基于术前检测指标构建Logistic回归模型在预测壶腹周围癌病理类型中的价值

doi:10.3877/cma.j.issn.2095-3232.2023.02.014

目的

探讨基于术前检测指标构建Logistic回归模型在预测壶腹周围癌病理类型中的价值。

方法

本研究对象为2015年1月至2018年12月在西安交通大学第一附属医院行根治性胰十二指肠切除术的256例壶腹周围癌患者。其中男159例，女97例；年龄39~88岁，中位年龄64岁。患者均签署知情同意书，符合医学伦理学规定。收集患者术前数据和术后病理类型，建立预测壶腹周围癌病理类型的Logistic回归模型、决策树模型和随机森林模型。随机选取192例作为训练集，64例为测试集，采用ROC曲线下面积（AUC）用于交互验证模型预测价值。

结果

所有患者病理类型中，肠型99例，胰胆管型157例。基于肿瘤直径、Hb变量构建多元Logistic回归模型，信息准则（AIC）值达到最低24。根据AdaBoost算法结果，CEA、CA125、ALB、肿瘤直径、CA19-9等是影响决策树的重要因素，分别设置随机种子为91和62得到相应决策树模型和随机森林模型。测试集验证Logistic回归、决策树、随机森林模型的准确度依次为0.810、0.766、0.766，敏感度依次为0.848、0.775、0.956，特异度依次为0.706、0.750、0.316，AUC依次为0.764、0.762、0.698。

结论

基于肿瘤直径、Hb建立的Logistic回归模型可在术前较为准确判别壶腹周围癌的病理类型，指导术前辅助化疗。

关键词: 壶腹周围癌, 病理类型, Logistic回归模型, 决策树模型, 随机森林模型

Objective

To evaluate the prediction value of constructing Logistic regression model based on the preoperative detection indexes in the pathological types of periampullary carcinoma.

Methods

256 patients with periampullary carcinoma who underwent radical pancreaticoduodenectomy in the First Affiliated Hospital of Xi'an Jiaotong University from January 2015 to December 2018 were recruited in this clinical trial. Among them, 159 patients were male and 97 female, aged from 39 to 88 years, with a median age of 64 years. The informed consents of all patients were obtained and the local ethical committee approval was received. Preoperative data and postoperative pathological types of all patients were collected, and Logistic regression model, decision tree model and random forest model were established to predict the pathological types of periampullary carcinoma. 192 cases were randomly selected into the training set and 64 cases into the test set. The predictive value of this model was assessed by the area under ROC curve (AUC).

Results

Among all patients, 99 patients were classified as the intestinal type and 157 cases of pancreaticobiliary type. Based on the tumor diameter and Hb variables, a multivariate Logistic regression model was constructed, and the Akaike information criterion (AIC) value reached the lowest 24. According to the results of AdaBoost algorithm, CEA, CA125, ALB, tumor diameter and CA19-9 were important factors affecting the decision tree. The corresponding decision tree model and random forest model were obtained by setting random seeds as 91 and 62, respectively. In the test set, the accuracy of Logistic regression, decision tree and random forest models was 0.810, 0.766 and 0.766, the sensitivity was 0.848, 0.775 and 0.956, the specificity was 0.706, 0.750 and 0.316, and the AUC was 0.764, 0.762 and 0.698, respectively.

Conclusions

Logistic regression model based on tumor diameter and Hb can accurately identify the pathological types of periampullary carcinoma before surgery, and provide guidance for the preoperative adjuvant chemotherapy.

