Experimental bone research often generates large amounts of histology and histomorphometry data, and the analysis of these data can be time-consuming and trivial. Machine learning offers a viable alternative to manual analysis for measuring e.g. bone volume versus total volume. The objective was to develop a neural network for image segmentation, and to assess the accuracy of this network when applied to ectopic bone formation samples compared to a ground truth. Thirteen tissue slides totaling 114 megapixels of ectopic bone formation were selected for model building. Slides were split into training, validation, and test data, with the test data reserved and only used for the final model assessment. We developed a neural network resembling U-Net that takes 512×512 pixel tiles. To improve model robustness, images were augmented online during training. The network was trained for 3 days on a NVidia Tesla K80 provided by a free online learning platform against ground truth masks annotated by an experienced researcher.Introduction
Method
Healing after bone fracture is assessed by frequent radiographs, which expose patients to radiation and lacks behind biological healing. This study aimed to investigate whether the electrical impedance using electrical impedance spectroscopy correlated to quantitative scores of bone healing obtained from micro-CT and mechanical bending test. Eighteen rabbits were subjected to tibial fracture that was stabilized with external fixator. Two electrodes were positioned, one electrode placed within the medullary cavity and the other on the lateral cortex, both three millimeters from the fracture site. Impedance was measured daily across the fracture site at a frequency range of 5 Hz to 1 MHz. The animals were divided into three groups with different follow-up time: 1, 3 and 6 weeks for micro-CT (Bone volume/tissue volume (BV/TV, %)) and mechanical testing (maximum stress (MPa), failure energy (kJ/cm3), young modulus (Mpa)). There was a statistically significant correlation between last measured impedance at 5 Hz frequency immediately prior to euthanasia and BV/TV of callus (−0.68, 95%CI: (−0.87; −0.31)). Considering the mechanical testing with three-point bending, no significant correlation was found between last measured impedance at 5 Hz frequency immediately prior to euthanasia and maximum stress (−0.35, 95%CI: (−0.70; 0.14)), failure energy (−0.23, 95%CI: (−0.63; 0.26)), or young modulus (−0.28, 95%CI: (−0.66; 0.22)). The significant negative correlation between impedance and BV/TV might indicate that impedances correlate with the relative bone volume in the callus site. The lack of correlation between impedance and mechanical parameters when at the same time observing a correlation between impedance and days since operation (0-42 days), might indicate that the impedance can measure biological changes at an earlier time point than rough mechanical testing.
The primary outcome was investigating differences in wear, as measured by femoral head penetration, between cross-linked vitamin E-diffused polyethylene (vE-PE) and cross-linked polyethylene (XLPE) acetabular component liners and between 32 and 36 mm head sizes at the ten-year follow-up. Secondary outcomes included acetabular component migration and patient-reported outcome measures (PROMs) such as the EuroQol five-dimension questionnaire, 36-Item Short-Form Health Survey, Harris Hip Score, and University of California, Los Angeles Activity Scale (UCLA). A single-blinded, multi-arm, 2 × 2 factorial randomized controlled trial was undertaken. Patients were recruited between May 2009 and April 2011. Radiostereometric analyses (RSAs) were performed from baseline to ten years. Of the 220 eligible patients, 116 underwent randomization, and 82 remained at the ten-year follow-up. Eligible patients were randomized into one of four interventions: vE-PE acetabular liner with either 32 or 36 mm femoral head, and XLPE acetabular liner with either 32 or 36 mm femoral head. Parameters were otherwise identical except for acetabular liner material and femoral head size.Aims
