Towards navigation in endoscopic kidney surgery based on preoperative imaging.

Area of Science:

• Medical Imaging
• Surgical Technology
• Urology

Background:

• Endoscopic renal surgeries face high re-operation rates due to limited visualization.
• Surgeons struggle to mentally map preoperative CT scans to the endoscopic view, leading to missed anatomy.
• Incomplete navigation of the intrarenal collecting system contributes to missed stones and tumors, increasing recurrence.

Purpose of the Study:

• To develop and validate a guidance system for estimating endoscope position within CT scans during renal surgery.
• To reduce re-operation rates by improving intraoperative navigation and reducing the cognitive burden on surgeons.
• To create a 3D map integrating endoscopic video with preoperative imaging for enhanced surgical guidance.

Main Methods:

• Utilized Structure from Motion (SfM) to reconstruct the kidney's collecting system from endoscopic videos.
• Employed 3D U-Net for segmenting the intrarenal collecting system from CT scans to generate a 3D model.
• Registered the SfM-derived and CT-derived collecting system models to determine relative endoscope position.

Main Results:

• Successfully reconstructed intrarenal anatomy and accurately localized the endoscope's position.
• Demonstrated the creation of a 3D map integrating RGB endoscopic images with the reconstructed anatomy.
• Validated the reconstruction pipeline for surgical guidance, showing potential to reduce mental mapping burden.

Conclusions:

• The proposed guidance system aids in reconstructing and localizing intrarenal anatomy during endoscopic surgery.
• This technology can decrease the cognitive load on surgeons, facilitating better navigation.
• It represents a significant advancement towards reducing re-operation rates in kidney stone surgery.