Computer Aided Intracranial Aneurysm Treatment Based on 2D/3D Mapping, Virtual Deployment and Online Distal Marker Detection.

Purpose: To introduce a computational tool for peri-interventional intracranial aneurysm treatment guidance that maps preoperative planning information from simulation onto real-time X-Ray imaging.

Methods: Preoperatively, multiple flow diverter (FD) devices are simulated based on the 3D mesh of the vessel to treat, to choose the optimal size and location. In the peri-operative stage, this 3D information is aligned and mapped to the continuous 2D-X-Ray scan feed from the operating room. The current flow diverter position in the 3D model is estimated by automatically detecting the distal FD marker locations and mapping them to the treated vessel. This allows to visually assess the possible outcome of releasing the device at the current position, and compare it with the one chosen pre-operatively.

Results: The full pipeline was validated using retrospectively collected biplane images from four different patients (5 3D-DSA datasets in total). The distal FD marker detector obtained an average F1-score of 0.67 ( $\pm 0.224$ ) in 412 2D-X-Ray scans. After aligning 3D-DSA + 2D-X-Ray datasets, the average difference between simulated and deployed positions was 0.832 mm ( $\pm 0.521$ mm). Finally, we qualitatively show that the proposed approach is able to display the current location of the FD compared to their pre-operatively planned position.

Conclusions: The proposed method allows to support the FD deployment procedure by merging and presenting preoperative simulation information to the interventionists, aiding them to make more accurate and less risky decisions.