Thibaut Bardyn

Atlas-based segmentation of brain tumor images using a Markov Random Field-based tumor growth model and non-rigid registration

We propose a new and clinically oriented approach to perform atlas-based segmentation of brain tumor images. A mesh-free method is used to model tumor-induced soft tissue deformations in a healthy brain atlas image with subsequent registration of the modified atlas to a pathologic patient image. The atlas is seeded with a tumor position prior and tumor growth simulating the tumor mass effect is performed with the aim of improving the registration accuracy in case of patients with space-occupying lesions. We perform tests on 2D axial slices of five different patient data sets and show that the approach gives good results for the segmentation of white matter, grey matter, cerebrospinal fluid and the tumor.

Prediction of dental implant torque with a fast and automatic finite element analysis: a pilot study

OBJECTIVES Despite its importance, implant removal torque can be assessed at present only after implantation. This paper presents a new technique to help clinicians preoperatively evaluate implant stability. STUDY DESIGN Planning software has been combined with an in-house finite element solver. Once the clinician has chosen the implant position on the planner, a finite element analysis automatically calculates the primary stability. The process was designed to be as simple and fast as possible for clinical use. This paper describes application of the method to the prediction of removal torque. A preliminary validation has been performed in both polyurethane foam and sheep bone. RESULTS The predicted torque is quantitatively equivalent to experimental values with correlation coefficients of >0.7 in both materials. CONCLUSIONS This preliminary study is a first step toward the introduction of finite element models in computer-assisted surgery. The fact that the process is fast and automatic makes it suitable for a clinical use.

Influence of Smoothing on Voxel-Based Mesh Accuracy in Micro-Finite Element

The interest in automatic volume meshing for finite element analysis (FEA) has grown more since the appearance of microfocus CT (μCT), due to its high resolution, which allows for the assessment of mechanical behaviour at a high precision. Nevertheless, the basic meshing approach of generating one hexahedron per voxel produces jagged edges. To prevent this effect, smoothing algorithms have been introduced to enhance the topology of the mesh. However, whether smoothing also improves the accuracy of voxel-based meshes in clinical applications is still under question. There is a trade-off between smoothing and quality of elements in the mesh. Distorted elements may be produced by excessive smoothing and reduce accuracy of the mesh. In the present work, influence of smoothing on the accuracy of voxel-based meshes in micro-FE was assessed. An accurate 3D model of a trabecular structure with known apparent mechanical properties was used as a reference model. Virtual CT scans of this reference model (with resolutions of 16, 32 and 64 μm) were then created and used to build voxel-based meshes of the microarchitecture. Effects of smoothing on the apparent mechanical properties of the voxel-based meshes as compared to the reference model were evaluated. Apparent Young’s moduli of the smooth voxel-based mesh were significantly closer to those of the reference model for the 16 and 32 μm resolutions. Improvements were not significant for the 64 μm, due to loss of trabecular connectivity in the model. This study shows that smoothing offers a real benefit to voxel-based meshes used in micro-FE. It might also broaden voxel-based meshing to other biomechanical domains where it was not used previously due to lack of accuracy. As an example, this work will be used in the framework of the European project ContraCancrum, which aims at providing a patient-specific simulation of tumour development in brain and lungs for oncologists. For this type of clinical application, such a fast, automatic, and accurate generation of the mesh is of great benefit.

Computer-assisted ankle joint arthroplasty using bio-engineered autografts

Bio-engineered cartilage has made substantial progress over the last years. Preciously few cases, however, are known where patients were actually able to benefit from these developments. In orthopaedic surgery, there are two major obstacles between in-vitro cartilage engineering and its clinical application: successful integration of an autologuous graft into a joint and the high cost of individually manufactured implants. Computer Assisted Surgery techniques can potentially address both issues at once by simplifying the therapy, allowing pre-fabrication of bone grafts according to a shape model, individual operation planning based on CT images and providing optimal accuracy during the intervention. A pilot study was conducted for the ankle joint, comprising a simplified rotational symmetric bone surface model, a dedicated planning software and a complete cycle of treatment on one cadaveric human foot. The outcome was analysed using CT and MRI images; the post-operative CT was further segmented and registered with the implant shape to prove the feasibility of computer assisted arthroplasty using bio-engineered autografts.

Computer-assisted arthroplasty using bio-engineered autografts

Recent advances in tissue-engineered cartilage open the door to new clinical treatments of joint lesions. Common to all therapies with in-vitro engineered autografts is the need for optimal fit of the construct, to allow screwless implantation and optimal integration into the live joint. Computer Assisted Surgery techniques are prime candidates to ensure the required accuracy while at the same time simplifying the procedure. A pilot study has been conducted aiming at assembling a new set of methods to support ankle joint arthroplasty using bio-engineered autografts. Computer assistance allows planning of the implant shape on a CT image, manufacturing the construct according to the plan and interoperatively navigating the surgical tools for implantation. A rotational symmetric model of the joint surface was used to avoid segmentation of the CT image; a new software was developed to determine the joint axis and make the implant shape parameterisable. A complete cycle of treatment from planning to operation was conducted on a human cadaveric foot, thus proving the feasibility of computer-assisted arthroplasty using bio-engineered autografts.

Automatic Finite Element Mesh Generation and Correction from 3D Image Data

In the framework of the ContraCancrum project, automatic generation of an accurate finite element mesh is necessary for minimum user-interaction. The interest on automatic volume meshing for finite element (FE) has grown more popular since the apparition of microfocus CT (μCT) due to its high resolution, allowing assessment of mechanical behavior at a high precision. However, the basic meshing approach of generating one hexahedron per voxel produces jagged edges. The Laplacian operator can be used to smooth the generated mesh, but this method produces mesh shrinkage and volume changes. In this paper an automatic meshing and smoothing algorithm for FE meshes from 3D image data is presented. The method includes a regularization step to assure good element’s shape based on a quality measure. The algorithm introduces a novel technique to combine hexahedron and prism elements in order to increase the degree of mesh smoothness while maintaining good quality of elements. The smoothing method is based on low-pass signal filtering using transfer functions approximated by Chebyshev polynomials, resulting in a fast and computationally efficient method being extended here for FE meshes. The smoothing process was evaluated on various data based on the quality of the elements after smoothing, and stress distribution.

In-vitro characterization of an implantable thermal flow sensor for hydrocephalus

An implantable thermal flow sen treatment of hydrocephalus has been developed.The sensor uses passive telemetry at 13.56 MHz for power supply and read out of the measured flow rate. The in vitro performance of the sensor has been characterized using artificial Cerebrospinal Fluid (CSF) with increased protein concentration and artificial CSF with 10% fresh blood. No drift could be observed in the flow sensor measurement which could be associated to a deposition of proteins at the sensitive surface walls of the packaged flow sensor.