基于干血滴质谱代谢组学方法区分结直肠癌和结直肠息肉的研究

发布时间：2018-09-13 06:19

【摘要】：目的结直肠癌(CRC)是癌症死亡的第三大原因,早期发现可以显著提高患者的生存率。在临床中,以易于操作的方式准确的筛选和诊断结直肠癌是非常有必要的。癌症可以影响重要的代谢通路,血液中充满了具有不同结构和性质的小分子代谢物,在许多研究中代谢物水平的变化被用作潜在的生物标志物。本研究旨在筛选具有结直肠癌诊断潜力的代谢标志物。方法基于干血滴(DBS)的直接注射质谱代谢组学方法被用于区分结直肠癌与结直肠息肉。应用85名结直肠癌患者和81名结直肠息肉患者的DBS样品进行检测,检测的目标代谢物由23个氨基酸和26个酰基肉碱组成,从而构建偏最小二乘判别分析(PLS-DA)模型,用于分析数据并检验结直肠癌组和结直肠息肉组之间的差异代谢,并找出对于区分两组贡献最大的代谢物,构建二元Logistic回归模型,用ROC曲线验证这些代谢物对结直肠癌组与结直肠息肉组的判别能力。结果通过多因素分析筛选出血液中与结直肠癌相关的21种代谢物。基于参数C16(P0.0001)、Arg(P0.0001)、C4/C8(P=0.0006)、C5/C3(P0.0001)、Val(P=0.0042)、Phe/Tyr(P=0.0018)、Ala(P=0.0187)、C4/C3(P=0.0057)构建二元Logistic回归模型,用十倍交叉验证(CV)法检验这个模型,用ROC曲线来评价诊断模型,该模型显示敏感度是81.18%,特异性是83.95%,AUC是0.909。结论在这项研究中,用8个血液参数,如C16、C4/C8、C5/C3、C4/C3、Arg、Val、Ala和Phe/Tyr,建立了一个合理的诊断回归模型,可用于区分结直肠癌与结直肠息肉,并回答基于干血滴的直接注射质谱技术可用于促进CRC筛选和早期诊断。
[Abstract]:Objective (CRC) is the third leading cause of cancer death in colorectal cancer. In clinical practice, it is necessary to screen and diagnose colorectal cancer accurately and easily. Cancer can affect important metabolic pathways and the blood is filled with small molecular metabolites with different structures and properties. In many studies changes in metabolite levels have been used as potential biomarkers. The aim of this study was to screen metabolic markers with diagnostic potential for colorectal cancer. Methods Direct injection mass spectrometry based on dry blood drop (DBS) was used to distinguish colorectal cancer from colorectal polyp. DBS samples from 85 patients with colorectal cancer and 81 patients with colorectal polyps were detected. The target metabolites were composed of 23 amino acids and 26 acyl carnitine. A partial least squares discriminant analysis (PLS-DA) model was constructed. To analyze the data and examine the differential metabolism between colorectal cancer group and colorectal polyp group, and to identify the metabolites that contribute most to the distinction between the two groups, a binary Logistic regression model was constructed. The discriminant ability of these metabolites between colorectal cancer group and colorectal polyp group was verified by ROC curve. Results 21 metabolites associated with colorectal cancer were screened by multivariate analysis. A binary Logistic regression model was constructed based on C16 (P0.0001) Arg (P0.0001) C4 / C8 (P0. 0006) C5 / C3 (P0. 0001) Val (P0. 0042) and P0. 0018) Ala (P0. 0187) C4 / C3 (P0. 0057). The model was tested by ten times cross-validated (CV) method. The diagnostic model was evaluated by ROC curve. The sensitivity of the model was 81.18 and the specificity was 83.95AUC was 0. 909. Conclusion in this study, a reasonable diagnostic regression model was established using eight blood parameters, such as C16 / C4 / C8 / C5 / C3 / Argor Vala and Phe/Tyr, which can be used to distinguish colorectal cancer from colorectal polyps. Direct injection mass spectrometry based on dry blood drops can be used to promote CRC screening and early diagnosis.
【学位授予单位】：锦州医科大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：R735.34

【相似文献】