Portal vein thrombosis (PVT) is one of the most common vascular disorders of the liver with significant morbidity and mortality. It consists of a blockage of the portal vein (the vessel that brings blood to the liver from the intestines) by a blood clot. This may cause chronic portal hypertension if recanalization is not obtained. Certain characteristics, such as the time since the onset of the thrombus (acute or chronic), the presence of collateral vessels, and other imaging aspects, are important components to decide the best therapeutic strategy. Up to 40% of cases of acute PVT show recanalization with early anticoagulation therapy. However, response to anticoagulation and long-term prognosis of PVT are not well-defined. 

The aim of this project is to identify imaging features (obtained from computer tomography scans) that along with clinical factors may predict the response of a thrombus to treatment. For this, radiomics (imaging) features and machine learning methods will be used to combine imaging and clinical data to better characterize thrombus and predict those that will respond to treatment. This project will be done in collaboration with Hospital Clínic de Barcelona.