The work-flow of three-dimensional visualization of extended endoscopic endonasal approaches using frame extraction.

Endoscopic endonasal approaches (EEAs) to the skull base have revolutionized neurosurgical access, however, as the spectrum of extended EEAs continues to expand, traditional two-dimensional imaging remains insufficient for comprehensive anatomical understanding. Three-dimensional (3D) photorealistic models derived from endoscopic video recordings of extended EEAs, offer immersive visualization for education. The objective of the study is to describe a standardized step-by-step workflow for generating 3D models from endoscopic endonasal cadaveric dissections. EEAs were performed to the sellar, parasellar region, cavernous sinus, medial wall of orbit, pterygoid fossa, and jugular foramen on cadaveric heads, and the final anatomical exposures obtained through these dissections were subsequently captured and reconstructed in 3D using this standardized workflow. Five embalmed human cadaveric heads underwent stepwise EEAs at Microneurosurgery Skull Base Laboratory at Indiana University. Videos were recorded using both 0 and 30 degree endoscopes, with systematic frame extraction via FFmpeg, post-processing in Adobe Photoshop, and 3D reconstruction in Agisoft Metashape. Models were hosted on Sketchfab for interactive viewing. Photorealistic, life-size 3D models were produced with millions of polygons, preserving anatomical textures and geometry. Offline and online visualizations enable scalable applications without resolution loss. This workflow provides a reproducible framework that can be adopted by skull base laboratories, bridging endoscopy and 3D modeling, enhancing skull base education and simulation. Future integration with AI-driven segmentation could further automate labeling and analysis.