Key words: Periampullary carcinoma, Pathological type, Logistic regression model, Decision tree model, Random forest model

表1 壶腹周围癌患者肿瘤病理类型影响因素的单因素分析结果

指标	肠型	胰胆管型	检验值	P值
肿瘤直径(cm，±s)	3.0±1.3	2.3±1.0	t=4.39	<0.001
Hb (g/L，±s)	124±18	117±20	t=2.77	0.010
HCT (±s)	0.38±0.05	0.36±0.06	t=2.36	0.020
WBC (×10⁹/L，±s)	5.7±2.1	6.5±2.8	t=-2.56	0.010
Ca²⁺ (mmol/L，±s)	2.26±0.13	2.22±0.17	t=1.96	0.050
ALB (g/L，±s)	38±5	37±6	t=2.52	0.010
Plt (>138×10⁹/L，例)	58	104	χ²=2.14	0.050
CEA (>1.72 μg/L，例)	19	14	χ²=4.88	0.030
CA19-9 (>18.53 kU/L，例)	19	14	χ²=4.87	0.030
IB (>20.5 μmol/L，例)	46	52	χ²=2.02	0.040
DB (>103.6 μmol/L，例)	45	53	χ²=4.60	0.030

表1 壶腹周围癌患者肿瘤病理类型影响因素的单因素分析结果

指标	肠型	胰胆管型	检验值	P值
肿瘤直径(cm，±s)	3.0±1.3	2.3±1.0	t=4.39	<0.001
Hb (g/L，±s)	124±18	117±20	t=2.77	0.010
HCT (±s)	0.38±0.05	0.36±0.06	t=2.36	0.020
WBC (×10⁹/L，±s)	5.7±2.1	6.5±2.8	t=-2.56	0.010
Ca²⁺ (mmol/L，±s)	2.26±0.13	2.22±0.17	t=1.96	0.050
ALB (g/L，±s)	38±5	37±6	t=2.52	0.010
Plt (>138×10⁹/L，例)	58	104	χ²=2.14	0.050
CEA (>1.72 μg/L，例)	19	14	χ²=4.88	0.030
CA19-9 (>18.53 kU/L，例)	19	14	χ²=4.87	0.030
IB (>20.5 μmol/L，例)	46	52	χ²=2.02	0.040
DB (>103.6 μmol/L，例)	45	53	χ²=4.60	0.030

表2 壶腹周围癌患者肿瘤病理类型的多元Logistic回归模型

指标	估计值	SE值	Z值	P值	OR值	95%CI
肿瘤直径	0.414	0.138	-2.990	0.003	1.715	1.334~2.205
Hb	0.021	0.009	-2.452	0.014	1.021	1.006~1.036
截距	-3.574	1.164	3.070	0.002	-	-

表2 壶腹周围癌患者肿瘤病理类型的多元Logistic回归模型

指标	估计值	SE值	Z值	P值	OR值	95%CI
肿瘤直径	0.414	0.138	-2.990	0.003	1.715	1.334~2.205
Hb	0.021	0.009	-2.452	0.014	1.021	1.006~1.036
截距	-3.574	1.164	3.070	0.002	-	-

图1 Adaboost算法各变量对决策树的影响程度

图2 壶腹周围癌患者肿瘤病理类型的决策树预测模型注：DD为D-二聚体，FDP为纤维蛋白原降解产物，TEG-CI为血栓弹力图综合指数，HCT为红细胞比容；0为胰胆管型壶腹周围癌，1为肠型壶腹周围癌

图3 三种壶腹周围癌病理类型预测模型的ROC曲线注：AUC为ROC曲线下面积；红色为多元Logistic回归模型，蓝色为决策树模型，绿色为随机森林模型

表3 三种壶腹周围癌病理类型预测模型的预测价值比较

模型	敏感度	特异度	准确度	阳性预测值	阴性预测值	AUC	95%CI	P值
Logistic回归模型	0.886	0.632	0.810	0.848	0.706	0.764	0.632~0.901	0.001
决策树模型	0.838	0.667	0.766	0.775	0.750	0.762	0.559~0.798	<0.001
随机森林模型	0.768	0.750	0.766	0.956	0.333	0.698	0.578~0.804	<0.001

表3 三种壶腹周围癌病理类型预测模型的预测价值比较

模型	敏感度	特异度	准确度	阳性预测值	阴性预测值	AUC	95%CI	P值
Logistic回归模型	0.886	0.632	0.810	0.848	0.706	0.764	0.632~0.901	0.001
决策树模型	0.838	0.667	0.766	0.775	0.750	0.762	0.559~0.798	<0.001
随机森林模型	0.768	0.750	0.766	0.956	0.333	0.698	0.578~0.804	<0.001

