In Clinica chimica acta; international journal of clinical chemistry
BACKGROUND AND AIMS : The vital metabolic signatures for IA risk stratification and its potential biological underpinnings remain elusive. Our study aimed to develop an early diagnosis model and rupture classification model by analyzing plasma metabolic profiles of IA patients.
MATERIALS AND METHODS : Plasma samples from a cohort of 105 participants, including 75 IA patients in unruptured and ruptured status (UIA, RIA) and 30 control participants were collected for comprehensive metabolic evaluation using ultra-high-performance liquid chromatography-mass spectrometry-based pseudotargeted metabolomics method. Furthermore, an integrated machine learning strategy based on LASSO, random forest and logistic regression were used for feature selection and model construction.
RESULTS : The metabolic profiling disturbed significantly in UIA and RIA patients. Notably, adenosine content was significantly downregulated in UIA, and various glycine-conjugated secondary bile acids were decreased in RIA patients. Enriched KEGG pathways included glutathione metabolism and bile acid metabolism. Two sets of biomarker panels were defined to discriminate IA and its rupture with the area under receiver operating characteristic curve of 0.843 and 0.929 on the validation sets, respectively.
CONCLUSIONS : The present study could contribute to a better understanding of IA etiopathogenesis and facilitate discovery of new therapeutic targets. The metabolite panels may serve as potential non-invasive diagnostic and risk stratification tool for IA.
Sun Kaijian, Zhang Xin, Li Xin, Li Xifeng, Su Shixing, Luo Yunhao, Tian Hao, Zeng Meiqin, Wang Cheng, Xie Yugu, Zhang Nan, Cao Ying, Zhu Zhaohua, Ni Qianlin, Liu Wenchao, Xia Fangbo, He Xuying, Shi Zunji, Duan Chuanzhi, Sun Haitao
2022-Nov-05
biomarkers, intracranial aneurysm, machine learning, mass spectrometry, metabolomics
