Prediction of Biomechanical Parameters of the Proximal Femur Using Statistical Appearance Models and Support Vector Regression

Abstract

Abstract.Fractures of the proximal femur are one of the principal causes of mortality among elderly persons. Traditional methods for the determination of femoral fracture risk use methods for measuring bone mineral density. However, BMD alone is not sufficient to predict bone failure load for an individual patient and additional parameters have to be determined for this purpose. In this work an approach that uses statistical models of appearance to identify relevant regions and parameters for the prediction of biomechanical properties of the proximal femur will be presented. By using Support Vector Regression the proposed model based approach is capable of predicting two different biomechanical parameters accurately and fully automatically in two different testing scenarios.

Assessment of the individual fracture risk of the proximal femur by using statistical appearance models

“…In addition, the prediction of FL could also be improved by using alternative attribute ranking algorithms and the usage of support vector regression instead of linear regression as presented in Fritscher et al 22,23 ACKNOWLEDGMENT This work has been supported by the AO Foundation ͑As-sociation of Osteosynthesis͒. a͒ Electronic mail: benedikt.schuler@umit.at 1 TABLE II.…”

Statistical model based analysis of bone mineral density of lumbar spine

“…Then the obtained correlation values were sorted according to their absolute value and the PCs were entered in the regression model in this order. For more detailed information on these methods presented above, please also refer to [1,3].…”

“…All calculations have been performed on these resampled datasets. Afterwards, the images were segmented automatically by the model-based level-set method described in [2,3], followed by manual corrections of the segmentation results. These segmented images formed the basis to apply the following methods:…”

“…These methods, especially the shape-intensity ("InShape")-model approach and its application, e.g. at the proximal femur has been described in detail previously [1,3]. First results of generating InShape models for the lumbar spine and their application to automated level set segmentation are presented in [2].…”

“…Based on our previous work on automated local bone quality assessment for the proximal femur [1,3], this paper will describe first attempts and results of adapting and extending this approach to the lumbar vertebral column. In general, the method consists of three successive steps:…”