TY  - JOUR
AU  - Rüttgers, Mario
AU  - Waldmann, Moritz
AU  - Hübenthal, Fabian
AU  - Vogt, Klaus
AU  - Tsubokura, Makoto
AU  - Lee, Sangseung
AU  - Lintermann, Andreas
TI  - Towards a widespread usage of computational fluid dynamics simulations for automated virtual nasal surgery planning
JO  - Future generation computer systems
VL  - 174
SN  - 0167-739X
CY  - Amsterdam [u.a.]
PB  - Elsevier Science
M1  - FZJ-2025-02797
SP  - 107935
PY  - 2025
AB  - Efficient computational approaches are crucial for advancing computational fluid dynamics (CFD)-based automated planning in nasal surgeries, such as septoplasties and turbinectomies. This study introduces a hybrid lattice-Boltzmann and level-set method to address the trade-off between computational cost and automation. By interpolating geometry variations in discrete steps between pre-surgical and target states, the approach achieves computational efficiency with only 21 surface variations per intervention. Previous methods rely on more costly coupling strategies, such as reinforcement learning or thermal modeling, which may still be appropriate for complex planning scenarios involving multiple intervention sites or thermal flow analysis. In contrast, the presented method reduces complexity while retaining key predictive capabilities, making it particularly suitable for widespread, time-sensitive clinical use focused on a single surgical intervention. Fluid mechanical metrics, including pressure loss and volume flow rate balance, are evaluated alongside tissue removal volume to recommend optimized surgical plans. Case studies on three patients demonstrate tissue savings of 12–25% without compromising key flow parameters. Additionally, a non-linear regression model trained on as few as 11 CFD simulations predicts pressure loss and flow rates with errors below 4%, and reduces computational costs by 50%. The proposed framework represents a significant step toward making CFD-based virtual nasal surgery planning more accessible and practical.
LB  - PUB:(DE-HGF)16
UR  - <Go to ISI:>//WOS:001512546300001
DO  - DOI:10.1016/j.future.2025.107935
UR  - https://juser.fz-juelich.de/record/1043189
ER  -