【摘要】：目的:探討血清腫瘤標(biāo)志物CEA、CA199、CA724、CA125、CA242和Tu M2-PK單項(xiàng)及聯(lián)合檢測(cè)對(duì)結(jié)腸癌的診斷價(jià)值,篩選出最優(yōu)的單項(xiàng)和聯(lián)合診斷結(jié)腸癌腫瘤的標(biāo)志物,并采用代謝組學(xué)的策略分別建立基于血清6種標(biāo)志物的主成分分析(PCA)、經(jīng)典判別分析、Logistic回歸分析和偏最小二乘-判別分析(PLS-DA)模型,評(píng)估各模型對(duì)結(jié)腸癌的診斷和預(yù)測(cè)價(jià)值,找出對(duì)結(jié)腸癌早期診斷最有幫助的模型。方法:收集結(jié)腸癌患者、結(jié)腸良性病變和健康對(duì)照組外周血,采用電化學(xué)發(fā)光法測(cè)定血清中CEA、CA199、CA724、CA125和CA242含量,酶聯(lián)免疫吸附試驗(yàn)(ELISA)檢測(cè)血清中Tu M2-PK的含量。組內(nèi)比較采用單因素方差分析,組間比較采用(Least-significant Difference)LSD檢驗(yàn),并以受試者工作特征曲線(ROC)下面積(AUC)的比較單個(gè)指標(biāo)的診斷效能,以Logistic回歸-ROC評(píng)估聯(lián)合指標(biāo)的診斷效能。采用PCA模型分析結(jié)腸癌、結(jié)腸良性疾病和健康對(duì)照組血清腫瘤標(biāo)志物的代謝譜,以經(jīng)典判別分析、Logistic回歸分析和偏最小二乘-判別分析(PLS-DA)模型評(píng)估結(jié)腸癌的診斷和預(yù)測(cè)價(jià)值。結(jié)果:結(jié)腸癌患者血清中CEA、CA199、CA125、CA242和Tu M2-PK水平均顯著高于良性疾病和健康對(duì)照,差異具有統(tǒng)計(jì)學(xué)意義(P0.05)。良性腫瘤組血清CEA、CA199、CA724、CA242和Tu M2-PK顯著高于健康對(duì)照組(P0.05)。六種血清腫瘤標(biāo)志物中CEA、CA199、CA724和Tu M2-PK的診斷效能較高(AUC≥0.70)。其中CA199具有最高的診斷效能,在臨界值為69.5U/L時(shí),具有64.1%的靈敏度和89.7%的特異度。單項(xiàng)標(biāo)志物中Tu M2-PK診斷結(jié)腸癌的靈敏度與CEA和CA199相當(dāng)。聯(lián)合檢測(cè)兩項(xiàng)標(biāo)志物,(CA199+Tu M2-PK)的診斷效能最高(AUC=0.87),具有77.5%的靈敏度和90.0%的特異度。聯(lián)合檢測(cè)三項(xiàng)血清腫瘤標(biāo)志物,(CEA+CA199+Tu M2-PK)的診斷效能最高,具有82.3%的靈敏度和91.5%的特異性。聯(lián)合檢測(cè)四項(xiàng)標(biāo)志物(CEA+CA72-4+Tu M2-PK+CA19-9),具有87.5%的靈敏度、83.0%的特異度和84.5%的準(zhǔn)確度。PCA模型表明,結(jié)腸癌患者血清6種腫瘤標(biāo)志物水平明顯紊亂,結(jié)腸癌患者個(gè)體分散,正常個(gè)體和結(jié)腸良性病患者個(gè)體變異小,三組個(gè)體有分離的趨勢(shì)。以AUC最高三項(xiàng)腫瘤標(biāo)志物(CEA+CA199+Tu M2-PK)建立Fisher經(jīng)典判別分析模型,該模型對(duì)結(jié)腸癌和非結(jié)腸癌(結(jié)腸良性疾病和正常對(duì)照組)的診斷正確率分別為(162/231)70.1%和(227/271)81.9%。以AUC最高三項(xiàng)腫瘤標(biāo)志物(CEA+CA199+Tu M2-PK)建立PLS-DA診斷模型,該模型對(duì)結(jié)腸癌和非結(jié)腸癌(結(jié)腸良性病組和對(duì)照組)診斷的準(zhǔn)確率分別為(179/231)77.5%和(259/271)95.6%。結(jié)論:六種血清腫瘤標(biāo)志物當(dāng)中,以CA199對(duì)結(jié)腸癌的早期診斷最有價(jià)值。而聯(lián)合檢測(cè)血清腫瘤標(biāo)志物CEA、CA199和Tu M2-PK是較為經(jīng)濟(jì)、有效的組合標(biāo)志物,有助于結(jié)腸癌的早期診斷�；谘迥[瘤標(biāo)志物CEA、CA199和Tu M2-PK的Logistic回歸和偏最小二乘判別分析(PLS-DA)診斷模型為結(jié)腸癌的診斷提供了可行有效的新模式,有助于結(jié)腸癌的早期診斷。
[Abstract]:Objective: to investigate the diagnostic value of serum tumor markers CEA,CA199,CA724,CA125,CA242 and Tu M2-PK in the diagnosis of colon cancer, to screen out the optimal markers for the diagnosis of colon cancer tumors, and to establish the classical discriminant analysis of (PCA), Logistic regression analysis and partial least squares discriminant analysis (PLS-DA) models based on the strategy of metabonomics. To evaluate the value of each model in the diagnosis and prediction of colon cancer, and to find the most helpful model for the early diagnosis of colon cancer. Methods: the peripheral blood of patients with colon cancer, benign lesions of colon and healthy control group were collected. The contents of CEA,CA199,CA724,CA125 and CA242 in serum were determined by electrochemiluminescence method, and the content of Tu M2-PK in serum was detected by enzyme-linked immunosorbent assay (ELISA). One factor analysis of variance was used in the group, (Least-significant Difference) LSD test), and the diagnostic efficacy of the single index was compared with the area (AUC) under the working characteristic curve (ROC) of the subjects, and the diagnostic effectiveness of the combined index was evaluated by Logistic regression ROC. PCA model was used to analyze the metabolic spectrum of serum tumor markers in colon cancer, benign colon diseases and healthy controls. Classical discriminant analysis, Logistic regression analysis and partial least squares discriminant analysis (PLS-DA) model were used to evaluate the diagnostic and predictive value of colon cancer. Results: the levels of CEA,CA199,CA125,CA242 and Tu M2-PK in serum of patients with colon cancer were significantly higher than those of benign diseases and healthy controls (P 0.05). Serum CEA,CA199,CA724,CA242 and Tu M2-PK in benign tumor group were significantly higher than those in healthy control group (P 0.05). Among the six serum tumor markers, CEA,CA199,CA724 and Tu M2-PK were more effective (AUC 鈮，

本文編號(hào)：2503944