Methods
The most frequent indication for revision surgery in total hip arthroplasty (THA) is aseptic loosening. Aseptic loosening is associated with polyethylene liner wear, and wear may be reduced by using vitamin E-doped liners. The primary objective of this study was to compare proximal femoral head penetration into the liner between a) two cross-linked polyethylene (XLPE) liners (vitamin E-doped (vE-PE)) versus standard XLPE liners, and b) two modular femoral head diameters (32 mm and 36 mm). Patients scheduled for a THA were randomized to receive a vE-PE or XLPE liner with a 32 mm or 36 mm metal head (four intervention groups in a 2 × 2 factorial design). Head penetration and acetabular component migration were measured using radiostereometric analysis at baseline, three, 12, 24, and 60 months postoperatively. The Harris Hip Score, University of California, Los Angeles (UCLA) Activity Score, EuroQol five-dimension questionnaire (EQ-5D), and 36-Item Short-Form Health Survey questionnaire (SF-36) were assessed at baseline, three, 12, 36, and 60 months.Aims
Methods
The metal on metal implants was introduced without the proper stepwise introduction. The ASR resurfacing hip arthroplasty (RHA) withdrawn due to high clinical failure rates and the large diameter head THA (LDH-THA) are also widely abandoned. Early (2 year) radiostereometry studies does not support early instability as cause of failure but more likely metal wear products. A possible advantage may be maintenance of bone mineral density (BMD). We present 5 year prospective follow up from a randomized series, aiming to report changes from baseline and to investigate links between implant micromotion, Cr & Co ions and BMD. Patients eligible for an artificial hip were randomized to RHA, Biomet LDH-THA or standard Biometric THA. 19, 17 and 15 patients completed 5 year follow-up. All followed with BMD of the femur, acetabulum and for RHA the collum. RHA and THA with whole blood Co and Co. LDH-THA only at 5 year. RHA had marker based RSA of both components, cup only for LDH-THA. Translations were compiled to total translation (TT= √(x2+y2+z2)). Data were collected at baseline, 8 weeks, 6 months, 1, 2 and 5 years. Statistical tests: ANCOVA for TT movement, Spearman's correlation for BMD, Cr, Co and BMI to TT at 5 years RSA: The 5 year median (25%to75%) RHA cup translations were X=-0.00(−0.49 to 0.19) Y=0.15(−0.03 to 0.20), z=0.24(−0.42 to 0.37) and TT 0.58 (0.16 to 1.82) mm. For the LDH-THA X=−0.33(−0.90 to 0.20) Y=0.28(0.02 to 0.54), z=0.43(−1.12 to −0.19) and TT 1.06 (0.97 to 1.72) mm. The TT was statistically different (p<0.05) for the two cups. The RHA femoral component moved X=0.37(0.21 to 0.56) Y=0.02(−0.07 to 0.11), z=-0.01(−0.07 to 0.26) and TT 0.48 (0.29 to 0.60) mm at 5 years. There was no TT movement from year 2. The mean (SD) acetabular BMD was diminished to 93(90–97)% for RHA and 97(93–99.9)% for THA, but LDH-THA maintained 99(95–103)%. Overall femoral BMD was unchanged at 5 years for all interventions, but both stemmed implants lost 17% at the calcar. Median (25%to75%) whole-blood Cr peaked in the LDH-THA group with 1.7 (0.9 to 3.1) followed by RHA 1.2 (0.8 to 5.0) and THA with 0.5 (0.4 to 0.7)ppb. For Co the highest levels were found in RHA with 1.6(0.8 to 4.7) followed by LDH-THA 1.2 (0.7–1.7) and THA 0.2 (0.2 to 0.6) ppb. The only correlations above +/−0.3 for TT were the RHA femoral component with a correlation of 0.47 to BMI, 0.30 to Co and Cr. The ASR cup conversely had a negative correlation of −0.60 to BMI and again, the LDH-THA cup had a negative correlation of −0.37 to Cr. In contrast to registered revision rates, we found significantly larger movement for the Biomet cup than the ASR cup. The metal ion levels were similar. The LDH-THA cup maintained the acetabular BMD best at 5 years, but the difference was small, we are limited by small numbers and the correlations between TT and the covariates showed no clear pattern.
