Photogrammetric Construction of 3D Models of Skull Base Cadaveric Dissections: A Novel Tool for Anatomical Education and Surgical Planning

Background The difficulty of skull base surgery can only be improved upon by repeated and substantial practice with skull base anatomy. There is a need for new neuroanatomical teaching methods for residents and young neurosurgeons that allow for self-directed practice, outside of the operation room to better develop their visual-spatial skills. We propose the use of three-dimensional (3D) photogrammetric models as a convenient tool in teaching skull base anatomy and demonstrate its value here with dissected models of the posterior and infratemporal fossa.

Methods Two embalmed and injected human cadaveric heads were dissected, at the posterior fossa, and the infratemporal fossa, and one dry human skull for the general anatomy of the skull. Photos were taken with a Nikon D810 SLR (Tokyo, Japan) camera and turntable. The images were then uploaded and rendered into a 3D model using Agisoft PhotoScan Pro (St. Petersburg, Russia) software. This software superimposes the images and measures the intersecting tie points to triangulate the location of points in 3D space and create a 3D object. Strategic points of interests were labeled and annotated within the posterior and inferior temporal fossa specifically, as well as on the skull in general and a video recording was created.

Results An average of 200 photographs were taken of each specimen at four angles in 360 degrees. We achieved accurate 3D stereoscopic models of the dissected specimens at the infratemporal and posterior fossa, and the complete skull. Dense clouds were created from tie points (115,000 on average) 22,000,000 points for dissected specimens, and 32,500,000 for the skull. Visual improvements, namely texture adjustments and polishing, were made using Autodesk MAYA (San Rafael, United States), a 3D animation and modeling software. The overall results were models that upheld a high level of clarity and structural integrity.

Discussion We successfully produced 3D, anatomically correct, cadaveric dissections of the skull base. The use of photogrammetry in neurosurgery has primarily focused on monitoring of postoperative intracranial changes in volume. We demonstrate here that photogrammetric models of skull base structures are also valuable as teaching models for resident education and as a preoperative surgical planning tool for skull base procedures. This novel technique produces versatile models that can be used alone in a self-directed learning program, or further used to develop 3D video lessons, 3D print models, or incorporated into augmented or virtual reality learning modules.

Conclusion Photogrammetric models are an exceptional learning tool for skull base anatomy. The learner can rotate the model 360 degrees in all axes to become familiar with the orientation of structures and key surgical points. The software is easy to use and allows for a high degree of personalization, including addition of annotations, narrations, animations, and videos. Future research will continue to explore the use of volumetric models of the skull base as a tool for anatomical education and surgical planning.