DVC allowed measurements of displacement and strain distribution in bone through the comparison of two, or more, 3D images. Hence, it has a potential as a diagnostic tool in combination with clinical CT. Currently, traditional computed tomography (CT) allows for a detailed 3D analysis of hard tissues, but imaging in a weight-bearing condition is still limited. PedCAT-CT (Curvebeam, USA) emerged as a novel technology allowing, for the first time, 3D imaging under full-weight bearing (Richter, Zech et al. 2015). Specifically, a PedCAT-CT based DVC was employed to establish its reliability through the strain uncertainties produced on bone structure targets, preliminarily to any further clinical studies. In addition, a reverse engineering FE modeling was used to predict possible force associated to displacement errors from DVC.

Three porcine thoracic vertebrae were used as bone benchmark for the DVC (Palanca, Tozzi et al. 2016, Tozzi, Dall'Ara et al. 2016). The choice of using porcine vertebrae (in a CT designed for foot/ankle) was driven by availability, as well as similar dimensions to the calcaneus. Each vertebra was immersed in saline solution and scanned twice without any repositioning (zero-strain-test) with a pedCAT-CT (Curvebeam, USA) obtaining an isotropic voxel size of 370 micrometers. Volumes of interest of 35 voxel were cropped inside the vertebrae. Displacement and strains were evaluated using DVC (DaVis-DC, LaVision, Germany), with different spatial resolution. The displacement maps were used to predict the force uncertainties via FE (Ansys Mechanical v.14, Ansys Inc, Canonsburg, PA). Each element was assigned a linear elastic isotropic constitutive law (Young modulus: 8 GPa, Poisson's ratio: 0.3, as in (Follet, Peyrin et al. 2007)). Overall, the precision error of strain measurement was evaluated as the average of the standard deviation of the absolute value of the different component of strain (Liu and Morgan 2007).

The force uncertainties obtained with the FE analysis produced magnitudes ranging from 231 to 2376 N. No clear trend on the force was observed in relation to the spatial resolution. Precision errors were smaller than 1000 microstrain in all cases, with the lowest ranging from 83 microstrain for the largest spatial resolution. Full-field strain on the bone tissue did not seem to highlight a preferential distribution of error in the volume.

The precision errors showed that the pedCAT-CT based DVC can be sufficient to investigate the bone tissue failure (7000–10000 microstrain) or, physiological deformation if well-optimized. FE analysis produced important force uncertainties up to 2376 N. However, this is a preliminary investigation. Further investigation will give a clearer indication on DVC based PedCAT-CT, as well as force uncertainties predicted. So far, the DVC showed its ability to measure displacement and strain with reasonable reliability with clinical-CT as well.

Knee arthroplasty with a rotating hinge knee (RHK) prosthesis has become an important clinical treatment option for knee revisions and primary patients with severe varus or valgus deformities and instable ligaments. The rotational axle constraints the anterior-posterior shear and varus-valgus moments, but currently used polyethylene bushings may fail in the mid-term due to insufficient creep and wear resistance of the material. Due to that carbon-fibre-reinforced (CFR) PEEK as an alternativ bushing material with enhanced creep, wear and fatigue behaviour has been introduced in a RHK design [Grupp 2011, Giurea 2014]. The objective of our study was to compare results from the pre-clinical biotribological characterisation to ex vivo findings on a series of retrieved implants.

In vitro wear simulation according to ISO 14243-1 was performed on rotating hinge knee devices (EnduRo® Aesculap, Germany) made out of cobalt-chromium and of a ZrN multi-layer ceramic coating for 5 million cycles. The mobile gliding surfaces were made out of polyethylene (GUR 1020, β-irradiated 30 ± 2 kGy). For the bushings of the rotational and flexion axles and the flanges a new bearing material based on CFR-PEEK with 30% PAN fiber content was used.

Analysis of 12 retrieved EnduRo^® RHK systems in cobalt-chromium and ZrN multi-layer in regard to

-

loosening torques in comparison with initial fastening torques

with a focus on the four CFR-PEEK components integrated in the EnduRo^® RHK design.

