切换至 "中华医学电子期刊资源库"

中华临床医师杂志(电子版) ›› 2023, Vol. 17 ›› Issue (11) : 1154 -1162. doi: 10.3877/cma.j.issn.1674-0785.2023.11.004

临床研究

结直肠癌患者周围神经侵犯预测模型的建立与评价
段福孝, 王鑫宇, 孙爽, 于知宇, 张成()   
  1. 110016 辽宁沈阳,北部战区总医院普通外科
  • 收稿日期:2023-07-25 出版日期:2023-11-15
  • 通信作者: 张成
  • 基金资助:
    辽宁省科学技术计划项目(2021JH2/10300106)

Establishment and evaluation of a predictive model for perineural invasion in colorectal cancer patients

Fuxiao Duan, Xinyu Wang, Shuang Sun, Zhiyu Yu, Cheng Zhang()   

  1. General Surgery Department, General Hospital of Northern Theater Command, Shenyang 110016, China
  • Received:2023-07-25 Published:2023-11-15
  • Corresponding author: Cheng Zhang
引用本文:

段福孝, 王鑫宇, 孙爽, 于知宇, 张成. 结直肠癌患者周围神经侵犯预测模型的建立与评价[J]. 中华临床医师杂志(电子版), 2023, 17(11): 1154-1162.

Fuxiao Duan, Xinyu Wang, Shuang Sun, Zhiyu Yu, Cheng Zhang. Establishment and evaluation of a predictive model for perineural invasion in colorectal cancer patients[J]. Chinese Journal of Clinicians(Electronic Edition), 2023, 17(11): 1154-1162.

目的

利用结直肠癌患者的临床信息及数据,通过XGBoost算法构建周围神经侵犯预测模型并评估其效能。

方法

选取我院178名结直肠癌患者真实世界临床数据并进行清洗及特征工程处理。训练及建立XGBoost模型,筛选最优特征子集并构建最终模型,利用SHAP 解释模型。应用Boruta算法筛选特征并构建线性判别分类器(linear discriminant analysis,LDA)、朴素贝叶斯(naive bayes)、KNN(K Nearest Neighbor,KNN)、分类与回归树(classification and regression trees,CART)、随机森林模型(random forest)五种模型,及过采样类平衡后再次构建模型。比较各模型间效能。

结果

XGBoost应用顺序特征选择器筛选包含9个特征数量的最优特征子集并构建最终模型,与其他机器学习模型相比,性能更加优越,训练速度更为快速。

结论

基于XGBoost算法成功构建了结直肠癌患者周围神经侵犯预测模型。

Objective

To construct a predictive model for perineural invasion using XGBoost based on clinical information and data of colorectal cancer patients and evaluate its effectiveness.

Methods

Real-world clinical data of 178 colorectal cancer patients in our hospital were selected and subjected to data cleaning and feature engineering processing. An XGBoost model was trained to select the optimal subset of features and construct the final model. The model was interpreted through SHapley Additive Explanation. The Boruta algorithm was applied to filter features and construct five models: Linear Discriminant Analysis, Naive Bayes, K Nearest Neighbor, Classification and Regression Trees, and Random Forest, and their oversampling class balance models were then generated. The performance of different models was compared.

Results

XGBoost applied Sequential Feature Selector to select the optimal feature subset containing 9 features and constructed the final model. Compared with other machine learning models, the performance of the XGBoost model was superior and the training speed was faster.

Conclusion

We have successfully constructed a predictive model for perineural invasion in colorectal cancer patients based on XGBoost. It can provide preoperative prediction of nerve invasion for clinical doctors, especially surgeons, and provide a basis for the development of comprehensive treatment plans.

