General Information
This study selected a 26-year-old male patient exhibiting a body weight of 70kg and a height of 172cm. In addition, CT spiral scan data of the hip of the patient were acquired. The thickness and spacing of the layer reached 0.5mm and 5mm, respectively. Furthermore, the resolution of each layer was 1024×1024 pixels, and the images generated were stored as DICOM format.
Analysis software
On the whole, Mimics software (version 16.0, Materialise NV, Leuven, Belgium), Geomagic-Studio software (version 11.0, Geomagic Corporation, USA), Solidworks software (version 2014, Dassault Systems S.A, USA) and Abaqus 6.10.1 (Simulia, Rhode Island, USA) were applied in this study as the analysis software.
Establish the model of ONFH
The DICOM CT data were imported into the Mimiucs software. The appropriate threshold segmentation was selected. The redundant connections between different bone cortex were deleted, the femoral cortical bone outline was obtained. Subsequently, regional growth, skin editor, cavity filling of cortical bone repair were conducted, and then a 3D model of cortical bone was built. After the mild smooth surface treatment, a STL format file was exported (Fig. 1). The STL file was imported into the Geomagic-Studio 11 software to conduct further surface smoothing, model peeling removal, mesh division, contour detection, curved surface construction and fitting, which was saved as STP format. Next, the STP format file was imported into the SoildWorks software for the assembly, and then the normal femoral head model was built. Afterwards, in accordance with the JIC classification, the 3D model of Osteonecrosis of the femoral head (JIC C1 and C2) were built.
Establish the model of allograft fibula and cannulated screw
The finite element model of allograft fibula was built in the SoildWorks software. Cortical bone, which was 100mm in length, 10mm in outer diameter and 5mm in diameter, of the hollow part was set up to implant allogeneic fibula column, and the model was saved as a PRT file. The finite element model of the cannulated screw was built in the SoildWorks software, and various specifications of the cannulated screw were set (e.g., length of 80mm, 90mm and 100mm). The hollow part had the diameter of overall 6.5mm, the diameter of the hollow part was all 5mm, and the thread length was 30mm, which was located at the distal end of the screw. Furthermore, the model was saved as PRT format (Fig. 2).
Establish the model of allograft fibula combined with cannulated screw for ONFH
The JIC type C1 and C2 ONFH model, allograft fibula and cannulated screw were assembled with the SoildWorks 2015 software. The allograft fibula was implanted at 15mm below the crest of the greater trochanter at the necrotic area of the anterolateral column of the femoral head. Besides, the depth was lower than 5mm from the subchondral bone of the femoral head, as well as nearly 5mm from the lateral femoral neck cortex. By using the allograft fibula implantation model, the cannulated screws were placed directly below, posteriorly below, anteriorly below and anteriorly above the allograft fibula, respectively. Moreover, the insertion depth of the cannulated screws, no less than 5mm, did not exceed the depth of the allograft fibula. 8 allograft fibulae combined with cannulated screw models were respectively generated (JIC C1 below model, JIC C1 posteriorly below model, JIC C1 anteriorly below model, JIC C1 anteriorly above model, JIC C2 below model, JIC C2 posteriorly below model, JIC C2 anteriorly below model, as well as JIC C2 anteriorly above model), which were saved as an X_T file (Fig. 3).
Material properties and meshing
The models were respectively imported into Abaqus software, and the proximal femur material acted as the isotropic elastic material. Table 1 lists the elastic modulus and Poisson's ratio of the respective material [17-20]. The femur, the cannulated screw and the fibula were all administrated with 10-node 4-hedral elements (C3D10), and Table 2 lists the nodes and elements of the respective model.
Table 1
Elastic modulus and Poisson's ratio of different materials
Material
|
Elastic modulus(MPa)
|
Poisson's ratio
|
Femoral cortical bone
|
15100
|
0.3
|
Femoral cancellous bone
|
445
|
0.22
|
Cannulated screw
|
110000
|
0.3
|
Allograft fibula
|
15100
|
0.3
|
The necrotic bone
|
125
|
0.15
|
Cartilage
|
10.5
|
0.45
|
Table 2
Nodes and units of each model (number)
Groups
|
Femoral cortical bone
|
Femoral cancellous bone
|
Allograft fibula
|
Cannulated screw
|
The necrotic bone of JIC C1
|
The necrotic bone of JIC C2
|
Cartilage
|
Number of nodes
|
145124
|
171453
|
10578
|
6957
|
35775
|
38784
|
22089
|
Unit number
|
86359
|
118008
|
6109
|
3495
|
24409
|
26480
|
11064
|
Boundary conditions and loads
All nodes were constrained at the distal end of the femur of 6 degrees freedom to zero. The simplified model of standing on one foot was employed, and the axial stress of 700N, which was axial downward, was applied in the bearing area of the femoral head (Fig. 4). The relationship between the allograft fibula and the femoral bone was set as the friction, and the coefficient of friction reached 0.2. Furthermore, the relationship between the cannulated screw and the femoral bone was set as the binding relation.
Evaluation criteria
First, the von Mises stress distribution of the subchondral cortical bone of femoral head was observed. Second, a comparison was drawn on the maximum stress value and mean stress value of the subchondral cortical bone of femoral head weight-bearing area.
Statistical analysis
Stata 13.0 statistical software was employed for the statistical analysis, and χ2 test or Fisher's exact probability method was adopted to count the data. The measurement data were expressed as mean ± standard deviation (x±s), and the normal distribution (Shapiro-Wilk) and the homogeneity of variance test (Bartlett method) were performed simultaneously. Under the normal distribution and the homogeneous variance, Bonferroni method will be applied to draw the pairwise comparison. Furthermore, under the unnormal distribution, or the uneven variance, non-parametric test would be performed, and P<0.05 was statistically significant.