Towards Predicting Biomechanical Consequences of Jaw Reconstruction Ian Stavness, Alan G. Hannam, John E. Lloyd, and Sidney Fels Abstract— We are developing dynamic computer models of surgical jaw reconstructions in order to determine the effect of altered musculoskeletal structure on the biomechanics of mastication. We aim to predict post-reconstruction deficits in jaw motion and force production. To support these research goals we have extended our biomechanics simulation toolkit, ArtiSynth [1], with new methods relevant to surgical plan- ning. The principle features of ArtiSynth include simulation of constrained rigid-bodies, volume-preserving finite-element methods for deformable bodies, contact between bodies, and muscle models. We are adding model editing capabilities and muscle activation optimization to facilitate progress on post- surgical simulation. Our software and research directions are focused on upper-airway and cranio-facial anatomy, however the toolset and methodology are applicable to other muscu- loskeletal systems. I. INTRODUCTION Cancer of the oral cavity can invade the mandible and nearby muscles. Treatment usually involves resection of the cancerous tissue and may include large sections of the mandible as well as tongue and jaw muscles. Reconstruction of the affected area varies and usually includes a mandibular graft to restore aesthetics and function. The procedure results in a dramatic alteration of orofacial anatomy and even with reconstruction, the loss of bone mass and musculature can produce functional deficits in mastication, speech, tongue mobility, jaw mobility, and bite force generation [2] [3]. Analyzing jaw biomechanics is difficult due to the in- ability to measure muscle and articulation forces directly. Advancements in dynamic computer simulation of biome- chanics permit the analysis of the complex interplay of forces and motion in the oromandibular system [4] [5] [6]. The method has already been used to study unilateral chewing [7], and a recent model has included a dynamic hyolaryngeal component [8]. In the present study, we are using our Java-based software platform, ArtiSynth, for simulating inframandibular (jaw, larynx, and tongue) biomechanics [1] [6]. Figure 1 shows our dynamic inframandibular model. ArtiSynth provides a graphical interface for interactive changes to musculoskeletal structure and properties. As such, it is ideal for simulating chewing in reconstructed mandibles. We believe that in the near future, detailed examination of the biomechanics of surgically reconstructed anatomy I. Stavness, J. Lloyd, and S. Fels are with the Department of Electrical and Computer Engineering, The University of British Columbia, Van- couver, Canada stavness@ece.ubc.ca, lloyd@cs.ubc.ca, ssfels@ece.ubc.ca A. G. Hannam is with Department of Oral Health Sciences, Faculty of Dentistry, The University of British Columbia, Vancouver, Canada ahannam@interchange.ubc.ca Fig. 1. Dynamic Jaw-Tongue-Hyoid model in ArtiSynth. A property editing panel is shown for the selected muscle through computer simulation will provide pre-operative ben- efit in planning reconstruction procedures and post-operative benefit by guiding rehabilitation. Simulating a variety of potential reconstructions, e.g. different mandible grafts or tissue reattachments, may inform the patient-specific proce- dure to be performed, and complement other factors such as clinician experience and intuition. Given a model of a spe- cific reconstruction, simulation of motor tasks with different muscle drive patterns may illuminate new motor strategies to compensate for the altered musculoskeletal structure. Knowl- edge of such motor strategies could potentially guide post- operative rehabilitation in order to retrain patients regaining motor function. In this paper, we present preliminary results of simulating chewing deficits associated with mandibular reconstruction by means of our biomechanical modeling toolkit. II. THE BIOMECHANICS TOOLKIT;ARTI SYNTH Our approach was to use our existing jaw model and chewing simulation developed in ArtiSynth and to modify it using ArtiSynth’s model editing tools to be consistent with the typical outcomes of jaw reconstruction surgery. We performed simulations with the new jaw-reconstruction model using nominal muscle drive patterns for chewing and observed the resulting motion and force production deficits. Here we discuss the tools in ArtiSynth that we used for model creation, modification and simulation.