Patient-specific, low-radiation Training setup for aortic aneurysm repair

• Applicants: Professor Dr. Bernhard Dorweiler; Anirban Mukhopadhyay, Ph.D.; Professor Dr.-Ing. Robert Heinrich Schmitt
• Subject Area: Cardiac and Vascular Surgery ; Image and Language Processing, Computer Graphics and Visualisation, Human Computer Interaction, Ubiquitous and Wearable Computing
• Term: since 2025
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 551593073

The most common treatment strategy for patients diagnosed with aortic abdominal aneurysms is an elective minimally invasive aortic repair procedure called EndoVascular Aortic Repair (EVAR) stent grafting. In this procedure, a stent is placed in the aneurysm to close it through stopping the blood flow and therefore clotting the blood inside the aneurysm. Training and performance of EVAR involves a high radiation dose being posed on the patients and the surgeons. Therefore, this proposal focuses on a low radiation training procedure and intervention for EVAR. The conventional fluoroscopy in EVAR is supplemented by electromagnetic tracking to minimize the use of X-ray imaging. For training the EVAR procedure as close to the patient’s anatomy as possible, 3D printed models of the patient’s aorta are used. During training, every step carried out by the surgeon or resident is filmed. To avoid using X-ray images for training, those video frames are used to generate artificial fluoroscopic images with Generative Adversarial Networks (GAN). To save time for experienced surgeons when evaluating the filmed training, machine learning is applied for an automated evaluation of the performance of the trained surgeon. The radiation-free training process and the low radiation operation process are evaluated concerning the risk they pose on the surgeon as well as the patient to ensure an increased safety compared to the conventional operation and training.

• DFG Programme: Research Grants