表1 纳入患者一般临床信息[例(%)]
表2 候选22个临床指标中文、英文名称及缩写
图1 候选连续性临床指标变量相关性。蓝色表示正相关,红色表示负相关,颜色深度及圆形标识大小反映相关程度
图2 结直肠癌患者神经侵犯预测模型构建研究流程
表3 患者不同特征数量下AUC值
图3 患者不同特征数量下AUC值
图4 XGBoost模型ROC曲线
图5 最优特征子集XGBoost模型条形图。表示最优特征子集的9个临床特征的全局重要性排序及其贡献
图6 最优特征子集XGBoost模型Summary Plot图。越红表示该变量对预测影响越高,正负值代表对结局的正负影响
图7 最优特征子集XGBoost模型热力图。表示预测值f(x)与正负相关特征的关系
图8 最优特征子集XGBoost模型个案瀑布图。预测值f(x)如何根据其变量值作出预测过程,其中条带红色代表对结果产生正向作用,蓝色代表负向作用,而宽度代表作用强度
图9 最优特征子集XGBoost模型个案力图。预测值f(x)如何根据其变量值作出预测过程,其中条带红色代表对结果产生正向作用,蓝色代表负向作用,而宽度代表作用强度
图10 Boruta算法筛选特征初步结果。提示特征评估为重要特征(绿色)、不重要特征(红色)及待定特征(黄色)
图11 五种机器学习模型及过采样类平衡后模型箱线图及ROC曲线
表4 XGBoost与各机器学习模型性能比较
1
Sung H, Ferlay J, Siegel RL, et al. Global cancer statistics 2020: Globocan estimates of incidence and mortality worldwide for 36 cancers in 185 countries [J]. CA Cancer J Clin, 2021, 71(3): 209-249.
2
Zheng R, Zhang S, Zeng H, et al. Cancer incidence and mortality in China, 2016 [J]. J Natl Cancer Cent, 2022, 2(1): 1-9.
3
Duchalais E, Guilluy C, Nedellec S, et al. Colorectal cancer cells adhere to and migrate along the neurons of the enteric nervous system [J]. Cell Mol Gastroenterol Hepatol, 2017, 5(1): 31-49.
4
Liebig C, Ayala G, Wilks JA, et al. Perineural invasion in cancer: a review of the literature [J]. Cancer, 2009, 115(15): 3379-3391.
5
Cao Y, Deng S, Yan L, et al. Perineural invasion is associated with poor prognosis of colorectal cancer: a retrospective cohort study [J]. Int J Colorectal Dis, 2020, 35(6): 1067-1075.
6
李洪波, 袁静静, 张佃良, 等. 周围神经侵犯与结直肠癌患者预后的相关性分析 [J]. 中国现代普通外科进展, 2022, 25(6): 432-435.
7
韩冬, 魏颖, 王曦迪, 等. 周围神经侵犯与结直肠癌临床病理因素及预后的关系 [J]. 中华胃肠外科杂志, 2017, 20(1): 62-66.
8
Benson AB, Venook AP, Al-Hawary MM, et al. Colon cancer, version 2.2021, NCCN clinical practice guidelines in oncology [J]. J Natl Compr Canc Netw, 2021, 19(3): 329-359.
9
Ahmed Z, Mohamed K, Zeeshan S, et al. Artificial intelligence with multi-functional machine learning platform development for better healthcare and precision medicine [J]. Database (Oxford), 2020, 2020: baaa010.
10
Eloranta S, Boman M. Predictive models for clinical decision making: Deep dives in practical machine learning [J]. J Intern Med, 2022, 292(2): 278-295.
11
Lundberg SM, Lee SI. A unified approach to interpreting model predictions. Proceedings of the Advances in Neural Information Processing Systems [C]. Cambridge: MIT Press, 2017: 4765-4774.
12
Kursa M B, Rudnicki W R. Feature Selection with the Boruta Package [J]. J Stat Softw, 2010, 36(11): 1-13.
13
夏启俊, 陈凛. 胃癌神经侵袭的细胞与分子机制 [J]. 中华胃肠外科杂志, 2015(2): 198-200.
14
Dong H, Ying W, Xidi W, et al. Association of peripheral nerve invasion with clinicopathological factors and prognosis of colorectal cancer [J]. Chin J Gastrointest Surg, 2017, 20(1): 62-66.
