Project Summary
Although neuronavigation systems are of crucial assistance during cerebrovascular surgery, they do not
integrate hemodynamics information needed to treat complex cerebrovascular malformations. The present
project aims at developing an Augmented Reality (AR) neuronavigation tool that will enable the visualization of
cerebral hemodynamics information in the surgical view. Our long-term goal is to contribute toward the
development and clinical adoption of visualization tools that allow for safe and accurate treatment of
cerebrovascular malformations. Our overall objectives in this project are to (i) develop a novel approach based
on deep neural networks that can classify and reconstruct 3D dynamic cerebral vasculature from 2D Digital
Subtraction Angiography (DSA) image series, (ii) compose an AR visualization that will enhance the surgical
view of the brain, and (iii) validate and evaluate our technology in real clinical settings. The rationale for this
project is that such technology will provide a clear and interpretable visualization tool to surgeons that will
support their decision-making process and reduce the time and complex spatial reasoning required to treat
cerebrovascular malformations. To attain the overall objectives, the following three specific aims will be
pursued: 1) develop and validate a method to classify artery and veins in DSA image series to visually
disentangle AVMs, 2) develop and validate a method to build dynamic, virtual 3D model of cerebral vasculature
from pairs of DSA image series and 3) build an AR visualization that aligns DSA image series with the surgical
view and assess its impact providing surgical guidance. In addition, we will examine, through a clinical
retrospective study, and through tests in the operating room on phantom data, the impact of this visualization in
providing surgeons with guidance during cerebrovascular surgery. The proposed project is innovative because
it will be possible to merge the true DSA-derived 3D cerebral hemodynamics with images of the brain surface
seen through a surgical microscope. The proposed project is significant because it will provide visual guidance
and confirmation to surgeons that will facilitate decision-making in the surgical treatment of complex AVMs.
The results are expected to have an important positive impact because they will provide novel neuronavigation
tools to improve the surgical treatment of cerebrovascular malformations and ultimately reduce the risks of
intraoperative hemorrhaging and postsurgical deficits. Furthermore, the methods described here are
cost-effective, adapted to low-resources settings, and can be easily implemented on a large scale, bringing
advanced imaging techniques to far more patients.