2.3. Micro-CT Imaging and In Situ Mechanics

The explanted sample was micro-CT scanned in a plastic container filled with PBS on the day of retrieval using a Versa 520 micro-CT scanner (Zeiss, Oberkochen, Germany) with coarse settings to provide a rapid evaluation of the defect site (Table 1). Each projection was of 2048 × 2048 resolution with no binning or filtering applied. Prior to in situ testing sample diameter was further reduced to approximately 10 mm using a precision sectioning saw (IsoMet, Buehler, Lake Bluff, IL, USA) to improve scan quality. The sample was stained in 1% w/v phosphotungstic acid (w/v, H₃PW₁₂O₄₀, PTA, #79690 Sigma-Aldrich, St. Louis, MO, USA) solution in 70% ethanol for 21 h following established protocol [26]. The sample was mounted to the bottom polyoxymethylene platen with polymethyl methacrylate bone cement (Simplex, Kemdent, Swindon, UK). The sample and platen were loaded into a custom-built in situ mechanical testing rig equipped with a load cell (LBS100, Interface Force Measurements Ltd, Crowthorne, UK). The plastic sample chamber was filled with 70% ethanol, the top compression platen was seated and silicone grease (404-124, RS Components, Corby, UK) used around the circumference to limit evaporation of the ethanol during scanning. The bottom platen was static, and all uniaxial loading was applied to the sample by the top platen. The custom-built top platen had a recessed diameter to avoid the risk of any released or formed gas bubbles in the sample vicinity being captured during scanning (Figure 1d). A small preload (<5 N) was applied to ensure good contact with the platens and two repeat reference scans were taken consecutively and thereafter in situ mechanical testing was performed at two levels of applied compression. Prior to each scan 30 min of equilibrium time was provided with the sample within the micro-CT scanner to ensure thermal and mechanical stability. The sample was kept in liquid at all times, including during scanning (Table 1).

X-ray micro-computed tomography parameters for scanning the extracted tissue-biomaterial sample. A coarse scan was taken to provide initial evaluation of the interior interface. In situ testing was then carried out with scans taken under successive load to permit study of deformation mechanics using DVC. SOD = Source to Object Distance, ODD = Object to Detector Distance.

Reconstruction of the projection images to produce 3D volumetric data sets was performed using the Reconstructor Scout-and-Scan software (Zeiss, Oberkochen, Germany). The reconstructed CT volumes were visualized and analysed using (Fiji Is Just) ImageJ software [27] (version 1.52g, NIH, Bethesda, MD, USA). For all scans, the optional rigid body movement correction module in DaVis was utilised to provide volumetric registration. After DVC calculation, the “subtract vector at reference position” process was utilised to restore displacements to the original orientation.

To evaluate error, the repeat pair of unloaded images were considered as constant-strain conditions [28]. Within Fiji, two separate pairs of image stacks of dimensions 450 × 450 × 450 voxels were manually aligned and then extracted to include all constituent parts of the tissue-biomaterial system (Figure 2), including both the implant (Figure 2d) and local tissue (Figure 2a,e–g). Images were imported into DaVis version 8.4.0 (LaVision, Göttingen, Germany) and the constant strain study was conducted across a range of subvolume sizes from 16–112 voxels (~30–205 µm) in increments of 16 for all three DVC processing methods available in the software 1) Direct correlation (DC) 2) Fast-Fourier transform (FFT) 3) A combined FFT + DC approach. For FFT + DC, the FFT pre-shift was set at 12 pixels larger than the DC step and three calculation passes were utilised.

Micro-CT reconstructed volumes of the extracted sample following six weeks in vivo. The entire extracted sample (a) with two volumes of interest digitally extracted (indicated with dashed boxes) to encompass all system components: newly formed tissue (b) covered the scaffold implant (d), within which there were also regions of newly formed tissue (c). Surrounding the defect was articular cartilage (g), the osteochondral transition between the cartilage and bone (e) and the subchondral bone (f). These components were analysed with DVC software between pairs of repeated constant strain micro-CT scans. The constant strain study provided error analysis on all components of the tissue-biomaterial system. No image processing was been applied to the micro-CT slices.

Error was quantified by calculating the mean absolute error (MAER, Equation (1)), representing the systematic error or accuracy, and the standard deviation of error (SDER, Equation (2)), representing the random error or precision inherent within the system [28]. MAER and SDER of strain were calculated through the cartilage using a MATLAB script (R2019a, The MathWorks Inc, Natick, MA, USA). The random error was also calculated for the displacement components. The systematic error of displacement was not calculated because the real displacement is unknown.

where “ε” represents the strain; “c” represents each of the six independent strain components; “k” represents the measurement point at each subvolume and N is the number of subvolumes measured.

Following optimisation of the DVC calculation technique and subvolume size using the error quantification, the loading scenario image stacks (unloaded-load1-load2) were processed in accordance with the maximum acceptable error set as 10% of the nominal strain [15]. For this osteochondral system, the highest nominal strain applied to a component of interest was approximately 2.5%; therefore, the maximum acceptable error was set as 0.25% strain (2500 µS). Using this threshold, the loaded DVC study was carried out using the FFT + DC method and a subvolume size of 80 voxels, providing spatial vector distancing of 181 µm. Analysis under loading was carried out with a volume of interest of dimensions 821 × 821 × 401 voxels (approximately 3.7 × 3.7 × 1.8 mm) at the interface surrounding the defect.