食管癌高危人群舌象特征分析与应用探索
发布时间：2018-08-17 13:08
【摘要】：研究背景河北省磁县是我国食管癌的高发区,食管癌的发病率为91.5/10万,明显高于全国平均水平。2005年磁县开始实施食管癌早诊早治项目,针对当地高危人群(40-69岁)进行内镜普查,对发现可疑病变进行组织病理学检查。然而,当地食管癌的发病率并没有得到有效控制。中医认为舌通过经络连接到人体内部器官,器官条件、属性和功能的变化可以通过观察舌象被发现,中医舌诊在食管癌筛查中的作用值得深入研究。第一部分：食管癌高危人群舌象特征分析目的分析食管癌高危人群不同病理诊断者之间舌象的差异,包括舌象颜色空间色值和舌象特征。方法在食管癌高发区河北省磁县纳入筛查者3053例,依据病理及胃镜诊断结果,分为正常组(包括炎症组、非炎症组)、食管低级别上皮内瘤变组(以下称食管低级别瘤变组)、食管高级别上皮内瘤变组(以下称食管高级别瘤变组)。收集筛查者的基本信息资料,利用舌面诊测信息采集系统(DS01-B),采集提取筛查者的舌象颜色空间色值(包括HSV、Lab、RGB三种颜色空间色值)及舌象特征(包括舌体颜色、舌苔颜色、舌体胖瘦等共10项)。结果1.正常组、食管低级别瘤变组、食管高级别瘤变组间性别、年龄的整体分布可见明显的统计学差异(P0.01)。2.在HSV颜色空间中,食管高级别瘤变组与正常组间S值比较,显示出明显的统计学差异(P0.05)。3.在Lab颜色空间中,正常组与食管低级别+高级别上皮内瘤变组间b值比较,显示出明显的统计学差异(P0.05)。4.正常组、食管低级别瘤变组、食管高级别瘤变组间舌体瘀斑的整体分布可见明显统计学差异(P0.01),舌体瘀斑在食管低级别瘤变组、食管高级别瘤变组中的比例,高于正常组。5.在正常组中,炎症组与非炎症组间舌体颜色整体分布可见明显统计学差异(P0.01)。6.癌+癌前病变组与非炎症组间舌体瘀斑的整体分布可见明显统计学差异(P0.01)。结论正常组与食管低级别瘤变组、食管高级别瘤变组之间舌象特征具有差异,其中舌体瘀斑具有明显统计学差异。第二部分：基于舌象特征的食管癌及癌前病变logistic回归模型的建立与验证目的利用舌象特征和性别、年龄,建立食管癌及癌前病变logistic回归模型,分析食管癌及癌前病变的危险因素,探讨食管癌的中医病因病机及中医证型,并进行模型在食管癌筛查中的应用探索。方法将纳入研究的3883例筛查者,依据入组时间分为训练组(3053例)及检测组(830例)。将训练组的舌象特征及性别年龄进行赋值,建立食管癌及癌前病变logistic回归模型,并通过检测组验证模型的准确性。结果1. Logistic回归分析显示,红舌、舌体瘀斑、黄白相兼苔是食管癌的高危因素。2. Logistic回归模型的内部诊断特异性为72.80%,敏感性为61.60%,总符合率为72.3%。外部诊断特异性为80.35%,敏感性为63.41%,总符合率为79.51%。结论基于舌象特征、性别、年龄建立的食管癌logistic回归模型,发现舌象特征中的红舌、舌体瘀斑、黄白相兼苔是食管癌及癌前病变的高危因素,将模型用于检验食管癌高危人群,与病理诊断间有较高的符合率。
[Abstract]:Background Cixian County, Hebei Province, is a high incidence area of esophageal cancer in China. The incidence of esophageal cancer is 91.5/100,000, which is significantly higher than the national average. In 2005, Cixian began to implement the project of early diagnosis and treatment of esophageal cancer. Endoscopic screening was carried out for the local high-risk population (40-69 years old) and the suspicious lesions were detected by histopathological examination. The incidence of esophageal cancer has not been effectively controlled. Traditional Chinese medicine holds that the tongue is connected to the internal organs, organ conditions, attributes and functions of the human body through meridians and collaterals. The role of tongue diagnosis in esophageal cancer screening deserves further study. Part I: Analysis of tongue characteristics in high-risk groups of esophageal cancer. Methods A total of 3053 cases were screened in Cixian County of Hebei Province, a high-risk area of esophageal cancer, and were divided into normal group (including inflammation group, non-inflammation group) and low-grade esophageal intraepithelial neoplasia group (hereinafter referred to as feeding group). The basic information of screeners was collected, and the tongue color space color values (including HSV, Lab, RGB) and tongue features (including tongue color) were collected and extracted by the tongue diagnosis information collection system (DS01-B). Results 1. There were significant differences in sex and age distribution among the normal group, low grade esophageal neoplasia group and high grade esophageal neoplasia group (P 0.01). 2. In the HSV color space, the S value between the high grade esophageal neoplasia group and the normal group was significantly different (P 0.05). 3. In Lab, there were significant differences in the S value between the high grade esophageal neoplasia group and the normal group (P 0.05). In color space, there were significant differences in B values between normal group and esophageal low-grade + high-grade intraepithelial neoplasia group (P 0.05). 4. There were significant differences in the overall distribution of tongue ecchymosis between normal group, esophageal low-grade neoplasia group and esophageal high-grade neoplasia group (P 0.01). The proportion of high-grade neoplasia group was higher than that of the normal group. 5. In the normal group, the overall distribution of tongue color between inflammation group and non-inflammation group was significantly different (P 0.01). 6. The overall distribution of tongue ecchymosis between cancer + precancerous lesion group and non-inflammation group was significantly different (P 0.01). There were significant differences in tongue features between the high-grade canal neoplasia group and the tongue ecchymosis group. Part 2: Establishment and validation of logistic regression model of esophageal cancer and precancerous lesions based on tongue features Objective To establish logistic regression model of esophageal cancer and precancerous lesions by using tongue features and gender, age, and analyze food intake. Methods 3883 cases of esophageal cancer screening were divided into training group (3053 cases) and testing group (830 cases) according to the time of enrollment. Results 1. Logistic regression analysis showed that red tongue, tongue ecchymosis, yellow and white fur were the high risk factors of esophageal cancer. 2. Logistic regression model had 72.80% specificity, 61.60% sensitivity and 72.3% overall coincidence rate. The diagnostic specificity was 80.35%, the sensitivity was 63.41%, and the total coincidence rate was 79.51%. There was a high coincidence rate between the two diagnoses.
