Abstract Body: Blackground：Due to the narrow therapeutic time window and strict indications, there are still a large number of patients unable to receive thrombolysis and endovascular therapy, resulting in a poor long-term functional prognosis. Moreover, it is unclear what effective therapeutic targets as well as molecular systemic changes in stroke in the chronic phase. Therefore, to provide useful insights into its pathogenesis, identification of therapeutic targets and biomarker discovery, we analyze the proteomic profile of blood in ischemic stroke patients in order to gain insights into the disease.
Method: We use proteomics technologies to examine serum at baseline and use the modified Rankin Scale (mRS) to evaluate neurological function at baseline and 3-month follow-up after ischemic stroke. Based on data-independent acquisition mass spectrometry and parallel reaction monitoring (PRM) technology, we implement a staged discovery-confirmatory pipeline in serum samples from better prognosis patients(n=48) and poor prognosis patients(n=50) according to mRS to comprehensively identify and validate more accurate biomarker performance characteristics of candidate proteins.
Results: In this proteomic study, we characterize 4,740 circulating proteins and 324 differentially expressed proteins in 98 patients. We identify circulating proteomic signatures for ischemic stroke prognosis, and correlate these with clinical data to develop a proteomic profile of 14 markers. Based on machine learning algorithms such as LASSO regression and exhaustive feature selection (EFS), we finally identify some biomarkers (IQGAP2, LYZ, PDGFD, EHD3, CETP, FYB1) are validated as contributors to the ischemic stroke prognosis model through verification of an additional 58 serum samples. (AUC range:0.65-0.99).
Conclusion: We identify potential biomarkers and therapeutic targets and offer an opportunity for more precise prognosis and treatment in ischemic stroke.