Related Experiment Video
Updated: Jun 16, 2025
09:53
Role of Diffusion MRI Tractography in Endoscopic Endonasal Skull Base Surgery
Published on: July 5, 2021
3.6K
3D reconstruction in endonasal pituitary surgery
Dannielle Lee1,2, Laurent Mennillo3,4, Emalee Burrows1,5
1UCL Hawkes Institute, University College London, London, UK.
Summary
This study introduces a 3D reconstruction pipeline for pituitary surgery using endoscopic videos. The method enhances visualization and navigation, showing potential for augmented reality (AR)-guided procedures.
Area of Science:
- Neurosurgery
- Computer Vision
- Medical Imaging
Background:
- Endoscopic transsphenoidal surgery for pituitary tumors faces challenges with limited visibility and maneuverability.
- The narrow nasal corridor increases the risk of surgical complications.
Purpose of the Study:
- To develop a 3D reconstruction pipeline from monocular endoscopic videos for enhanced intraoperative visualization and navigation.
- To address visibility and maneuverability limitations in pituitary surgery.
Main Methods:
- A user study with trainee surgeons used 3D printed phantom devices.
- Learned feature detectors and matchers were employed to extract information from textureless surfaces.
- COLMAP was used for feature reconstruction, and the iterative closest point algorithm evaluated surface accuracy against CAD models.
Main Results:
- Accurate 3D reconstructions were achieved with moderate variability, even with blur or occlusions.
- The best methods achieved average RMSE values of 0.33 mm and 0.41 mm.
- Dense Kernelized Feature Matching showed higher computation time compared to SuperPoint with LightGlue.
Conclusions:
- The pipeline accurately reconstructs 3D models of surgical anatomy.
- Learned feature detectors and matchers show potential for real-time AR-guided pituitary surgery.
