基于1H-NMR的代谢组学技术建立老年癌性急腹症病理分期预测模型的研究

发布时间：2018-05-27 03:17

本文选题：1H-NMR + 代谢组技术　；参考：《川北医学院》2017年硕士论文

【摘要】：目的:基于1H--NMR代谢组学平台,探索老年癌性急腹症患者的代谢表型特征,并结合代谢网络分析,揭示老年癌性急腹症患者代谢网络扰动机制。进一步,采用计算生物学方法建立一种快速、精准的老年癌性急腹症术前病理分期的评估技术。方法:采用队列研究设计,纳入在四川人民医院急诊外科收治的由结直肠癌而导致的急腹症患者,纳入标准为:(1)年龄60岁及以上;(2)急诊以急腹症入院,临床初诊为结直肠癌的患者;(3)患者及家属同意参加本研究,并签字确认;排除标准为:(1)糖尿病以及其他的内分泌问题;(2)合并其他肿瘤;(3)妊娠期或哺乳期的女性;(4)精神状态无法有效的合作,存在一定的精神性问题,缺乏良好的自制力以及高效的表述;(5)参加其他临床试验的患者:分别在入组后当日立即进行抽血操作(2ml),并提取其血浆备用。随后对粗提血浆进行高速离心操作,并按照本实验室建立的氢核磁共振(1H-NMR)代谢组学标本前处理常规操作流程(standard operation process,SOP)进行样本前处理。前处理完成的样本置入核磁检测管后,即放入BrukerAvvance DRx600傅立叶变换核磁共振仪中进行检测。检测结果获取一维NMR谱图。原始NMR谱图数据,先进行傅立叶变换,然后使用MsetReNova软件进一步处理,获得199个ppm段进行桶装积分。而后将获得数据导入MATLAB软件,使用PLS-tool box工具进行偏最小二乘法验算操作。将PLS-DA聚类信息与患者的病理TNM分期进行比对。进而着手构建对应的回溯分析技术并建立代谢模型,然后通过主成分分析以及偏最小二乘法以及支持向量进行数据分析。对于临床资料,计量资料采用Wilcoxon秩和检验(非正态分布分析)或方差分析(正态分布)来确定不同因素之间的相关性。结果:2014年10月到2016年11月,共纳入62例不同分期的患者。人口学资料表明,患者的年龄、体重、便血无差异,根据血浆核磁共振图谱来对不同时期的老年癌性急腹症患者进行区分,其血浆磁共振波谱指纹图谱可特征性区分不同肿瘤分期。在不同TNM分期患者中,其血浆代谢池中存在明显的代谢物差异,III、IV期老年癌性急腹症患者相对于I、II期患者,在血浆代谢池中存在明显的代谢物差异,其中包括了多种氨基酸如:亮氨酸、缬氨酸、丙氨酸、谷氨酰胺、谷氨酸、组氨酸、酪氨酸、甘氨酸、酪氨酸、异亮氨酸、等含量显著下降,(P0.05),而在血浆代谢池中的有关乳酸、肌酸以及胆碱等物质也存在极为显著的降低(P0.05),但是,I、II期患者的有关血浆代谢池中的极低密度脂蛋白、胆固醇以及丙酮等物质却存在一定的提升,(P0.05)。经过计算,最终筛选出谷氨酸、异亮氨酸、胆碱、缬氨酸、a-葡萄糖、β-葡萄糖6个关键性代谢标志物是100对酶和基因。通过对其中主要的代谢标志物进行分析,以此建立术前快速临床分期预测模型,该预测模型的均方根误(RMSE)为5.2,R2=0.47(P=0.036),结果显示预测效果良好。结论:基于1H--NMR代谢组学技术建立老年癌性急腹症患者病理分期预测模型,能够有效的对术前老年癌性急腹症患者进行临床分期。
[Abstract]:Objective: Based on the 1H--NMR metabonomics platform, the metabolic phenotypic characteristics of the elderly patients with cancer acute abdomen were explored and metabolic network analysis was used to reveal the metabolic network disturbance mechanism of the elderly patients with cancer acute abdomen. Further, a rapid and accurate preoperative pathological stage evaluation of the elderly patients with cancer acute abdomen was established by using computational biology methods. Methods: a cohort study was designed to include patients with acute abdomen caused by colorectal cancer at the emergency surgery in the people's Hospital of Sichuan. The criteria were: (1) age 60 and above; (2) emergency patients were hospitalized with acute abdomen, patients were first diagnosed as colorectal cancer patients; (3) patients and family members agreed to participate in the study and signed confirmation; exclusion criteria. (1) diabetes and other endocrine problems; (2) combined with other tumors; (3) women in pregnancy or lactation; (4) unable to cooperate effectively in the mental state; there are certain mental problems, lack of good self-control and efficient expression; (5) patients who participated in other clinical trials were immediately taken blood exercises on the same day after entering the group. To make (2ml) and extract its plasma reserve. Then, the high speed centrifugation was performed on the crude extract plasma, and the pre processing of the routine operation process (standard operation process, SOP) was carried out according to the hydrogen nuclear magnetic resonance (standard operation process, SOP) set up in our laboratory. The pre processed samples were placed in the NMR tube, which was put into the BrukerAvv Ance DRx600 Fu Liye transform nuclear magnetic resonance (NMR) detector. The results obtain one dimensional NMR spectrum. The original NMR spectrum data, first carry out the Fu Liye transform, and then use MsetReNova software further processing, get 199 ppm segments for barrel integration. Then the data is imported to MATLAB software, using PLS-tool box tool to minimize the deviation. Two multiplication checking operation. Compare the PLS-DA clustering information with the patient's pathological TNM stage, and then build the corresponding backtracking analysis technique and establish the metabolic model, and then analyze the data through the principal component analysis and the partial least squares and support vector. The Wilcoxon rank and test for the clinical data are used. Normal distribution analysis) or variance analysis (normal distribution) to determine the correlation between different factors. Results: from October 2014 to November 2016, 62 patients with different stages were included. Demographic data showed that the age, weight, and blood pressure of the patients were not different, according to the plasma nuclear magnetic resonance spectrum, the elderly patients with cancer of different periods were diagnosed as acute abdomen. In different TNM staging patients, there are significant differences in metabolites in the plasma metabolites of the patients with different TNM stages. There are significant differences in metabolites in III and IV elderly patients with I, II, and plasma metabolites, including a variety of different types of metabolites. Amino acids, such as leucine, valine, alanine, glutamine, glutamic acid, histidine, tyrosine, glycine, glycine, tyrosine, isoleucine, were significantly decreased, (P0.05), while in the plasma metabolite, lactic acid, creatine and choline were also significantly reduced (P0.05), but, I, II patients related to the plasma metabolism pool. The extremely low density lipoprotein, cholesterol and acetone have a certain increase, (P0.05). After calculation, the 6 key metabolic markers of glutamic acid, isoleucine, choline, valine, a- glucose and beta glucose are 100 pairs of enzymes and genes. The preoperatively fast clinical staging model, the root mean square error (RMSE) of the predicted model was 5.2, R2=0.47 (P=0.036), and the results showed good prediction results. Conclusion: Based on the 1H--NMR metabolic histopathology technique, the pathological staging model of the elderly patients with cancer acute abdomen can be established, which can be effective in the clinical staging of the elderly patients with cancer acute abdomen.
【学位授予单位】：川北医学院
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：R735.34

【参考文献】