Hepatocarcinoma Discrimination by Ratiometric Lipid Profiles Using Tip-Contact Sampling/Ionization Mass Spectrometry

Precise diagnosis at the molecular level is essential for the improvement of surgery and treatment. High-throughput and spatial-resolved mass spectrometric (MS) methods for in situ detection of metabolites on tissue samples can reveal the dysregulation of metabolism in abnormal tissue and help identification of tumor. We here report a nondestructive MS method named as tip-contact sampling/ionization (TCSI)-MS technology which can quickly acquire lipidomic information from liver tissue and thereby realize tumor identification. Using this technology, fatty acids and lipids at the liver tissue surface can be rapidly imprinted onto a silicon nanowire tip attached with reduced graphene oxide (rGO) and sensitively detected by on-chip MS. With proper data pretreatment and statistical analysis, the clinical primary hepatocellular carcinoma (HCC) tissues can be discriminated from the nontumor parts. In addition, we found that a panel of adjacent dual peaks' ratio can be used to build a prediction model in artificial neural networks (ANN), resulting in high accuracy (91.7-98.3%) for tumor discrimination. Ratiometric TCSI-MS imaging using a selected dual peaks' ratio can greatly enhance the spatial resolution of tumor margin. The feature ratiometric data of lipid molecules may guide the study of metabolism pathways involved in hepatocarcinoma and ultimately become new metabolic biomarkers in clinical diagnosis. The present work demonstrated that the TCSI-MS technology may pave a novel way for surgery guidance and precision diagnosis in tissue biopsy.