Interlayer-aware postoperative facial appearance prediction in orthognathic surgery with bio-geometric guidance

Clinical Snapshot

Key Findings

• P Value: p < 0.05 versus all baseline comparators

• Effect Size: Statistically significant improvement over all baseline methods; specific effect sizes not reported in abstract

• Primary Outcome: Mean whole-face prediction error of 1.39 mm, below the 2 mm clinical acceptability threshold

• Nnt Or Sensitivity: Clinician acceptance rate: 94% (pragmatic clinical validation metric); no sensitivity/specificity or NNT reported as this is a predictive modelling study

• Confidence Interval: Not reported

Clinical Application

Clinical feasibility is currently limited. The model requires paired 3D imaging data (pre- and postoperative), specialised computational infrastructure, and integration into existing surgical planning workflows (e.g., Dolphin Imaging, ProPlan CMF). The 88-patient training set is insufficient for robust clinical deployment. Prospective validation in diverse populations and regulatory approval would be prerequisites for clinical adoption. Processing time and workflow integration requirements are not described. Orthognathic surgery in Australia is performed by oral and maxillofacial surgeons and is partially covered under Medicare for patients with documented functional indications (Item Numbers 45617–45638). The RACGP and ANZAOMS do not currently have specific guidelines on computational facial prediction tools. The TGA would classify such software as a medical device (Software as a Medical Device, SaMD) under the Therapeutic Goods Act 1989, requiring inclusion on the ARTG before clinical use. The Australian patient population's ethnic diversity (including significant East Asian, South Asian, Pacific Islander, and Indigenous Australian representation) means that a model trained exclusively on Chinese patients would require substantial external validation before Australian deployment. No PBS implications apply directly, though reduced surgical revision rates could have downstream cost benefits. Patients undergoing orthognathic surgery (skeletal malocclusion correction) where preoperative soft-tissue outcome prediction is used to support surgical planning and patient counselling. Most applicable to patients with significant dentofacial deformity requiring bimaxillary or single-jaw osteotomy.

Abstract

Objective.Accurate prediction of postoperative facial appearance is essential for orthognathic surgical planning, yet remains challenging due to the nonlinear biomechanical coupling between bone and soft tissue. While deep learning methods offer a faster alternative to traditional biomechanical simulation, they typically map bony displacements directly to facial surface deformation, overlooking the intervening soft-tissue layers with distinct biomechanical properties through which displacement is progressively transmitted. Moreover, the many-to-one effect of regional bony movements on each soft-tissue point remains insufficiently captured.Approach.We propose a novel network that introduces an interlayer mixture-of-experts mechanism to decouple bone-to-surface deformation propagation into three biomechanically inspired proxy representations at different tissue depth levels. Since the contribution of each depth level varies across individuals and facial regions, a gated routing network adaptively weights each layer's contribution, providing a data-driven approximation of spatially heterogeneous, patient-specific deformation transmission. Additionally, a bio-geometric convolution module captures regional bony influences through elastic-weighted neighborhood aggregation.Main results.Evaluations on 88 orthognathic surgery patients demonstrate that the proposed method achieves a mean whole-face error of 1.39 mm, below the 2 mm clinical threshold, with statistically significant improvements over all baselines (p<0.05) and the highest clinician acceptance rate (94%).Significance.By incorporating bio-geometric priors into a deep learning framework, our approach enables a more physically grounded and interpretable prediction paradigm, supporting efficient and clinically reliable surgical planning.

References

1. 1.Zhang, X., Bao, H., Zhong, J., Senhadji, L., Shu, H., Wu, J., Liu, L., & Yan, B. (2026). Interlayer-aware postoperative facial appearance prediction in orthognathic surgery with bio-geometric guidance. Physics in Medicine and Biology. https://doi.org/10.1088/1361-6560/ae752e