[1]	Siegel RL, Miller KD, Jemal A. Cancer statistics, 2018[J]. CA Cancer J Clin, 2018, 68(1):7-30.
[2]	Williams JL, Chan CK, Toste PA, et al. Association of histopathologic phenotype of periampullary adenocarcinomas with survival[J]. JAMA Surg, 2017, 152(1):82-88.
[3]	赵静, 宋彬. 壶腹周围癌影像学诊断方法的研究现状与进展[J]. 中国普外基础与临床杂志, 2022, 29(2):239-242.
[4]	Radojkovic M, Stojanovic M, Radojković D, et al. Histopathologic differentiation as a prognostic factor in patients with carcinoma of the hepatopancreatic ampulla of Vater[J]. J Int Med Res, 2018, 46(11): 4634-4639.
[5]	Acharya A, Markar SR, Sodergren MH, et al. Meta-analysis of adjuvant therapy following curative surgery for periampullary adenocarcinoma[J]. Br J Surg, 2017, 104(7):814-822.
[6]	祝琼洁, 张冰, 黎琪, 等. MRI术前鉴别诊断肠型和胰胆管型壶腹周围癌的价值研究[J]. 南京医科大学学报(自然科学版), 2021, 41(6):892-898.
[7]	Baghmar S, Agrawal N, Kumar G, et al. Prognostic factors and the role of adjuvant treatment in periampullary carcinoma: a single-centre experience of 95 patients[J]. J Gastrointest Cancer, 2019, 50(3):361-369.
[8]	Manohar KM, Sasikala M, Anuradha S, et al. MUC1, CK20, and CDX2 immunohistochemical markers can sub-classify periampullary carcinomas into pancreaticobiliary, intestinal, and mixed subtypes[J]. Indian J Pathol Microbiol, 2021, 64(4):693-701.
[9]	Sree UD, Prayaga AK, R Reddy VV, et al. Differential expression of CK7, CK20, CDX2 in intestinal and pancreatobiliary types of preriampullary carcinoma[J]. Indian J Pathol Microbiol, 2022, 65(1): 42-48.
[10]	van Roessel S, Soer EC, Daamen LA, et al. Preoperative misdiagnosis of pancreatic and periampullary cancer in patients undergoing pancreatoduodenectomy: a multicentre retrospective cohort study[J]. Eur J Surg Oncol, 2021, 47(10):2525-2532.
[11]	Dang C, Wang M, Zhu F, et al. Comparison of laparoscopic and open pancreaticoduodenectomy for the treatment of nonpancreatic periampullary adenocarcinomas: a propensity score matching analysis[J]. Am J Surg, 2021, 222(2):377-382.
[12]	Bi L, Dong Y, Jing C, et al. Differentiation of pancreatobiliary-type from intestinal-type periampullary carcinomas using 3.0T MRI[J].J Magn Reson Imaging, 2016, 43(4):877-886.
[13]	Duan Z, Zhang Y, Tang Y, et al. Adjuvant therapy for periampullary carcinoma and the significance of histopathological typing: a systematic review[J]. Transl Oncol, 2022(20):101414.
[14]	Komine R, Kojima M, Ishi G, et al. Recognition and pathological features of periampullary region adenocarcinoma with an indeterminable origin[J]. Cancer Med, 2021, 10(11):3499-3510.
[15]	Hank T, Klaiber U, Sahora K, et al. Surgery for periampullary pancreatic cancer[J]. Chirurg, 2021, 92(9):776-787.
[16]	Tang Y, Zheng Y, Chen X, et al. Identifying periampullary regions in MRI images using deep learning[J]. Front Oncol, 2021(11):674579.
[17]	Poruk KE, Gay DZ, Brown K, et al. The clinical utility of CA19-9 in pancreatic adenocarcinoma: diagnostic and prognostic updates[J]. Curr Mol Med, 2013, 13(3):340-351.
[18]	曹洪滔. 合并梗阻性黄疸的壶腹周围癌患者术前减黄的回顾性临床对照研究[D]. 北京: 中国医学科学院, 2021.
[19]	Peng X, Jiao X, Zhao P, et al. Influence of non-jaundice stage at diagnosis on clinicopathological features and long-term survival of patients with periampullary carcinomas[J]. Medicine, 2019, 98(45): e17673.
[20]	Schmuck RB, Brokat C, Andreou A, et al. Clinicopathological stratification and long-term follow-up of patients with periampullary carcinomas[J]. Anticancer Res, 2018, 38(9):5379-5386.

