Erwan Jolivet

An anatomical subject-specific FE-model for hip fracture load prediction

In order to reduce the socio-economic burden induced by osteoporotic hip fractures, finite element models have been evaluated as an additional diagnostic tool for fracture prediction. For a future clinical application, the challenge is to reach the best compromise between model relevance and computing time. Based on this consideration, the current study focused on the development and validation of a subject-specific FE-model using an original parameterised generic model and a specific personalization method. A total of 39 human femurs were tested to failure under a quasi-static compression in stance configuration. The corresponding FE-models were generated and for each specimen the numerical fracture load (F FEM) was compared with the experimental value (F EXP), resulting in a significant correlation (F EXP = 1.006 F FEM with r 2 = 0.87 and SEE = 1220 N, p < 0.05) obtained with a reasonable computing time (30 mn). Further in vivo study should confirm the ability of this FE-model to improve the fracture risk prediction.

Volumic patient-specific reconstruction of muscular system based on a reduced dataset of medical images.

Three-dimensional mechanical modelling of muscles is essential for various biomechanical applications and clinical evaluation, but it requires a tedious manual processing of numerous images. A muscle reconstruction method is presented based on a reduced set of images to generate an approximate parametric object from basic dimensions of muscle contours. A regular volumic mesh is constructed based on this parametric object. The approximate object and the corresponding mesh are deformed to fit the exact muscles contours yielding patient-specific geometry. Evaluation was performed by comparison of geometry to that obtained by contouring all computed tomography (CT) slices, and by quantification of the mesh quality criteria. Muscle fatty infiltration was estimated using a threshold between fat and muscle. Volumic fat index (VFI) of a muscle was computed using first all the complete CT scan slices containing the muscle (VFIref) and a second time only the slices used for reconstruction (VFIrecons). Mean volume error estimation was 2.6% and hexahedron meshes fulfilled quality criteria. VFIrecons respect the individual variation of fat content.

Three-dimensional stereoradiographic modeling of rib cage before and after spinal growing rod procedures in early-onset scoliosis

BACKGROUND Early-onset scoliosis frequently leads to major thoracic deformity and pulmonary restrictive disease. Growing rods surgical techniques were developed to achieve a satisfactory correction of the spinal curves during growth. The effect on the rib cage deformity has not yet been documented. The purpose of this study was to analyze the changes of the thoracic geometry after implantation of a growing rod, and to evaluate a stereoradiographic reconstruction method among young scoliotic patients. METHODS Four patients were enrolled in the study, and four additional patients in the reproducibility study. Three-dimensional spine and rib cage models were generated after low-dose stereoradiographic imaging (EOS). Three-dimensional parameters were computed before and after surgery. Intra and inter-observer reproducibility was calculated, and the accuracy was assessed in comparison to volumetric CT-scan. FINDINGS The average Cobb angle was reduced from 50.8 degrees to 26 degrees . The surgery resulted in a complex 3D effect on the rib cage, combining frontal, lateral, and axial rotation. This effect was dependent of the side (concave or convex), and the position relative to the apical vertebra. Mean errors in comparison to CT-scan were 3.5mm. INTERPRETATION The results on the spinal deformity are comparable to other series. The effect on the rib cage is of a smaller magnitude than in the case of a spinal arthrodesis. A longer follow-up is necessary to confirm the positive effect on the rib cage deformity. Further research should be performed to improve the reproducibility of 3D parameters.

Subject-specific musculoskeletal model of the lower limb in a lying and standing position

Accurate estimation of joint loads implies using subject-specific musculoskeletal models. Moreover, as the lines of action of the muscles are dictated by the soft tissues, which are in turn influenced by gravitational forces, we developed a method to build subject-specific models of the lower limb in a functional standing position. Bones and skin envelope were obtained in a standing position, whereas muscles and a set of bony landmarks were obtained from conventional magnetic resonance images in a lying position. These muscles were merged with the subject-specific skeletal model using a nonlinear transformation, taking into account soft tissue movements and gravitational effects. Seven asymptomatic lower limbs were modelled using this method, and results showed realistic deformations. Comparing the subject-specific skeletal model to a scaled reference model rendered differences in terms of muscle length up to 4% and in terms of moment arm for adductor muscles up to 30%. These preliminary findings enlightened the importance of subject-specific modelling in a functional position.