15
王俊, 何平. 神经侵犯在新辅助放疗后ypⅠ~Ⅱ期直肠癌中的预后预测价值 [J]. 中国癌症杂志, 2022, 32(12): 1229-1234.
16
Kaya T, Dursun A. Can lymphovascular and perineural invasion be additional staging criteria in colorectal cancer? [J]. J Coll Physicians Surg Pak, 2021, 30(6): 657-662.
17
Qin L, Heng Y, Deng S, et al. Perineural invasion affects prognosis of patients undergoing colorectal cancer surgery: a propensity score matching analysis [J]. BMC Cancer, 2023, 23(1): 452.
18
Suzuki T, Suwa K, Ogawa M, et al. Adjuvant chemotherapy for the perineural invasion of colorectal cancer [J]. J Surg Res, 2015, 199(1): 84-89.
19
刘书豪, 侯新月, 张宪祥, 等. 进展期胃癌神经侵犯列线图预测模型的构建与验证 [J]. 中华胃肠外科杂志, 2020, 23(11): 1059-1066.
20
Chen T, Guestrin C. XGBoost: A scalable tree boosting system [C]//Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. 2016: 785-794.
21
Beck FK, Rosenthal TC. Prealbumin: a marker for nutritional evaluation [J]. Am Fam Physician, 2002, 65(8): 1575-1578.
22
Tu M-Y, Chien T-W, Chou M-T. Using a nutritional screening tool to evaluate the nutritional status of patients with colorectal cancer [J]. Nutr Cancer, 2012, 64(2): 323-330.
23
Lai CC, You JF, Yeh CY, et al. Low preoperative serum albumin in colon cancer: a risk factor for poor outcome [J]. Int J Colorectal Dis, 2011, 26(4): 473-481.
24
Boonpipattanapong T, Chewatanakornkul S. Preoperative carcinoembryonic antigen and albumin in predicting survival in patients with colon and rectal carcinomas [J]. J Clin Gastroenterol, 2006, 40(7): 592-595.
25
Borda F, Borda A, Jiménez J, et al. Valor predictivo de la hipoalbuminemia pre-tratamiento sobre el pronóstico del cáncer colorrectal resecado [Predictive value of pre-treatment hypoalbuminemia in prognosis of resected colorectal cancer] [J]. Gastroenterol Hepatol, 2014, 37(5): 289-295. Spanish.
26
沈小钢, 郭志义, 周晓刚. 血清前白蛋白与结肠癌患者预后的关系研究 [J]. 西南医科大学学报, 2017, 40(3): 292-295.
27
严征远, 张恒, 尹光平. 术前纤维蛋白原与白蛋白比值联合全身炎症反应指数对可切除胃癌患者预后的评估价值 [J/OL]. 中华普通外科学文献(电子版), 2020, 14(4): 266-270.
28
Yang M, Zhang Q, Ruan G-T, et al. Association Between Serum Creatinine Concentrations and Overall Survival in Patients With Colorectal Cancer: A Multi-Center Cohort Study[J]. Frontiers in Oncology, 2021, 11: 710423.
29
Brown J C, Caan B J, Meyerhardt J A, et al. The deterioration of muscle mass and radiodensity is prognostic of poor survival in stage I-III colorectal cancer: a population-based cohort study (C-SCANS)[J]. Journal of Cachexia, Sarcopenia and Muscle, 2018, 9(4): 664-672.
30
Armstrong L E, Muñoz C X, Armstrong E M. Distinguishing Low and High Water Consumers-A Paradigm of Disease Risk[J]. Nutrients, 2020, 12(3): 858.