Despite high revision rates, the mean two year migration of the ASRTM cup is within an acceptable threshold. Slightly higher migration rates found for the M2a- Magnum™ Porous Coated Acetabular Component but longer follow up is needed to establish if this implant is at risk. RSA can detect the migration of an implant, and continuous migration is a predictor for failure (1). The ASRTM resurfacing implant was withdrawn from the marked due to excessive failure rate but showed initial femoral component stability. The aim of this study was to investigate the initial implant stability for the ASR cup as a possible explanation for the high revision rate, and to compare it to another metal on metal (MoM) cup.Summary
Introduction
The combined incubation of a composite scaffold with bone marrow stromal cells in a perfusion bioreactor could make up a novel hybrid graft material with optimal properties for early fixation of implant to bone. The aim of this study was to create a bioreactor activated graft (BAG) material, which could induce early implant fixation similar to that of allograft. Two porous scaffold materials incubated with cells in a perfusion bioreactor were tested in this study. Two groups of 8 skeletally mature female sheep were anaesthetized before aspiration of bone marrow from the iliac crest. For both groups, mononuclear cells were isolated, and injected into a perfusion bioreactor (Millenium Biologix AG, Switzerland). Scaffold granules Ø∼900–1500 μm, ∼88% porosity) in group 1, consisted of hydroxyapatite (HA, 70%) with -tricalcium-phosphate (−TCP, 30%) (Danish Technological Institute, Denmark). The granules were coated with poly-lactic acid (PLA) 12%, in order to increase the mechanical strength of the material (Phusis, France). Scaffold granules Ø∼900–1400 μm, 80% porosity) in group 2 consisted of pure HA/-TCP (Fin Ceramica, Italy). For both groups, cells were incubated in the bioreactor for 2 weeks. Fresh culture medium supplemented with dexamethasone and ascorbic-acid was added every third or fourth day. Porous titanium alloy implants with diameter=length=10mm (Biomet, USA) were inserted bilaterally in each of the distal femurs of the sheep; thus 4 implants in each sheep. The concentric gap (2 mm) surrounding the implant was filled with 1) BAG (autogenous), 2) granules, 3) granules+bone marrow aspirate (BMA, autologous) or 4) allograft. The sheep were euthanized after 6 weeks. Distal femurs were removed and implant-bone samples were divided in two parts. The superficial part was used for mechanical testing and micro-CT scanning, and the profound part for histomorphometry. Push-out tests were performed on an 858 Bionix MTS hydraulic materials testing machine. Shear mechanical properties between implant and newly generated bone were calculated to assess implant fixation. Results were assessed by One-way ANOVA. P-values less than 0.05 were considered significant.Introduction
Methods and Materials
Osteoporosis (OP), osteoarthrosis (OA), and rheumatoid arthritis (RA) are the most common age-related degenerative bone diseases, and major public health problems in terms of enormous amount of economic cost. RA is considered as a major cause of secondary osteoporosis. At late stage, OP often leads to skeletal fractures, and OA and RA result in severe joint disability. Over the last a few decades, much significant research on the properties has been carried out on these diseases, however, a detailed comparison of the microarchitecture of cancellous bones of these diseases is not available. In this study, we investigated three-dimensional (3-D) microarchitectural properties of OP, OA and RA cancellous bone. We hypothesized that there were significant differences in microarchitecture among OP, OA and RA bone tissues that might lead to different bone quality. Twenty OP, fifty OA, and twelve RA femur heads were harvested from patients undergone total hip replacement surgery. Cubic cancellous bone samples (8∗8∗8 mm3) were prepared and scanned with a high resolution microtomographic system (vivaCT 40, Scanco Medical AG., Brüttisellen, Switzerland). Then micro-CT images were segmented using individual thresholds to obtain accurate 3-D data sets. Detailed microarchitectural properties were evaluated based on novel unbiased, model-free 3-D methods. For statistical analysis, one-way ANOVA was used, and a p<0.05 was considered significant.Introduction
Materials and Method
For bone grafting procedures, the use of autologous bone is considered the gold standard, as it is has a better healing capacity compared to other alternatives as allograft and synthetic bone substitutes. However, as there are several drawbacks related to autografting (infection, nerve- or vascular damage, chronic pain problems, abdominal herniation), there has been a targeted effort to improve the healing capacities of synthetic bone substitutes. To evaluate the performance of a carbonated osteoionductive hydroxyapatite (CHA) scaffold of clinical relevant size (Ø=15mm, H=50mm) in a sheep model of multi level posterolateral intertransverse lumbar spine fusion after activation with autologous bone marrow nuclear cells (BMNC) in a flow perfusion bioreactor.Background
Aim
There is a great need for suitable large animal models that closely resemble osteoporosis in humans, and that they have adequate bone size for bone prosthesis and biomaterial research. This study aimed to investigate effects of a 7 month glucocorticoid (GC) treatment alone without ovariectomy on the properties of sheep cancellous bone. Eighteen female sheep were randomly allocated into 3 groups: group 1 (GC-1) received GC (0.60mg/kg/day methylprednisolone) 5 days weekly for 7 months; group 2 (GC-2) received the same treatment regime for 7 months, and further observed for 3 months without GC; and group 3 served as the control group, and left untreated for 7 months. The sheep received restricted diet. After 7 months of GC treatment. Cancellous bone volume fraction of the 5th lumbar vertebra in the GC-1 group was reduced by −35%, trabecular thickness by −28%, and changed from typical plate structure to a combination of plate and rod structure with increased connectivity by 202%. Bone strength was reduced by 52%. Bone formation marker, serum osteocalcin of GC-1, was reduced by 71% at 7 months, but recovered with an increase of 45% at 10 month in the GC-2 group. Similar trends were also seen in the femur and tibia. At 10 months, the GC-2 group had microarchitectural and mechanical properties similar to the level of the control sheep. We have demonstrated in this study that 7 month high-dose GC on bone density and microarchitecture are comparable with those observed in human after long-term GC treatment. Moreover, we have shown that the bone quality with regard to strength and microarchitecture recovers after 3 months further observation without GC. This suggests that a prolonged administration of GC is needed for long-term observation to keep osteopenic bone. The model will be useful in pre-clinical studies.
