Geometric Morphometric-Based Estimation as an Aid for Naso-Orbito-Ethmoidal Trauma Management

Area of Science:

• Craniofacial surgery
• Medical imaging
• Computer-aided design

Background:

• Naso-orbito-ethmoid (NOE) fractures pose reconstruction challenges, often leading to re-deformities.
• Restoring pretraumatic facial form is critical but difficult.

Purpose of the Study:

• To evaluate computer-aided methods for estimating pretraumatic facial form in silico.
• To compare the accuracy of thin-plate spline (TPS) and multivariate regression for facial landmark estimation.

Main Methods:

• Utilized 3D facial images of adult Caucasians.
• Applied geometric morphometric methods (TPS and regression) to estimate 6 facial landmarks based on 18 others.
• Analyzed facial form changes using principal component analysis and permutation testing.

Main Results:

• Thin-plate spline (TPS) demonstrated superior accuracy in landmark estimation.
• Regression-based methods introduced statistically significant artificial deformities.
• Mean estimation error for TPS was highest at the nasion (up to 3.34±1.7 mm).

Conclusions:

• TPS-based estimation shows promising accuracy for reconstructive surgery.
• The computer-assisted pipeline is time-feasible and can generate 3D-printable surgical aids.
• Further clinical translation and replication are needed to validate these in silico findings.