Digital Image Processing Applied in the Deformation Analysis of Hip Prosthesis: Multivariate Regression Analysis

Total Hip Arthroplasty (THA) is one of the most performed surgical procedures to treat osteoarthritis of the hip joint or other skeletal disorders, such as Developmental Dysplasia of the Hip (DDH) or malformations. The purpose of the THA is to replace damaged bone/cartilage and recover the biomechanics of the joint in the best possible way. Placing hip prostheses is common practice in the treatment of THA, but the effect of this treatment must be evaluated before surgery. This work proposes a method to assess hip deformation with an implanted acetabular cup, using digital image processing and multivariate linear regression. Images were acquired when a hemipelvis was subjected to mechanical load ranging from 0 to 1800 Newtons. Eighteen geometric characteristics were analyzed, and only seven were significant in the multivariate linear regression model. To decide on the variables that are considered important, three criteria were used on the multivariate linear regression model established in the different characteristics. The first criterion considers the significance of the model coefficients; the second criterion considers a coefficient of determination greater than 75%, and the third criterion considers that the proposed model is significant. If all three criteria are met, the response variable is considered significant. This analysis gives specialists a tool to make decisions before performing THA treatment. Moreover, this methodology could be applied to assess the deformation of the hip with different pathologies such as DDH, malformations, and overweight conditions. Finally, a relationship is generated between the extracted geometric characteristics and the experimental results obtained in the laboratory.