For the rotating hinge knee design with flexion bushing and flanges out of CFR-PEEK the volumetric wear rates were 2.3 ± 0.48 mm³/million cycles (cobalt-chromium) and 0.21 ± 0.02 mm³/million cycles (ZrN multi-layer), a 10.9-fold reduction (p = 0.0016). The UHMWPE and CFR-PEEK particles were comparable in size and morphology and predominantly in submicron size [5]. The biological response to representative sub-micron sized CFR-PEEK particles has been demonstrated in vivo based on the leucoyte-endothelian-cell interactions in the synovia of a murine intra-articular knee model by Utzschneider 2010. Schwiesau 2013 extracted the frequency of daily activities in hip and knee replacement patients from literature and estimated an average of 1.76 million gait cycles per year. Thus, the 5 million cycles of in vitro wear testing reflect a mean in vivo service life of 2.9 years, which fits to the time in vivo of 12–60 months of the retrieved RHK devices. The in vitro surface articulation pattern of the wear simulation tests are comparable to findings on retrieved CFR-PEEK components for both types of articulations – cobalt-chromium and ZrN multi-layer coating.

For the rotating hinge knee design the findings on retrieved implants demonstrate the high suitability of CFR-PEEK as a biomaterial for highly loaded bearings, such as RHK bushings and flanges in articulation to cobalt-chromium and to a ZrN multi-layer coating.

The human musculoskeletal system is a biological composite of hard and soft material phases organized into a complex 3D structure. The replication of mechanical properties in 3-dimensional space, so called ‘4D’ techniques, therefore promises next-generation of prosthetics and engineering structures for the musculoskeletal system. Approaches using in situ indentation of tissue correlated with micro computed tomography (μCT) are used here to provide a 4D data set that is representative of the native tissue at high fidelity. Multi-material 3D printing is exploited to realize the collected 4D data set by using materials with a wide range of mechanical properties and printing structures representative of native tissue. We demonstrate this correlative approach to reproduce bone structures and highlight a workflow approach of indentation, μCT and 3D printing to potentially mimic any structure found in the musculoskeletal system.

Structures in the human musculoskeletal system, such as bone [1] and tendon-bone connective tissue [2], can be considered as complex composites of hard and soft materials. Development of prosthetics capable of replacing body parts lost to trauma, disease or congenital conditions requires the accurate replication of the required body part. 3D printing promises considerable advantages over other manufacturing methods in mimicking native tissue, including the ability to produce complex structures [3]. However, accurate representation of whole body parts down to tissue microstructures requires correlative approaches where mechanical properties in 3-dimensional space are known. The objective of this study is to apply in situ indentation, correlate to 3D imaging of bone using μCT and finally 3D print mimicked structures.

Samples of bovine compact bone were imaged at high resolution using μCT (Xradia Versa 510, Zeiss, USA). A custom build in situ micro indentation setup within the μCT was used to map the mechanical properties of the bone at multiple positions. Correlation between sample x-ray attenuation and corresponding elastic modulus found from indentation was established. Data was converted to a 4D data set of elastic modulus values in 3D space, segmented and exported to the 3D printer. An inkjet 3D printer (Projet 5500X, 3D Systems, USA) was used to print materials with a range of mechanical properties that approach those found in the native bone material. Macroscopic testing on both bone samples and 3D printed samples were carried out using standard compression (Instron, UK).

Preliminary results indicated similarity between 3D printed structures and native bone tissue. Macroscopic testing of bone samples and 3D printed equivalents showed additional similarities in stress-strain behaviour.

Our preliminary work presented here indicates that the workflow of 3D imaging correlated to point mechanical measurements using indentation is suitable to give a 4D dataset that is representative of the native bone tissue. 3D printing is able to produce structures that start to mimick bone but are critically dependent on the data segmentation, particularly averaging imaging data to a resolution that is appropriate for the 3D printer.

3D printing can be used for the regeneration of complex tissues with intricate 3D microarchitecture. Trabecular bone is a complex and porous structure with a high degree of anisotropy. Changes in bone microarchitecture are associated with pathologies such as osteoporosis [1]. The objective of this study is to determine the viability of using 3D printing to replicate trabecular bone structures with a good control over the microarchitecture and mechanical properties.

