Wang Xiaojie, Chi Pan, Lin Huiming, Lu Xingrong, Huang Ying, Xu Zongbin, Huang Shenghui, Sun Yanwu, Ye Daoxiong
Department of Colorectal Surgery, Union Hospital, Fujian Medical University, Fuzhou 350001, China.
Zhonghua Wei Chang Wai Ke Za Zhi. 2017 Jun 25;20(6):654-659.
To establish a nomogram to predict long-term survival in non-metastatic colorectal cancer patients.
A retrospective analysis was conducted in patients with non-metastatic colorectal cancer who underwent radical surgery in the Department of Colorectal Surgery of Affiliated Union Hospital of Fujian Medical University between January 2000 and December 2014. Univariate and multivariate analyses on disease-free survival (DFS) were performed using the Cox proportional regression model. Based on the multivariate analysis results, a prognostic nomogram was formulated to predict the probability for DFS. Concordance index was applied in predictive evaluation of the nomogram and calibration curves were drawn to test the nomogram's prediction and actual observation of the 5-year DFS rate. The predictive ability of nomogram was compared with AJCC-7 staging system.
A total of 2 641 patients were identified. The median age was 59.3 years old, and 60.3% of cases were men. The number of patients with TNM stage 0, I(, II( and III( was 96, 505, 923 and 1043, respectively. The most common tumor site was the rectum, accounting for 43.2%. A total of 413 (15.6%) patients underwent neoadjuvant treatment. The most common gross type of tumor was ulcerative type, accounting for 79.5%. The 3- and 5-year DFS rate was 85.8% and 79.8%, respectively. Based on the Cox proportional regression model, the following six factors were independently associated with reduced DFS rate and were selected for the nomogram: older age, higher pathologic T stage, higher pathologic N stage, higher preoperative serum CEA level, infiltrative gross type and perineural invasion. The results of the nomogram showed that the score of T0, T1, T2, T3 and T4 stage was 0, 2.2, 3.9, 4.1 and 6, respectively, and the score of N0, N1 and N2 was 0, 3.8 and 9.3, respectively. For gross type, the score of expanding type, ulcerative type and infiltrative type was 6, 9 and 10, respectively. The score of perineural invasion was 5.2. Higher scores were added to older age and higher CEA level. The total scores were calculated by taking the sum of the points from all predictors. Higher total score was associated with poor DFS. The prognostic nomogram differentiated well and showed a concordance index of 0.718, which was better than AJCC-7 staging system (concordance index=0.683). Also, the calibration of nomogram predictions was good.
A nomogram based on 6 independently prognostic factors to predict long-term survival in non-metastatic colorectal cancer patients is established successfully. The nomogram can be conveniently used to facilitate the accurate individualized prediction of DFS rates in patients with non-metastatic colorectal cancer.
建立一种列线图以预测非转移性结直肠癌患者的长期生存情况。
对2000年1月至2014年12月在福建医科大学附属协和医院结直肠外科接受根治性手术的非转移性结直肠癌患者进行回顾性分析。采用Cox比例回归模型对无病生存期(DFS)进行单因素和多因素分析。基于多因素分析结果,制定预后列线图以预测DFS概率。应用一致性指数对列线图进行预测评估,并绘制校准曲线以检验列线图对5年DFS率的预测与实际观察情况。将列线图的预测能力与AJCC-7分期系统进行比较。
共纳入2641例患者。中位年龄为59.3岁,60.3%为男性。TNM分期0、I、II和III期的患者数量分别为96、505、923和1043例。最常见的肿瘤部位是直肠,占43.2%。共有413例(15.6%)患者接受了新辅助治疗。最常见的肿瘤大体类型是溃疡型,占79.5%。3年和5年DFS率分别为85.8%和79.8%。基于Cox比例回归模型,以下六个因素与DFS率降低独立相关,并被选入列线图:年龄较大、病理T分期较高、病理N分期较高、术前血清CEA水平较高、浸润性大体类型和神经周围侵犯。列线图结果显示,T0、T1、T2、T3和T4期的评分分别为0、2.2、3.9、4.1和6,N0、N1和N2的评分分别为0、3.8和9.3。对于大体类型,膨胀型、溃疡型和浸润型的评分分别为6、9和10。神经周围侵犯的评分为5.2。年龄较大和CEA水平较高者增加较高分数。通过将所有预测因素的得分相加计算总分。总分较高与DFS较差相关。预后列线图区分良好,一致性指数为0.718,优于AJCC-7分期系统(一致性指数=0.683)。此外,列线图预测的校准良好。
成功建立了一种基于6个独立预后因素的列线图,用于预测非转移性结直肠癌患者的长期生存情况。该列线图可方便地用于促进对非转移性结直肠癌患者DFS率的准确个体化预测。