3DMed | Set of FE models for trunk orthotics
post-template-default,single,single-post,postid-16911,single-format-standard,ajax_fade,page_not_loaded,,qode-theme-ver-17.1,qode-theme-bridge,disabled_footer_bottom,qode_header_in_grid,wpb-js-composer js-comp-ver-5.5.5,vc_responsive

Set of FE models for trunk orthotics

Back spine pain

Set of FE models for trunk orthotics

The group of Mechanics of Materials and Structures of Ghent University is involved in WP1 and WP4 of the 3DMED project. Our main responsibility is to develop FE models for trunk orthotics, in particular for scoliosis braces. Scoliosis brace are orthotic devices used to stabilize and prevent the progress of mild spinal deformities observed in Adolescent Idiopathic Scoliosis (AIS). Such condition affects adolescents between 10 and 18 years old who typically have to wear a tailored made brace for full-time, except only for bathing and sporting. Currently there is a lack of evidence to support the design and prove the effectiveness of scoliosis braces. Typically, braces are designed by a Certified Prosthetist Orthotist (CPO) who is guided by general correction principles and empirical knowledge. There is no consensus or standards for the design of scoliosis braces and comfort is often neglected.

Finite Element (FE) models have been used to understand the biomechanical working principles of scoliosis braces and thus help to drive brace design towards more objective and effective treatments.  The aim of this demonstrator is to develop a methodology for topology optimization of scoliosis braces taking into account function, comfort and 3D printing technologies.

Our work in the last 8 months involved data collection and the development of preliminary patient-specific brace-torso FE models and 3D reconstruction methods of the upper trunk.  New contact was established with a Children’s Orthopedic Surgeon from the University Hospital of Ghent and together with Vigo we collected 5 datasets from patient’s with AIS. The datasets included multiview self-calibrated 2D x-ray images, body scans the reference scoliosis braces designed for each patient’s treatment. Preliminary brace-torso FE models were developed for a certain patient data provided by Vigo. Such model was a starting point for modelling brace installation over a patient’s torso. The complexity of the braces designed by Vigo lead us to develop a method, different from what can be found in the literature, for brace installation. Brace installation involved a number of steps including (S1) a preprocessing mapping step used to resolve the initial interpenetrations between brace and torso (S0), (S2) a contact step used to define brace-torso surfaces of contact, (S3) a strapping step to tighten the brace over the torso, (S4) a relaxation step of the displacements applied over the torso at the previous mapping step, (S5-S6)  the uplifting of the arm’s sections to fit the armpit pad in the intended place, following their relaxation (S7-8) and (S9) the relaxation of all boundary conditions applied to the brace but the strapping loads  (Fig. 1). The model was checked in terms of contact pressures, and stress levels over the brace. This preliminary model showed to be very challenging in terms of contact and computational times. The models were shared with M2i to help to improve the efficiency of the preliminary model foreseeing a possible problem for modelling more complex brace-trunk models required for topology optimization.

Fig. 1 -Main step of brace installation.


A 3D reconstruction method of the upper trunk is currently on going. Such reconstruction method uses as input the 2D frontal and lateral x-ray views of a patient’s upper trunk and a 3D reference model of the upper trunk of an healthy subject. The 3D shape and size of the patient’s trunk is approximated based on the use of landmark points equally defined between the 2D and 3D datasets, together with a  set of rigid and elastic registration functions (Fig. 2). However, in particular for the lateral view x-ray image it is often difficult to recognize the shapes of the vertebrae, due to the overlay of the arms over the thoracic spine. To overcome this problem we are currently working on the addition of parametric models to the reconstruction step to help in the reconstruction of the hard to see upper thoracic vertebrae based on statistical shape inferences.

Fig. 2 – Example of the ongoing 3D reconstruction method of the spine applied to L5-L3.


Next steps are: to continue with the 3D reconstruction of the trunk, and generate a complete model of a patient’s trunk ( including both soft and hard tissues)to model the effect of reference scoliosis braces over the spine and torso. Afterwards, shape and topology optimization methods will be developed for the design of functional, comfortable and 3D printed braces.

Cristiana Costa (Mechanics of Materials and Structures group, UGent)