Cylindrical samples of bovine trabecular bone were used in this study. Micro-computed tomography (microCT) was carried out and an isotropic voxel size of 22 µm was obtained (Xradia Versa 520, Zeiss, USA). After 3D reconstruction the main microstructure characteristics were analysed using ImageJ (NIH, US).

The 3D printed bone replicas were created by segmenting the microCT imaged bone tissue and then converted into a STL file using Avizo (FEI, US). The 3D printer used for this study was the ProJet 5500X (3D Systems, US), which allows a number of different materials to be printed in the same built with a resolution of 25 µm. Preliminary results were obtained using one single material (VisiJet CR-WT, Tensile Modulus: 1–1.6 GPa, Tensile Strength: 37–47 MPa). The 3D printed bone replicas followed a critical cleaning step to remove any remaining support material in the pores. MicroCT was then carried out for the bone replicas obtaining the same isotropic voxel size as for their biological counterparts. ImageJ was used to obtain the main microstructure characteristics.

The values of bone volume fraction (BV/TV), mean trabecular thickness (Tb.Th), mean trabecular spacing (Tb.Sp), and degree of anisotropy (DA) were measured for bone samples and their 3D printed replicas [2].

Preliminary results on the first bone sample with its 3D printed replica showed similar apparent trabecular structures. Their respective BV/TV was found to be 0.24 (bone) and 0.43 (replica). The Tb.Th and Tb.Sp were 0.222 mm and 0.750 mm respectively for the bone and 0.376 mm and 0.575 mm for the replica. Finally, their respective DA was found to be 0.68 (bone) and 0.66 (replica).

The main microstructure characteristics analyzed showed some differences between the bone sample and the 3D printed replica. In particular, the 3D microstructures resulted over-dimensioned mainly due to factors such as microCT voxel size, resolution of the 3D printer and supporting material removal. However this is a preliminary investigation. Further analysis will focus on optimizing the microCT imaging as well as the 3D printing process to achieve more accurate bone replicas. In addition, multi-material printing will be employed to optimize some of the mechanical properties obtained through in situ microCT testing and FE subject-specific modelling.

Prophylactic augmentation is meant to reinforce the vertebral body (VB), but in some cases it is suspected to actually weaken it. To elucidate the biomechanical efficacy of prophylactic augmentation, the full-field three-dimensional strain distributions were measured for the first time inside prophylactic-augmented vertebrae.

Twelve thoracic porcine vertebrae were assigned to three groups: 4 were augmented with bone cement for vertebroplasty (Mendec-Spine, Tecres), 4 were treated with another bone cement for vertebroplasty (Calcemex-Spine, Tecres) while the other 4 were tested untreated as a control. Destructive tests were carried out under axial compression, in a step-wise fashion (unloaded, 5%, 10% and 15% compression). At each loading step, μCT-images were acquired. The internal strain distribution was investigated by means of DVC analysis.

Some augmented specimens were stronger than the respective control, while others were weaker. In most of the specimens, the strain distribution in the elastic regime (5% compression) seemed to predict the location of the micro-damage initiation before it actually became identifiable (at 10% and 15% compression). The measured strain had the same order of magnitude for all groups. However, in the control vertebrae, the highest strain would unpredictably appear at any location inside the VB. Conversely, for both augmentation groups, the highest strains were measured in the regions adjacent to the injected cement mass, whereas the cement-interdigitated-bone was less strained. Localization of high strains and failure was consistent between specimens, but different between the two cement types: with Mendec-Spine failure the highest strains were mainly localized at mid-height and at the same level where the cement mass was localized; with Calcemex-Spine failure the highest strains were mainly cranial and caudal to the cement mass.

Both the micro-CT images, and the DVC strain analysis highlighted that:

The cement mass was less strained than any other regions in the vertebra. Failure never started inside the cement mass. This can be explained with the additional stiffening and reinforcement associated with the infiltration of the cement inside the trabecular bone.

This study has demonstrated the potential of DVC in measuring the internal strain and failure in prophylactic-augmented vertebrae. It has been shown that failure starts inside the augmented VB, next to the injected cement mass. This can help establishing better criteria (in terms of localization of the cement mass) in order to improve clinical protocols for vertebroplasty surgical procedures.