Gas chromatography-mass spectrometry untargeted profiling of non-Hodgkin’s lymphoma urinary metabolite markers

• Published in: Analytical and bioanalytical chemistry 2020-11, Vol.412 (27), p.7469-7480
• Main Authors: Bueno Duarte, Gustavo Henrique, de Piloto Fernandes, Anna Maria Alves, Silva, Alex Aparecido Rosini, Zamora-Obando, Hans R., Amaral, Alan Gonçalves, de Sousa Mesquita, Alessandra, Schmidt-Filho, Jayr, Cordeiro de Lima, Vladmir C., D’Almeida Costa, Felipe, Andrade, Victor Piana, Porcari, Andreia M., Eberlin, Marcos Nogueira, Simionato, Ana Valéria Colnaghi
• Format: Article
• Language: English
• Subjects: Adult; Aged; Algorithms; Amino acids; Analytical Chemistry; Biochemistry; Biomarkers, Tumor - metabolism; Biomarkers, Tumor - urine; Characterization and Evaluation of Materials; Chemistry; Chemistry and Materials Science; Chromatography; Data analysis; Female; Food Science; Gas chromatography; Gas Chromatography-Mass Spectrometry - methods; Humans; Information management; Invasiveness; Laboratory Medicine; Lymphatic system; Lymphocytes; Lymphocytes T; Lymphoma; Lymphoma, Non-Hodgkin - metabolism; Lymphoma, Non-Hodgkin - urine; Male; Mass spectrometry; Mass spectroscopy; Metabolites; Metabolome; Metabolomics - methods; Middle Aged; Model accuracy; Monitoring/Environmental Analysis; Multivariate analysis; Neoplasia; Neoplasms; Non-Hodgkin's lymphoma; Non-Hodgkin's lymphomas; Organic acids; Research Paper; Scientific imaging; Spectroscopy; Spectrum analysis; Statistical analysis; Statistical methods; Sugar; Support vector machines; Vitamins; Volatile compounds
• ISSN: 1618-2642; 1618-2650
• DOI: 10.1007/s00216-020-02881-5

Non-Hodgkin’s lymphoma (NHL) is a cancer of the lymphatic system where the lymphoid and hematopoietic tissues are infiltrated by malignant neoplasms of B, T, and natural killer lymphocytes. Effective and less invasive methods for NHL screening are urgently needed. Herein, we report an untargeted gas chromatography-mass spectrometry (GC-MS) method to investigate metabolic changes in non-volatile derivatized compounds from urine samples of NHL patients ( N  = 15) and compare them to healthy controls ( N  = 34). Uni- and multivariate data analysis showed 18 endogenous metabolites, including amino acids and their metabolites, sugars, small organic acids, and vitamins, as statistically significant for group differentiation. A receiver operating characteristic curve (ROC) generated from a support vector machine (SVM) algorithm-based model achieved 0.998 of predictive accuracy, displaying the potential and relevance of GC-MS-detected urinary non-volatile compounds for predictive purposes. Furthermore, a specific panel of key metabolites was also evaluated, showing similar results. All in all, our results indicate that this robust GC-MS method is an effective screening tool for NHL diagnosis and it is able to highlight different pathways of the disease. Graphical Abstract