Institute of Chemical Process Fundamentals of the CAS, 165 00 Prague 6, Czech Republic.
Department of Analytical Chemistry, University of Chemistry and Technology, Prague, 166 28 Prague 6, Czech Republic.
J Proteome Res. 2023 Jun 2;22(6):1936-1946. doi: 10.1021/acs.jproteome.3c00047. Epub 2023 Apr 5.
Nuclear magnetic resonance (NMR) metabolomics was used for identification of metabolic changes in pancreatic cancer (PC) blood plasma samples when compared to healthy controls or diabetes mellitus patients. An increased number of PC samples enabled a subdivision of the group according to individual PC stages and the construction of predictive models for finer classification of at-risk individuals recruited from patients with recently diagnosed diabetes mellitus. High-performance values of orthogonal partial least squares (OPLS) discriminant analysis were found for discrimination between individual PC stages and both control groups. The discrimination between early and metastatic stages was achieved with only 71.5% accuracy. A predictive model based on discriminant analyses between individual PC stages and the diabetes mellitus group identified 12 individuals out of 59 as at-risk of development of pathological changes in the pancreas, and four of them were classified as at moderate risk.
利用核磁共振（NMR）代谢组学技术，对胰腺癌（PC）患者的血浆样本与健康对照组或糖尿病患者组进行比较，以确定代谢变化。增加了 PC 样本数量，使得能够根据个体 PC 阶段对该组进行细分，并构建预测模型，以便对新诊断为糖尿病的患者中招募的高危个体进行更精细的分类。正交偏最小二乘法（OPLS）判别分析的高性能值用于区分各个 PC 阶段和两个对照组。早期和转移性阶段之间的区分仅达到 71.5%的准确率。基于个体 PC 阶段与糖尿病组之间的判别分析的预测模型，在 59 名个体中确定了 12 名有发生胰腺病变风险的高危个体，其中 4 名被归类为中度风险。
