Skip to content

EndoPredictor – Prediction system for complications after endovascular aortic repair based on geometric-biomechanical data

Artificial intelligence for predicting complications in aortic aneurysm treatments.

Aortic aneurysms, i.e. dangerous protrusions of the aorta, are among the most common vascular diseases and cause thousands of deaths every year in Europe and the USA alone. While open surgery involves more stress during the operation and a longer recovery time, endovascular treatment is associated with a number of complications after the operation. In endovascular treatment, a stent graft – a type of tube – is inserted through the blood vessel and placed directly at the aneurysm. This is intended to prevent the aneurysm from rupturing by relieving the pressure on the vessel at this point and allowing the blood to flow through the stent graft.

EndoPredictor

In the EndoPredictor project, researchers from the Department of Medical Informatics, together with physicians from the Kepler University Hospital Linz and MATTES Medical Imaging GmbH, developed methods that allow the properties of the abdominal aorta and the aneurysm to be extracted from medical image data and digitally mapped. For this purpose, 50 anonymized patient data sets consisting of computed tomography angiography (CT-A) images before the procedure and from several follow-up examinations after the endovascular treatment were used. The aortas and stent grafts were modeled from this image data, the blood flow through the stent graft was simulated and a possible change in the position and shape of the stent graft during the follow-up examination was calculated.

Automated prediction of complications 

A total of 201 CT-A image datasets were analyzed and 42 measures were calculated for each. These measures describe the shape of the aorta and stent-graft as well as properties obtained from the simulated blood flow. It was investigated which measures show statistical correlations with complications such as leaks, vasoconstrictions or vascular occlusions. These complications were predicted using a specially developed method. The result is a software system for the automated prediction of postoperative complications after endovascular treatments. A prediction accuracy of up to 88 % was achieved.

The prediction method developed is based on machine learning methods, automatically recognizes relevant features in the data and learns from previous data sets. A new method was implemented to validate the feature selection while simultaneously training the prediction system.

This project was supported by the strategic economic and research programme “Innovatives OÖ 2020” of the province of Upper Austria.

Project partners

Project details

  • Project short title: EndoPredictor
  • Projekt-Long title: Prediction system for complications after endovascular aortic repair based on geometric-biomechanical data
  • Funding call: Innovatives Oberösterreich 2020 – Ausschreibung 2015: Medizintechnik (Gesundheitswesen, alternde Gesellschaft)
  • Project partners:
    • RISC Software GmbH, Research Unit Medial Informatics (consortium management)
    • Kepler Universitätsklinikum Linz
    • MATTES Medical Imaging GmbH
  • Budget volume (total): EUR 356 thousand
  • thereof funding (total): EUR 177 thousand
  • Duration: 09/2015 – 08/2018

Contact person









    Project management

    Dr. Michael Giretzlehner

    Head of Unit Medical Informatics

    Read more