Fig.2 Calibration curves of 1-, 3-, and 5-year OS rate in the training group and the validation group A-C: training group; D-F: validation group
    图1 基于训练组Cox分析的列线图预测模型Fig.1 Predicted nomogram based on multivariate Cox regression of training group
    图2 训练组与验证组预测1、3、5年OS率的校准曲线 A-C: 训练组;D-F: 验证组Fig.2 Calibration curves of 1-, 3-, and 5-year OS rate in the training group and the validation group A-C: training group; D-F: validation group
    表 1 训练组与验证组胃癌患者的临床特征比较[n(%)]Table 1 Comparison of clinical characteristics between training group and validation group [n (%)]
    参考文献
    相似文献
    引证文献
引用本文

李吴寒,张营,潘晶晶,王高生.胃癌患者预后相关影响因素的列线图模型构建及验证[J].中国普通外科杂志,2022,31(10):1381-1388.
DOI:10.7659/j. issn.1005-6947.2022.10.014

复制
分享
文章指标
  • 点击次数:
  • 下载次数:
历史
  • 收稿日期:2021-09-09
  • 最后修改日期:2022-04-06
  • 录用日期:
  • 在线发布日期: 2022-10-31