Machine Learning Analysis of the Cerebrovascular Thrombi Lipidome in Acute Ischemic Stroke.

OBJECTIVE

The aim of this study was to identify a signature lipid profile from cerebral thrombi in acute ischemic stroke (AIS) patients at the time of ictus. METHODS: We performed untargeted lipidomics analysis using liquid chromatography-mass spectrometry on cerebral thrombi taken from a nonprobability, convenience sampling of adult subjects (≥18 years old, n = 5) who underwent thrombectomy for acute cerebrovascular occlusion. The data were classified using random forest, a machine learning algorithm. RESULTS: The top 10 metabolites identified from the random forest analysis were of the glycerophospholipid species and fatty acids. CONCLUSION: Preliminary analysis demonstrates feasibility of identification of lipid metabolomic profiling in cerebral thrombi retrieved from AIS patients. Recent advances in omic methodologies enable lipidomic profiling, which may provide insight into the cellular metabolic pathophysiology caused by AIS. Understanding of lipidomic changes in AIS may illuminate specific metabolite and lipid pathways involved and further the potential to develop personalized preventive strategies.