New method for 3D reconstruction of the human cranial vault from CT-scan data

This study presents a new method for the 3D reconstruction of the human cranial vault from routine Computed Tomography (CT) data. The reconstruction method was based on the conceptualization of the shape of the cranial vault with a parametric description. An initialization was first realized with the identification of anatomical landmarks and contours on Digitally Reconstructed Radiographs (DRR) in order to obtain a pre-personalized reconstruction. Then an optimization of the reconstruction was performed to segment the internal and external surfaces of the cranial vault for thickness computation. The method was validated by comparing final reconstructions issued from our approach and from a manual slice-by-slice segmentation method on ten CT-scans. Errors were comparable to the CT image resolution, and less than 2 min were dedicated to the operator-dependent marking step. The reconstruction of internal and external surfaces of the cranial vault allows quantifying and visualizing of thickness throughout the cranial vault. This thickness mapping is useful for clinical purposes as additional pre-surgical information. Moreover, this study constitutes a first step in the personalized characterization of skull resistance directly from routine exams.

CT-based semi-automatic quantification of vertebral fracture restoration

Minimally invasive surgeries aiming to restore fractured vertebral body are increasing; therefore, our goals were to create a 3D vertebra reconstruction process and design clinical indices to assess the vertebral restoration in terms of heights, angles and volumes. Based on computed tomography (CT)-scan of the vertebral spine, a 3D reconstruction method as well as relevant clinical indices were developed. First, a vertebra initial solution requiring 5 min of manual adjustments is built. Then an image processing algorithm places this solution in the CT-scan images volume to adjust the models nodes. On the vertebral bodys anterior and posterior parts, nine robust heights, volume and endplate angle measurement methods were developed. These parameters were evaluated by reproducibility and accuracy studies. The vertebral body reconstruction accuracy was 1.0 mm; heights and volume accuracy were, respectively, 1.2 and 179 mm3. In conclusion, a 3D vertebra reconstruction process requiring little user time was proposed as well as 3D clinical indices assessing fractured and restored vertebra.

Distribution and variability study of the femur cortical thickness from computer tomography

In the context of patient-specific 3D bone reconstruction, enhancing the surface with cortical thickness (COT) opens a large field of applications for research and medicine. This functionality calls for database analysis for better knowledge of COT. Our study provides a new approach to reconstruct 3D internal and external cortical surfaces from computer tomography (CT) scans and analyses COT distribution and variability on a set of asymptomatic femurs. The reconstruction method relies on a short (∼5 min) initialisation phase based on 3D reconstruction from biplanar CT-based virtual X-rays and an automatic optimisation phase based on intensity-based cortical structure detection in the CT volume, the COT being the distance between internal and external cortical surfaces. Surfaces and COT show root mean square reconstruction errors below 1 and 1.3 mm. Descriptions of the COT distributions by anatomical regions are provided and principal component analysis has been applied. The first mode, 16–50% of the variance, corresponds to the variation of the mean COT around its averaged shape; the second mode, 9–28%, corresponds to a fine variation of its shape. A femur COT model can, therefore, be described as the averaged COT distribution in which the first parameter adjusts its mean value and a second parameter adjusts its shape.

Prediction of the vertebral strength using a subject-specific finite-element model

Given the ageing population, osteoporosis and vertebral fractures are considered as a major public health problem. One key issue is to predict the vertebral strength. Mechanical approaches based on finite-element models (FEM) have been proposed in vitro (Buckley et al. 2007). However, these models, based on quantitative computed tomography (QCT), may be restricted for in vivo analyses of the whole spine because of the high radiation dose. The EOS system (Biospace Med, Paris, France) takes simultaneously bi-planar perpendicular head-to-feet radiographs, in standing position, using low-dose X-rays detectors. Based on previous works, a fast and accurate 3D reconstruction of the spine is available (Pomero et al. 2004; Humbert et al. forthcoming). Moreover, using a dual-energy modality, in vitro accurate bone mineral densities (BMD) can be assessed (Sapin et al. 2008) and their relationships with the mechanical properties of the vertebral cancellous bone have been established (Sapin et al. 2009). Hence, the present study proposes subject-specific finite-element models based low-dose imaging investigations.

Subject specific 3D lower limbs muscle reconstructions for children with cerebral palsy

Cerebral palsy (CP) results from a lesion in the immature brain. It results in abnormalities of muscle strength and tone (spasticity), and joint movement. Soft tissue treatments and surgeries concern in general botulinum toxin injection, muscle lengthening, tendon transfer, etc. Information on musculo-skeletal geometry is then necessary for surgery planning (Figures 1 and 2). Three-dimensional (3D) skeletal reconstructions for lower limbs are possible using stereoradiography technique (Jolivet et al. 2008). The aim of this study is to assess specific subject 3D muscles reconstructions for children with CP.