【学位授予单位】：北京中医药大学
【学位级别】：博士
【学位授予年份】：2016
【分类号】：R735.1
[Abstract]:Background Cixian County, Hebei Province, is a high incidence area of esophageal cancer in China. The incidence of esophageal cancer is 91.5/100,000, which is significantly higher than the national average. In 2005, Cixian began to implement the project of early diagnosis and treatment of esophageal cancer. Endoscopic screening was carried out for the local high-risk population (40-69 years old) and the suspicious lesions were detected by histopathological examination. The incidence of esophageal cancer has not been effectively controlled. Traditional Chinese medicine holds that the tongue is connected to the internal organs, organ conditions, attributes and functions of the human body through meridians and collaterals. The role of tongue diagnosis in esophageal cancer screening deserves further study. Part I: Analysis of tongue characteristics in high-risk groups of esophageal cancer. Methods A total of 3053 cases were screened in Cixian County of Hebei Province, a high-risk area of esophageal cancer, and were divided into normal group (including inflammation group, non-inflammation group) and low-grade esophageal intraepithelial neoplasia group (hereinafter referred to as feeding group). The basic information of screeners was collected, and the tongue color space color values (including HSV, Lab, RGB) and tongue features (including tongue color) were collected and extracted by the tongue diagnosis information collection system (DS01-B). Results 1. There were significant differences in sex and age distribution among the normal group, low grade esophageal neoplasia group and high grade esophageal neoplasia group (P 0.01). 2. In the HSV color space, the S value between the high grade esophageal neoplasia group and the normal group was significantly different (P 0.05). 3. In Lab, there were significant differences in the S value between the high grade esophageal neoplasia group and the normal group (P 0.05). In color space, there were significant differences in B values between normal group and esophageal low-grade + high-grade intraepithelial neoplasia group (P 0.05). 4. There were significant differences in the overall distribution of tongue ecchymosis between normal group, esophageal low-grade neoplasia group and esophageal high-grade neoplasia group (P 0.01). The proportion of high-grade neoplasia group was higher than that of the normal group. 5. In the normal group, the overall distribution of tongue color between inflammation group and non-inflammation group was significantly different (P 0.01). 6. The overall distribution of tongue ecchymosis between cancer + precancerous lesion group and non-inflammation group was significantly different (P 0.01). There were significant differences in tongue features between the high-grade canal neoplasia group and the tongue ecchymosis group. Part 2: Establishment and validation of logistic regression model of esophageal cancer and precancerous lesions based on tongue features Objective To establish logistic regression model of esophageal cancer and precancerous lesions by using tongue features and gender, age, and analyze food intake. Methods 3883 cases of esophageal cancer screening were divided into training group (3053 cases) and testing group (830 cases) according to the time of enrollment. Results 1. Logistic regression analysis showed that red tongue, tongue ecchymosis, yellow and white fur were the high risk factors of esophageal cancer. 2. Logistic regression model had 72.80% specificity, 61.60% sensitivity and 72.3% overall coincidence rate. The diagnostic specificity was 80.35%, the sensitivity was 63.41%, and the total coincidence rate was 79.51%. There was a high coincidence rate between the two diagnoses.
【学位授予单位】：北京中医药大学
【学位级别】：博士
【学位授予年份】：2016
【分类号】：R735.1
【相似文献】
相关期刊论文 前10条
1 高金金;张葆青;;中医舌象研究概况[J];辽宁中医药大学学报;2012年03期
2 童国tT;陈利铭;林宝q，
本文编号：2187728
本文链接：https://www.wllwen.com/yixuelunwen/zlx/2187728.html
