Pre-interventional neurovascular analysis

Pre-interventional neurovascular analysis system based on MRA or CTA data sets

Summary

In current neurovascular clinical practice, patients often undergo an MRA or CTA scan for diagnostic purposes. To decide whether an interventional procedure such as coiling of aneurysms or stenting of stenoses in the neurovascular tree is feasible and to assess the dimensions of the devices that should be used, catheter angiography and 3D-rotational angiography, acquired during the intervention, is considered the most accurate imaging modality. However, in our experience over the last few years, the efficiency and outcome of neurovascular interventions can be improved significantly when accurate pre-operative measurements about the vessel sizes and morphology, vessel trajectory and pathology such as aneurysm volume and shape, are available without the use of invasive imaging. Unfortunately, comprehensive and robust analytical software is not yet commercially available.

In our group, an analysis application has been developed over the last few years called ‘CAMARA’, which was designed to perform measurements on CTA and MRA data sets to plan neurointerventional procedures. Clinical practice has demonstrated that based on such quantitative data the invasive pre-operative 3D angiography could be avoided in a number of cases. This is an important benefit given the potential risks of the invasive imaging procedures for the individual patient in comparison to the non-invasive MRA and CTA. The results from our first validation and feasibility study are very promising, and based on these experiences, a great need has arisen to further develop and extend the current software to a comprehensive analysis platform for planning various types of neurovascular interventions. Such approaches only have a chance of introduction into clinical practice if these are sufficiently fast, robust and accurate.

The goal of this project is to develop and validate a system for pre-operative planning of intra-cranial aneurysms from MRA and CTA data. The specific focus is on quantitative analysis of the cerebral vasculature, detection of pathologies, personalized selection of stents and seamless integration in the clinical workflow.

The project has a number of sub-goals:

  • Development of fully automated extraction of the cerebral tree from CTA and MRA datasets, and subsequent detection of vascular abnormalities.
  • The development of a comparison and classification metric for cerebral tree topology, and the construction of a database of cerebral tree topology types.
  • The integration of manufacturer specifications of mechanical stent properties in virtual stenting procedures, and quantitative validation by comparison of virtual stent deployment with actual stent deployment in the same patient.
  • The development of interactive visualization methods for fast and intuitive exploration of the cerebral tree and aneurysm characteristics.
  • Elaborate clinical testing and validation of the developed software in routine clinical care.

There is a great potential for the actual usage of the result of this project, i.e. the ready-to-use application for planning neurovascular interventions, in clinical practice. The prototype demo application CAMARA is already used by the neurointerventionalists in the LUMC, despite its limited functionality. Dr Brouwer has the CAMARA application on his laptop and uses the software at other hospitals (a.o. Toronto, Montreal, Sherbrook, Winnipeg, ErasmusMC, Australia) where he teaches flow diverter procedures to colleagues. The responses to the application are very encouraging and the neurointerventionalists recognize the value of a more general neurointerventional planning platform, that can be applied to multiple types of interventions. This expanding network of colleagues that are interested in the application ensures valuable feedback on prototype versions throughout this project.

Medis medical imaging systems B.V. has expressed interest in the outcome of this research. The core business of Medis is the development and commercialization of (cardio-) vascular applications. In this field the Division of Image Processing (LKEB) at the Department of Radiology, LUMC and Medis have a strong history of collaboration, which has led to a number of commercially available application packages for vascular structures, such as QAngio XA that is based on X-ray arteriography, QIvus for analyses of plaque burden and composition in intravascular pullback sequences acquired by ultrasound or OCT (optical coherence tomography) techniques,  and QPlaque, based on MR angiography and vessel wall images. However, the detection and measurements of aneurysms and stenoses are not incorporated in these programs since they are specific to the neurovascular field. The outcome of this research would be a useful supplement to the current Medis software and is therefore of great interest to Medis.

Project Team: