Modernizing Geppetto's Workshop: Predicting Aesthetic Outcomes Of 3D-printed, Patient-specific Dorsal Nasal Implants Utilizing A Novel Imaging Protocol
Nicholas A. Vernice, A.B., Carly A. Askinas, B.S., Sabrina Shih, B.A., Xue Dong, M.D., Ph.D., Ryan J. Bender, B.S., James Shin, M.D., Jason A. Spector, M.D..
Weill Cornell Medicine, New York, NY, USA.
PURPOSE: To develop a low-cost, rapid protocol for the design and production of custom-fabricated dorsal nasal implants.
METHODS: For protocol validation, we utilized facial CT data to provide "ground truth" patient anatomy. Using these data, 3D models of the subject's skull and soft tissue were 3D-printed with polylactic acid (PLA) and cast in silicone, respectively (A). This "face phantom" was imaged to generate a 3D photograph utilizing commercially available photogrammetry software (B). Desired augmentation was determined via virtual deformation of this model. A corresponding, custom-designed dorsal nasal implant was 3D-printed in PLA, implanted on the phantom, and reimaged as above. To demonstrate fidelity, the photogrammetrically-derived model with and without augmentation was co-registered and compared to CT-derived "ground truth" (C-D).
RESULTS: Photogrammetric comparison between the 3D photograph and CT-derived "ground truth" revealed an average Hausdorff distance of 0.198 mm (95% 0.640 mm; Dice coefficient=0.989). Dorsal nasal augmentation revealed an average Hausdorff distance of 0.381 mm (95% 1.56 mm; Dice coefficient=0.978) compared to "ground truth." Comparison between expected and actual augmentation revealed an average Hausdorff distance of 0.276 mm (95% 1.24 mm; Dice coefficient=0.985). Heatmap analysis demonstrated high congruence in all relevant anatomical areas, with variation exclusively noted along the nasal dorsum as expected (D).
CONCLUSION: Our imaging protocol produces a highly accurate means of capturing critical facial anatomy necessary for design of custom-fabricated dorsal nasal implants.
Back to 2022 Posters