Accurate Discrimination of Benign Biliary Diseases and Cholangiocarcinoma with Serum Multiomics Revealed by High-Throughput Nanoassisted Laser Desorption Ionization Mass Spectrometry

Cholangiocarcinoma (CCA) is an aggressive malignant tumorwitha poor prognosis. Carbohydrate antigen 19-9 is an essential biomarkerfor CCA diagnosis, but its low sensitivity (72%) makes the diagnosisunreliable. To explore potential biomarkers for the diagnosis of CCA,a high-throughput nanoassisted laser desorption ionization mass spectrometrytechnique was constructed. We performed serum lipidomics and peptidomicsanalyses from 112 patients with CCA and 123 patients with benign biliarydiseases. Lipidomics analysis showed that various lipids, such asglycerophospholipids, glycerides, and sphingolipids, were perturbed.Peptidomics analysis revealed perturbations of multiple proteins involvedin the coagulation cascade, lipid transport, and so on. After datamining, 25 characteristic molecules including 20 lipids and 5 peptideswere identified as potential diagnostic biomarkers. After screeningvarious machine learning algorithms, artificial neural network wasselected to construct a multiomics model for CCA diagnosis with 96.5%sensitivity and 96.4% specificity. The sensitivity and specificityof the model in the independent test cohort were 93.8 and 87.5%, respectively.Furthermore, integrated analysis with transcriptomic data in the cancergenome atlas confirmed that genes altered in CCA significantly affectedmultiple lipid- and protein-related pathways. Data are available viaMetaboLights with the identifier MTBLS6712.

Automated summary

Cholangiocarcinoma is a malignant tumor with poor prognosis. A high-throughput nanoassisted laser desorption ionization mass spectrometry technique was constructed to explore potential biomarkers for CCA diagnosis. Lipidomics and peptidomics analyses identified 25 characteristic molecules as potential diagnostic biomarkers. An artificial neural network model achieved high sensitivity and specificity for CCA diagnosis. Integrated analysis confirmed the significant impact of genes altered in CCA on lipid and protein-related pathways. Data are available in the MetaboLights database with the identifier MTBLS6712.