Evaluation of Software Tools for Segmentation of Temporal Bone Anatomy

Surgeons are increasingly relying on 3D medical image data for planning interventions. Virtual 3D models of intricate anatomy, such as that found within the temporal bone, have proven useful for surgical education, planning, and rehearsal, but such applications require segmentation of surgically relevant structures in the image data. Four publicly available software packages, ITK-SNAP, MITK, 3D Slicer, and Seg3D, were evaluated for their efficacy in segmenting temporal bone anatomy from CT and MR images to support patient-specific surgery simulation. No single application provided efficient means to segment every structure, but a combination of the tools evaluated enables creation of a complete virtual temporal bone model from raw image data with reasonably minimal effort.