AnurSim: Preoperative simulation of coiled aneurysms for their optimal clipping

Socio-economic and scientific context

An intracranial aneurysm (IA) is a deformation of the cerebral artery wall, leading to a fragile arterial zone that may rupture. Aneurysmal rupture most often causes a subarachnoid hemorrhage (SAH), which has a poor prognosis. It is estimated that one-third of patients with this condition will die, one-third will survive with a bad neurologic outcome and one-third will have a favorable outcome. IA is found in approximately 2–8% of the general population, with incidental detections increasing due to advancements and the growth of medical imaging. Aneurysmal rupture affects about 5,000 people per year in France. The historical treatment for IA is surgical clipping, which allows for aneurysm exclusion through an extravascular approach via craniotomy. The first case was described in 1937 by Dr. Walter Dandy, and the technique has evolved into a modern approach with the refinement of micro-instrumentation and the use of an operating microscope. The first treatments using coils guided by endovascular imaging in interventional neuroradiology (INR) were introduced in the early 1990s. Today, this method plays a predominant role in the preventive treatment of unruptured aneurysms and the management of ruptured, due to ISAT trial results. However, compared to surgical treatment, endovascular treatment has a higher rate of secondary aneurysm recanalization. A 2009 meta-analysis reported a recanalization rate of 20.8% over a follow-up period of 4.7 to 38 months. Although with improvements in embolization techniques and materials, the recanalization rate is likely lower today. The type of materials used, and the endovascular treatment technique are among the factors influencing recanalization. Secondary microsurgical treatment of previously embolized aneurysms presents additional challenges compared to primary microsurgical treatment. The presence of metallic coils within the aneurysmal sac creates mechanical constraints and limits wall deformation at the aneurysm neck, making clipping impossible in some cases without risk to the parent artery or its branches. In certain cases, it may be necessary to wait for increased recanalization before performing microsurgery, which exposes the patient to the risk of aneurysmal rupture and its potentially severe clinical consequences.

Working hypothesis and aims

Currently, 3D reconstruction of recanalized aneurysms is possible using data from angiographic imaging (3D Digital Subtraction Angiography (DSA)). However, this reconstruction has limitations due to variations in aneurysm sac morphology, branching patterns, wall thickness, and the rigidity and positioning of endovascular materials within the aneurysmal sac. Moreover, morphological measurements performed manually on these 3D reconstructions guide the therapeutic decision-making but can be significantly variable from one operator to the other. There is therefore a need to develop reliable automated approach allowing the optimal planning of clipping intervention for aneurysms recanalization. Several algorithms have been developed for this purpose, but none are currently specific to previously embolized aneurysms. Segmenting a recanalized coiled aneurysm can be complex and time-consuming, highlighting the need for the development of an automated segmentation tool, similar to those already available for untreated intracranial aneurysms. Moreover, the presence of metallic coils within the aneurysmal sac creates mechanical constraints and limits wall deformation at the aneurysm neck, sometimes making clipping impossible without risking damage to the parent artery or its branching vessels. There is there a need to integrate modelling approaches in order to better consider the wall deformation of the aneurysmal sac and better assist the surgeon about the optimal clipping strategy to perform.

Main milestones of the thesis

The doctoral thesis will therefore be divided into three main tasks: (1) Finalize the development of the automatic morphometry analysis of the aneurysmal sac based on a previous work, that has been performed during a previous Master’s internship, and validate the accuracy of the algorithms, (2) integrate Deep-Learning based finite element modelling to better consider wall deformation with an efficient computation time to be used during the planning phase, and (3) compare the results with 3D printing realistic models.

The PhD student will be hosted at the Laboratory of Medical Information Processing (LaTIM-UMR INSERM 1101). Born from the complementarity between health sciences and communication sciences, LaTIM (Laboratory of Medical Information Processing) conducts multidisciplinary research led by members of the University of Western Brittany (UBO), IMT Atlantique, INSERM, and Brest University Hospital (CHRU). Information is at the core of the unit’s research project; due to its multimodal, complex, heterogeneous, shared, and distributed nature, it is integrated by researchers into methodological solutions aimed at improving medical care. The student will therefore have access to its dedicated platforms, including PLACIS (https://latim.univ-brest.fr/platforms) and PLaTIMed (www.platimed.fr). The PLACIS platform will provide access to storage and high-performance computing solutions, essential for the development, training, and evaluation of algorithms designed for aneurysm analysis and optimal clipping calculation. On the other hand, the PLaTIMed platform will offer a realistic surgical environment, enabling initial concrete trials and testing of our approaches to assist the medical team in clip positioning. In addition to PLaTIMed, a collaboration with the W.Print platform of the CHU de Brest will also be consolidated to continue producing realistic 3D-printed models of recanalized aneurysms. This will allow us to simulate clipping procedures in a realistic manner and validate our approaches. Several 3D prints have already been carried out during a previous Master’s internship. It has been demonstrated that 3D printing facilitates clipping procedures and helps in selecting the optimal clip. 

Holding a Master’s degree (Master 2/Engineering School, BAC+5), the candidate must have strong skills in applied mathematics, statistical and deep learning, finite element modelling, and software development (C++ / Python), as well as a strong interest in the medical field. Proficiency in English is essential.