Hand Motion Analysis Illustrates Differences When Drilling Cadaveric and Printed Temporal Bone.

OBJECTIVE: Temporal bone simulation is now commonly used to augment cadaveric education. Assessment of these tools is ongoing, with haptic modeling illustrating dissimilar motion patterns compared to cadaveric opportunities. This has the potential to result in maladaptive skill development. It is hypothesized that trainee drill motion patterns during printed model dissection may likewise demonstrate dissimilar hand motion patterns. METHODS: Resident surgeons dissected 3D-printed temporal bones generated from microCT data and cadaveric simulations. A magnetic position tracking system (TrakSTAR Ascension, Yarraville, Australia) captured drill position and orientation. Skill assessment included cortical mastoidectomy, thinning procedures (sigmoid sinus, dural plate, posterior canal wall) and facial recess development. Dissection was performed by 8 trainees (n = 5 < PGY3 > n = 3) using k-cos metrics to analyze drill strokes within position recordings. K-cos metrics define strokes by change in direction, providing metrics for stroke duration, curvature, and length. RESULTS: T-tests between models showed no significant difference in drill stroke frequency (cadaveric = 1.36/s, printed = 1.50/s, P < .40) but demonstrate significantly shorter duration (cadaveric = 0.37 s, printed = 0.16 s, P < .01) and a higher percentage of curved strokes (cadaveric = 31, printed = 67, P < .01) employed in printed bone dissection. Junior staff used a higher number of short strokes (junior = 0.54, senior = 0.38, P < .01) and higher percentage of curved strokes (junior = 35%, senior = 21%, P < .01). CONCLUSIONS: Significant differences in hand motions were present between simulations, however the significance is unclear. This may indicate that printed bone is not best positioned to be the principal training schema.

Generation of a 3D printed temporal bone model with internal fidelity and validation of the mechanical construct.

OBJECTIVE: To generate a rapid-prototyped temporal bone model from computed tomography (CT) data with a specific focus on internal anatomic fidelity. STUDY DESIGN: Research ethics board-approved prospective cohort study. SETTING: Current iterations of a rapid-prototyped temporal bone model are complicated by absent void spaces and inconsistent bone density due to limited infiltrant exposure. The creation of a high-fidelity model allows surgical trainees to practice in a standardized and reproducible training environment. This learning paradigm will significantly augment resident understanding of surgical approaches and techniques to prevent adverse outcomes. SUBJECTS AND METHODS: We describe a technique for generating internally accurate rapid-prototyped anatomical models with solid and hollow structures, including void spaces. The novel slicing algorithm digitally deconstructs a model into segments and permits removal of extraneous print material and allows infiltrant penetration of the entire bone structure. Precise reassembly is facilitated by digitally generated fiducials. Infiltrant of choice was determined by expert assessment and subjected to objective mechanical property assessment with comparison to cadaveric sheep bone. RESULTS: The printed bone models are highly realistic. Void space representation was excellent with 88% concordance between cadaveric bone and the resultant rapid-prototyped temporal bone model. Ultimately, cyanoacrylate with hydroquinone was determined to be the most appropriate infiltrant for both cortical and trabecular simulation. The mechanical properties of all tested infiltrants were similar to real bone. CONCLUSION: This model serves as an excellent replica of a human temporal bone for training and preoperative surgical rehearsal and can be dissected in a true-to-life fashion.