Omics analysis and automated clot detection for thrombus histology in acute ischemic strokes using high performance computing AI systems.

Abstract:

The overall objective of the project is to develop innovative and accurate tools based on radiomics and metabolomics to assess prognosis and best therapeutic strategy for Acute Ischemic Stroke (AIS). The framework will be based on 1) automated clot detection at in hospital Computed Tomography Angiography (CTA), 2) extraction and evaluation of a radiomic signature for clots composition related to high or low probability of intracranial vessel recanalization 3) identification of a metabolomic signature of intracranial Large Vessel Occlusion (LVO) 4) identification of a metabolomic signature associated to the probability of recanalization 5) association between clot histology and radiomics/metabolomics profile.

The project will leverage on the setup of a Biobank in acute stroke (a worldwide rare asset). The stroke biobank will be further exploited for storing vascular clot samples after Mechanical Thrombectomy (MT), and radiomics features from CTA which could be used to train machine learning models to predict the histo-pathological characteristics of the clot and the probability of recanalization.

The final output of the project is a radiomic and metabolomic based prediction of histological clot composition and patient’ outcome in patients with AIS. Despite the high feasibility of the project, it has high innovation potential, aiming to deliver a new diagnostic stroke model towards a more personalized medicine for AIS management from its earliest stages.

Given the tremendous social and economic burden of stroke survivors (2% of population), the new “omics” findings would help in the identification of new insights about LVO clot composition and may help in the identification of new strategies to optimize both MT therapies in the acute phase and implement secondary prevention of AIS, with an economical gain for the healthcare systems.