Finite element modeling of proximal femur with quantifiable weight-bearing area in standing position

Background: The positional distribution and size of the weight-bearing area of the femoral head in the standing position as well as the direct active surface of joint force can directly affect the result of finite element (FE) stress analysis. However, the division of this area was vague, imprecise, and un-individualized in most studies related to separate FE models of the femur. The purpose of this study was to quantify the positional distribution and size of the weight-bearing area of the femoral head in standing position by a set of simple methods, to realize individualized reconstruction of the proximal femur FE model.

Methods: Five adult volunteers were recruited for an X-ray and CT examination in the same simulated bipedal standing position with a specialized patented device. We extracted these image data, calculated the 2D weight-bearing area on the X-ray image, reconstructed the 3D model of the proximal femur based on CT data, and registered them to realize the 2D weight-bearing area to 3D transformation as the quantified weight-bearing surface. One of the 3D models of the proximal femur was randomly selected for finite element analysis (FEA), and we defined three different loading surfaces and compared their FEA results.

Results: A total of 10 weight-bearing surfaces in 5 volunteers were constructed, and they were mainly distributed on the dome and anterolateral of the femoral head with a crescent shape, in the range of 1218.63-1,871.06 mm². The results of FEA showed that stress magnitude and distribution in proximal femur FE models among three different loading conditions had significant differences, and the loading case with the quantized weight-bearing area was more in accordance with the physical phenomenon of the hip.

Conclusion: This study confirmed an effective FE modeling method of the proximal femur, which can quantify the weight-bearing area to define a more reasonable load surface setting without increasing the actual modeling difficulty.