Related Experiment Video
Updated: May 28, 2026

Treatment of Facial Deformities using 3D Planning and Printing of Patient-Specific Implants
Published on: May 23, 2020
Comparison of Virtual and 3D-Printing Methods to Personalize Nasoseptal Flap Reconstruction
Beatrice Katsnelson1, Alyssa Glennon2, S Bulent Omay3,4
1Yale School of Medicine, New Haven, Connecticut, USA.
Objectives:
3D-printed models can be used to personalize the length of nasoseptal flap (NSF) reconstruction of anterior skull base defects to preserve more anterior nasal septal mucosa. The objective of this study is to evaluate the correlation between a previously described 3D-printed model method with an entirely virtual method of personalizing NSF length.
Methods:
Twenty-eight patients with anterior skull base pathology were included. CT scans were segmented to produce a cavity that would replicate the surgical changes associated with endoscopic endonasal skull base surgery. In 3D-printed models, NSF length was estimated with felt strips as previously described. For the virtual method, a curve was generated using the Materialise 3-Matic Curve function along the lateral sphenoid sinus walls and clival recess corridors. Estimations were made in a blinded fashion and comparisons of NSF lengths were made between virtual and 3D-printed methods.
Results:
Measurements between the 3D-printed and virtual methodologies were comparable. Mean virtual measurement was greater than the 3D-printed estimate by approximately 1 mm and was statistically significant (p 0.006) for the right lateral sphenoid wall route, but not for other routes. Linear regression analysis between the virtual and 3D-printed models demonstrated strong correlation coefficients for the right (0.85) and left (0.87) lateral sphenoid wall routes, and very strong correlation coefficients for the right (0.94) and left (0.93) clival recess comparisons.
Conclusion:
High correlation was demonstrated between the virtual and 3D-printed methods of NSF length estimation. Importantly, the virtual method does not generally underestimate NSF length compared to the 3D-printed method.