[1]	赵洪峰, 王淑颖, 胡炜, 聂世姣, 费莹, 石尚世, 储华英, 王剑荣. 体外膜肺氧合相关血流感染危险因素及预测模型建立[J]. 中华危重症医学杂志(电子版), 2023, 16(02): 98-104.
[2]	何玥薇, 田杰, 易岂建, 刘晓燕, 吕铁伟, 计晓娟, 李谧. 血清降钙素原联合C反应蛋白对儿童川崎病并发冠状动脉损害的预测价值[J]. 中华妇幼临床医学杂志(电子版), 2018, 14(01): 31-37.
[3]	李闯, 林湘杰, 樊敬文, 杨雪菲. 内镜下逆行阑尾炎治疗术临床应用现状的范围综述[J]. 中华普通外科学文献(电子版), 2022, 16(05): 366-370.
[4]	葛卫卫, 李伟, 李荣兴, 成亚会, 李青松. MSCT与HRCT在不同病理类型肺磨玻璃结节中的临床应用[J]. 中华肺部疾病杂志(电子版), 2023, 16(04): 572-574.
[5]	徐春玲, 张雪, 史曙光, 李树芸, 苗重昌. 肺磨玻璃肺结节CT三维重在早期肺癌及病理类型的鉴别意义[J]. 中华肺部疾病杂志(电子版), 2022, 15(03): 381-384.
[6]	赵芬田, 于洪福, 李坤泽, 赵家亮. 不同病理类型周围型肺癌CT影像征象分析[J]. 中华肺部疾病杂志(电子版), 2022, 15(03): 370-372.
[7]	陈国标, 杜巍, 周建平, 卢星照, 黄锡安, 胡伟成, 刘鑫. 肺癌病例分布及病理特征趋势分析[J]. 中华肺部疾病杂志(电子版), 2021, 14(05): 590-592.
[8]	陈文, 单奔, 陈小宇. 高分辨率CT下肺磨玻璃结节影像学特征以及术后病理比较[J]. 中华肺部疾病杂志(电子版), 2021, 14(03): 288-292.
[9]	武帅, 吴含雪, 莫建涛, 杨沛泽, 程亚丽, 仵正. 术前胆道引流对壶腹周围癌患者术后并发症和远期生存的影响[J]. 中华肝脏外科手术学电子杂志, 2023, 12(01): 82-86.
[10]	梁嘉毅, 张旭峰, 何晓, 谢斌辉, 刘清泉. 腹腔镜胰十二指肠切除术13例临床分析[J]. 中华肝脏外科手术学电子杂志, 2020, 09(05): 462-465.
[11]	李乐乐, 林沛楠, 孙婧, 陈晓锋, 李萍, 顾艳宏. 青年结直肠癌患者临床及病理特征分析[J]. 中华结直肠疾病电子杂志, 2017, 06(04): 310-315.
[12]	王姚, 张川. 胰胆管定量分析对壶腹周围癌的鉴别诊断价值[J]. 中华临床医师杂志(电子版), 2018, 12(07): 400-404.

阅读次数

全文

摘要

选择文件类型/文献管理软件名称

选择包含的内容

Prediction value of Logistic regression model based on preoperative detection indexes in pathological types of periampullary carcinoma

模态框（Modal）标题

选择文件类型/文献管理软件名称

选择包含的内容

Prediction value of Logistic regression model based on preoperative detection indexes in pathological types of periampullary carcinoma