31
Gao Y, Wang J, Zhou Y, et al. Evaluation of Serum CEA, CA19-9, CA72-4, CA125 and Ferritin as Diagnostic Markers and Factors of Clinical Parameters for Colorectal Cancer[J]. Scientific Reports, 2018, 8(1): 2732.
32
Tumay V, Guner O S. The utility and prognostic value of CA 19-9 and CEA serum markers in the long-term follow up of patients with colorectal cancer. A single-center experience over 13 years[J]. Annali Italiani Di Chirurgia, 2020, 91: 494-503.
33
王振勇, 孟宇, 李金超, 等. 胰腺癌术后局部复发特点及影响因素分析[J]. 中华普通外科杂志, 2022, 37(8): 592-596.
34
付蕾, 李潜, 卢创新, 等. 尿酸联合CA199检测对结直肠癌TNM分期和预后判断的价值[J]. 肿瘤, 2020, 40(7): 496-502.
35
黄利军, 方佳峰, 吴珏堃, 等. 术前血清肿瘤标志物联合血常规指标对结直肠癌患者预后评估的价值[J]. 中华胃肠外科杂志, 2018, 21(12): 1421-1426.
36
刘映华, 符国宏, 郑杨慈, 等. KLKB1联合CEA、CA199水平对老年结直肠癌的诊断价值[J]. 中国老年学杂志, 2023, 43(11): 2601-2604.
37
郭杰, 刘海东, 韦琴, 等. 基于检验大数据的结直肠癌风险预测模型建立与验证[J]. 中华检验医学杂志, 2021, 44(10): 914-920.
38
郝俊, 王帅, 朱军, 等. 神经网络预测模型辅助诊断结直肠癌微卫星状态的研究[J]. 中国普通外科杂志, 2023, 32(4): 488-496.
39
罗治文, 陈晓, 张业繁, 等. 同时性结直肠癌肝转移术后早期复发的影响因素[J]. 中华肝胆外科杂志, 2020, 26(10): 741-747.
[1] 范帅华, 郭伟, 郭军. 基于机器学习的决策树算法在血流感染预后预测中应用现状及展望[J]. 中华实验和临床感染病杂志(电子版), 2023, 17(05): 289-293.
[2] 葛云鹏, 崔红元, 宋京海. 人工智能在原发性肝癌诊断、治疗及预后中的应用[J]. 中华肝脏外科手术学电子杂志, 2023, 12(04): 367-371.
[3] 陈润芝, 杨东梅, 徐慧婷. 信迪利单抗联合索凡替尼后线治疗MSS型BRAF突变的转移性结肠癌:个案报道并文献复习[J]. 中华结直肠疾病电子杂志, 2023, 12(05): 431-435.
[4] 鄂一民, 孙司正, 范小彧, 喻春钊. 结直肠癌粪便筛查的现状与展望[J]. 中华结直肠疾病电子杂志, 2023, 12(04): 331-336.
[5] 陈显金, 吴芹芹, 何长春, 张庆华. 利用多模态医学数据和机器学习构建脑出血预后预测模型的研究[J]. 中华脑科疾病与康复杂志(电子版), 2023, 13(04): 193-198.
[6] 梁文龙, 曹杰, 黄庆, 林泳, 黄红丽, 杨平, 李冠炜, 胡鹤. 信迪利单抗联合瑞戈非尼治疗晚期结直肠癌的疗效与安全性分析[J]. 中华消化病与影像杂志(电子版), 2023, 13(06): 409-413.
[7] 王小娜, 谭微, 李悦, 姜文艳. 预测性护理对结直肠癌根治术患者围手术期生活质量、情绪及并发症的影响[J]. 中华消化病与影像杂志(电子版), 2023, 13(06): 525-529.
[8] 刘代江, 蒋俊艳, 万晓强, 马莎英. 结直肠癌肝转移患者生存状况及预后影响因素分析[J]. 中华消化病与影像杂志(电子版), 2023, 13(05): 284-288.
[9] 郭震天, 张宗明, 赵月, 刘立民, 张翀, 刘卓, 齐晖, 田坤. 机器学习算法预测老年急性胆囊炎术后住院时间探索[J]. 中华临床医师杂志(电子版), 2023, 17(09): 955-961.
[10] 李永胜, 孙家和, 郭书伟, 卢义康, 刘洪洲. 高龄结直肠癌患者根治术后短期并发症及其影响因素[J]. 中华临床医师杂志(电子版), 2023, 17(09): 962-967.
[11] 王飞飞, 王光林, 孟泽松, 李保坤, 曹龙飞, 张娟, 周超熙, 丁源一, 王贵英. 敲低IMPDH1对结肠癌SW480、HT29细胞生物功能的影响[J]. 中华临床医师杂志(电子版), 2023, 17(08): 884-890.
[12] 孔凡彪, 杨建荣. 肝脏基础疾病与结直肠癌肝转移之间关系的研究进展[J]. 中华临床医师杂志(电子版), 2023, 17(07): 818-822.
[13] 王亚丹, 吴静, 黄博洋, 王苗苗, 郭春梅, 宿慧, 王沧海, 王静, 丁鹏鹏, 刘红. 白光内镜下结直肠肿瘤性质预测模型的构建与验证[J]. 中华临床医师杂志(电子版), 2023, 17(06): 655-661.
[14] 王延召, 牛鹏飞, 丁长民, 高庆坤, 高兆亚, 安柯, 翟志超, 曾庆敏, 黄文生, 雷福明, 顾晋. 结直肠癌致腹壁巨大缺损的一期修补经验(附13例报告)[J]. 中华临床医师杂志(电子版), 2023, 17(05): 557-561.
[15] 王俊杰, 尹晓亮, 刘二腾, 陆军, 祁鹏, 胡深, 杨希孟, 陈鲲鹏, 张东, 王大明. 机器学习对预测颈内动脉非急性闭塞患者血管内再通术成功的潜在价值[J]. 中华脑血管病杂志(电子版), 2023, 17(05): 464-470.
阅读次数
全文


摘要