Introduction Clear delineation of intracranial aneurysms (IA) on angiographic working projections is essential for endovascular treatment to achieve successful occlusion and avoid complications. However, only little is known about what different interventional neuroradiologists (INRs) understand as optimal biplane IA working projections and how much the chosen working projections differ for a given anatomy.

Aim of Study The aim of our study is to gain a better understanding and objectification of biplane working projections, chosen for endovascular aneurysm treatment between different INRs, to develop an AI-assisted treatment projection optimization considering the possibility of improving head positioning.

Methods Five INRs used a self-developed software solution to simulate the best possible, IA working projection for 20 aneurysms. Due to the heterogeneity between the INRs, 17 biplane and 3 monoplane representative median projections were selected. In the next step, the usability of different viewing angels from 5° to 30° were rated as yes-no question.

Results The average proportion of agreement to remain usable per angle across all raters are as follows: 81.5% (5°), 66.4% (10°) down to 34.4% (30°).

Conclusion There is a wide variation between the raters for the initial treatment projection. Even if the angle of an IA projection differs by only 10°, already 33.6% of the resulting projections are considered as inappropriate compared to the primary selected projection. Up to a 5° angular change, the resulting IA projection is still feasible. Based on these results, an automated head repositioning suggestion can be considered in anatomically difficult conditions despite possible manual inaccuracies.

Disclosure of Interest Nothing to disclose.