子宫内膜癌淋巴结转移预测模型的初步建立

发布时间：2018-06-13 12:36

本文选题：子宫内膜癌 + 淋巴结转移　；参考：《上海交通大学》2015年博士论文

【摘要】：目的:探讨临床病理参数与子宫内膜癌淋巴结转移的关系;分析联合多个肿瘤免疫组化标志物在内膜癌淋巴结转移中的诊断价值;建立并评估用于个体化预测内膜癌淋巴结转移风险的列线图模型。方法:回顾性分析从2008年1月1日至2014年12月31日在上海市交通大学附属第六人民医院收治的子宫内膜癌患者的临床病理资料,收集符合纳入条件的分期手术病人182例。详细记录患者术前血清CA125、组织学分级、肌层浸润深度、淋巴血管间隙浸润、淋巴结转移等临床病理资料以及术后石蜡切片组织免疫组化标志物ER、PR、p53、EGFR、c-erB-2、ki-67的表达情况。单因素分析临床病理参数与内膜癌淋巴结转移的相关性;通过多变量Logistic回归对上述六种标志物进行分析筛选,绘制ROC曲线,计算联合免疫组化标志物与各临床病理参数的ROC曲线下面积;建立预测内膜癌淋巴结转移的列线图模型,并对模型的预测准确性及符合度进行内部验证和校正。结果:182例子宫内膜癌患者中,156例行盆腔淋巴结切除术,26例同时行盆腔和腹主动脉旁淋巴结切除术。术后病理结果证实有盆腔淋巴结转移14例(占病例总数7.7%),同时合并腹主动脉旁淋巴结转移4例(2.2%),无单独腹主动脉旁淋巴结转移。术前血清CA125在预测淋巴结转移的ROC曲线下面积为0.804(95%Cl:0.697-0.901),当临界值取40U/mL时,敏感度和特异度分别为83.6%和80.3%。单因素分析结果显示不同病理类型、不同组织学分级、伴或不伴淋巴血管间隙浸润、不同肌层浸润深度之间淋巴结转移率不同,差异均具有统计学意义(P均0.05)。PR、p53、EGFR、c-erB-2、ki-67的蛋白表达在淋巴结阴性和阳性患者之间的差异也具有统计学意义(P均0.05)。Logistic多重回归分析筛选EGFR、c-erB-2、ki-67蛋白表达是淋巴结转移的危险因素,回归系数分别为2.086、1.853和1.337,绘制联合上述标志物的相应ROC曲线,曲线下面积为0.897(95%Cl:0.809-0.933)均大于独立预测因子。成功构建用于预测内膜癌淋巴结转移风险的列线图模型,并采用Bootstrap自抽样方法对列线图模型进行内部验证后得到的C-index为0.743。结论:免疫组化标志物可用于预测子宫内膜癌淋巴结转移,联合预测价值高于各独立预测因子。成功构建基于内膜癌临床病理特征及免疫组化结果的列线图模型,有一定的预测准确性,具有临床应用价值。
[Abstract]:Objective: To investigate the relationship between the clinicopathological parameters and lymph node metastasis of endometrial carcinoma; to analyze the diagnostic value of combined multiple tumor immuno histochemical markers in lymph node metastasis of endometrial carcinoma; to establish and evaluate the line graph model for individualized prediction of the risk of lymph node metastasis in endometrial carcinoma. Methods: retrospective analysis from January 1, 2008 to 2014 The clinicopathological data of endometrial cancer patients in the Sixth People's Hospital Affiliated to Shanghai Jiaotong University in December 31st were collected and 182 cases were collected. The preoperative serum CA125, histological grade, depth of myometrium infiltration, infiltration of lymphatic space, lymph node metastasis and other clinicopathological data were recorded in detail. And the expression of tissue immuno histochemical markers ER, PR, p53, EGFR, c-erB-2, Ki-67 in paraffin section. The correlation between the clinicopathological parameters and lymph node metastasis of endometrial carcinoma was analyzed by single factor analysis. The six markers were screened by multivariable Logistic regression, the ROC curve was plotted, and the joint immuno histochemical markers were calculated and the various faces were calculated. The area under the ROC curve of the pathological parameters of the bed; the establishment of a line map model for predicting lymph node metastasis of endometrial carcinoma, and internal verification and correction of the predictive accuracy and coincidence of the model. Results: of the 182 patients with endometrial carcinoma, 156 cases were performed pelvic lymphadenectomy and 26 cases were performed simultaneously with pelvic and abdominal aortic dissection. Pathological findings confirmed 14 cases of pelvic lymph node metastasis (7.7% of the total cases), and 4 cases (2.2%) with paracal lymph node metastases (2.2%) and no solitary abdominal lymph node metastasis. The preoperative serum CA125 was 0.804 (95%Cl: 0.697-0.901) under the ROC curve of predicting lymph node metastasis. When the critical value was 40U/mL, sensitivity and specificity The results of 83.6% and 80.3%. single factor analysis showed that different pathological types, different histological grades, with or without infiltration of lymphatic space, the rate of lymph node metastasis was different between different myometrium infiltration depth, and the difference was statistically significant (P 0.05).PR, p53, EGFR, c-erB-2, Ki-67 protein expression in lymph node negative and positive patients The difference was also statistically significant (P 0.05).Logistic multiple regression analysis screening EGFR, c-erB-2, Ki-67 protein expression was a risk factor for lymph node metastasis, the regression coefficient was 2.086,1.853 and 1.337 respectively, and the corresponding ROC curve of the combined markers was plotted, and the area under the curve was 0.897 (95%Cl:0.809-0.933) larger than the independent prediction cause. A line map model was successfully constructed to predict the risk of lymph node metastasis in endometrial carcinoma, and the C-index was 0.743. conclusion after the internal verification of the line map model by Bootstrap self sampling. The immuno histochemical marker could be used to predict lymph node metastasis of endometrial carcinoma, and the combined predictive value was higher than that of the independent predictors. Based on the nomogram of clinicopathological features and immunohistochemical results of endometrial cancer, we have a certain predictive accuracy and clinical value.
【学位授予单位】：上海交通大学
【学位级别】：博士
【学位授予年份】：2015
【分类号】：R737.33

【相似文献】