结直肠癌术后肝转移的危险因素分析及预测模型构建

Analysis of risk factors for postoperative liver metastasis in colorectal cancer andconstruction of a predictive model

    摘要
  • 摘要:
      目的  分析结直肠癌术后肝转移的危险因素并构建预测模型。
      方法  选取2011年1月至2015年12月于苏州市第九人民医院行根治性手术切除的247例结直肠癌患者作为观察对象，依据随访是否出现肝转移分为转移组、非转移组。对两组各项指标进行单因素筛选，再行Logistic多因素回归分析，并建立预测模型，应用受试者工作特征（receiver operating characteristic，ROC）曲线检测其区分度，应用拟合优度检验评价其校准度。另选本院102例结直肠癌患者进行该模型的临床验证。
      结果  247例结直肠癌患者术后随访中出现肝转移81例（转移组），占32.79%，其中分化程度、淋巴结转移、血清癌胚抗原（carcinoembryonic antigen，CEA）、D-二聚体、血清白蛋白（serum albumin，Alb）、可溶性细胞间黏附分子-1（soluble intercellular adhesion molecule-1，sICAM-1）、miRNA-203表达量均是结直肠癌术后肝转移的危险因素（P<0.05）。根据危险因素得出预测模型表达方程为：Logit（P）＝1.203×分化程度+2.065×淋巴结转移+1.257×CEA+0.829×D-二聚体+0.395×Alb+1.156×sICAM-1+1.484×miRNA-203表达量-34.187。该模型ROC曲线下的面积为0.895（95% CI：0.830～0.959），最大约登指数（0.722）对应的灵敏度为0.896，特异度为0.904。经拟合优度检验，χ2=3.014，P=0.421。临床验证模型预测的灵敏度为86.21%，特异度为89.04%，准确率为88.24%。
      结论  结直肠癌术后肝转移的危险因素分别为分化程度低、淋巴结转移、CEA高表达、D-二聚体高表达、Alb降低、sICAM-1升高、低表达miRNA-203。以此构建预测模型，区分能力良好，可以有效评估结直肠癌术后肝转移风险。

     

    Abstract:
      Objective  To analyze the risk factors for liver metastasis after colorectal cancer surgery and construct a predictive model.
      Methods  In total, 247 colorectal cancer who underwent radical surgical resection at Suzhou Ninth People's Hospital from January 2011 to December 2015 were selected for postoperative observation. These patients were assigned into groups in which metastases did or did not appear. One-way ANOVA of prognostic indicators among both groups was performed. Logistic multiple factor regression analysis was employed, and a predictive model was constructed. Receiver operating characteristic (ROC) curves were analyzed to assess the discrimination of the model, and a goodness of fit test was used to evaluate its calibration. Another 102 colorectal cancer patients in this hospital were selected for clinical validation of the model.
      Results  During follow-up of 247 colorectal cancer patients, 81 cases of liver metastasis (metastasis group) occurred (32.79%). Among these, the degree of differentiation, lymph node metastasis, and expression of serum carcinoembryonic antigen (CEA), D-dimer, albumin (Alb), soluble intercellular adhesion molecule-1 (sICAM-1), and miRNA-203 are all risk factors in colorectal cancer patients for postoperative liver metastasis (P<0.05). According to these risk factors, the equation expressing our predictive model is: Logit (P)=1.203×differentiation degree + 2.065 × lymph node metastasis + 1.257 × CEA + 0.829 × D-dimer + 0.395 × Alb + 1.156 × sICAM-1 + 1.484× miRNA-203 expression level-34.187. The area under the ROC curve for the model was 0.895 (95% CI: 0.830～0.959). The maximum Youden index (0.722) corresponded to a sensitivity of 0.896 and a specificity of 0.904. Goodness of fit test yielded χ2=3.014 and P=0.421. The prediction sensitivity of the clinical validation model was 86.21%. The specificity and accuracy were 89.04% and 88.24%, respectively.
      Conclusions  The risk factors for postoperative liver metastasis in colorectal cancer include reduced differentiation, lymph node metastasis, high CEA expression, high D-dimer expression, decreased Alb level, increased sICAM-1 level, and low degree of miRNA-203 expression. These findings aided in the construction of a predictive model with good discrimination ability that can effectively assess the risk of liver metastasis after colorectal cancer surgery.

     

    • HTML全文
    • 参考文献(25)
    • 相关文章
    • 施引文献
  • 资源附件(0)

/

返回文章
返回