Structure model type and trabecular thickness are important characteristics in describing cancellous bone architecture. It has been qualitatively observed that a radical change of trabeculae from plate-like to rod-like occurs in aging, bone remodeling, and osteoporosis. Thickness of trabeculae has traditionally been measured using model-based histomorphometric methods on two-dimensional (2-D) sections. However, no quantitative study has been published based on three-dimensional (3-D) methods on the age-related changes in structure model type and trabecular thickness for human peripheral (tibial) cancellous bone. In this study, 160 human proximal tibial cancellous bone specimens from 40 normal donors, aged 16 to 85 years, were collected. These specimens were micro-CT scanned, then the micro-CT images were segmented using optimal thresholds. From accurate 3-D data sets, structure model type and trabecular thickness were quantified by means of novel 3-D methods. Structure model type was assessed by calculating the structure model index (SMI). The SMI was quantified based on a differential analysis of the triangulated bone surface of a structure. This technique allowed quantification of structure model type, such as plate, rod objects or mixture of plates or rods. Trabecular thickness was calculated directly from 3-D images, which is especially important for an a priori unknown or changing structure. Furthermore, 2-D trabecular thickness was also calculated based on the plate model. Our results showed that structure model type changed towards more rod-like in the elderly, and that trabecular thickness declined significantly with age. These changes become significant after 80 years of age for human tibial cancellous bone, whereas both properties seem to remain relatively unchanged between 20 and 80 years. Although a fairly close relationship was seen between 3-D trabecular thickness and 2-D trabecular thickness, real 3-D trabecular thickness was significantly underestimated using 2-D method.
Osteoarthrosis (OA) stands alongside cancer and heart disease as one of the major causes of suffering and disability amongst the elderly. Changes related to OA occur in all elements of the joint, and there are indications that sub-chondral cancellous bone plays a primary role in the cartilage degeneration in OA. Most previous investigations have been focused on moderate and late OA, whereas little is known about the changes in cancellous bone microstructure in human early OA. This study quantified cancellous bone microstructure in early-stage OA using three-dimensional (3D) methods. Subchondral cancellous bone specimens, produced from 10 human post-mortem early-stage osteoarthrotic (OA) proximal tibiae and 10 normal age- and gender-matched proximal tibiae, were allocated to 4 groups: medial OA, lateral control, normal medial control, and normal lateral control. OA initiates mostly at the medial condyle, and histological analysis was done to confirm this change. The cylindrical specimens were micro-computed tomography (micro-CT) scanned. From accurate 3D data sets, structural parameters were determined by means of true, unbiased and assumption-free 3D methods. The data were assessed statistically, and a p<
0.05 was considered significant. Our data supported the hypothesis that significant microstructural changes – other than density changes – occur in early-stage OA cancellous bone. OA cancellous bone is markedly plate-like, less anisotropic, less interconnected, but lower in mechanical properties which suggests a disorganisation in the microstructure as OA initiation. Structure model type best explains the mechanical properties for the OA and the normal controls. However, the determination coefficients (R2) for the OA group are largely reduced. These results indicate significant property and quality deterioration in early-stage OA subchondral cancellous bone.
