Elsevier

Bone

Volume 44, Issue 4, April 2009, Pages 579-584
Bone

Trabecular bone strength predictions using finite element analysis of micro-scale images at limited spatial resolution

https://doi.org/10.1016/j.bone.2008.11.020Get rights and content

Abstract

Advances in micro-scanning technology have led to renewed clinical interest in the ability to predict bone strength using finite element (FE) analysis based on images with resolutions in the range of 80 µm. Using such images, we sought to determine whether predictions of yield stress provided by nonlinear FE models could improve correlations with bone strength as compared to the use of predictions of elastic modulus from linear FE models, and if this effect depended on voxel size or bone volume fraction. Linear and nonlinear FE analyses were conducted for 46 trabecular cores from three human anatomic sites using element sizes ranging from 20 to 120 µm to obtain measures of apparent yield stress and elastic modulus, and these measures were correlated against the predicted yield stress from the 20 µm models (assumed to be the gold standard strength for this study). Results indicated that when considering all samples and any resolution, yield stress and elastic modulus were both excellent predictors of strength (R2 > 0.99). When only low-density samples (BV/TV < 0.15) were considered, yield stress was better correlated with 20 µm-strength than was elastic modulus (R2 increased from 0.93 to 0.99 at 40 µm and from 0.90 to 0.95 at 80 µm). However, at a voxel size of 120 µm, the predictive ability of yield stress was slightly less than that of stiffness, likely due to the large convergence-related errors that could develop with larger element sizes. As expected, a limit was observed in the ability of elastic modulus to predict strength—the predictive ability of elastic modulus measured at 20 µm was comparable to that of yield strength at 80 µm. We also found that strength predictions from FE models at clinical-type resolutions had nearly the same power to detect bone quality effects via variations in strength–density relationships as did high-resolution models. We conclude that nonlinear FE models can account for additional variations in strength relative to linear models when using images at resolutions of ∼ 80 µm and less, and such models offer a promising method for examining microarchitecture-related bone quality effects associated with aging, disease, and treatment.

Introduction

Finite element models have the potential to provide non-invasive in vivo estimates of bone biomechanical properties [1], [2], [3], which could aid in osteoporotic fracture risk prediction and assessment of therapeutic efficacy. Recent advances in micro-CT and micro-MRI scanning technology have made in vivo clinical imaging possible at voxel sizes of about 80 µm at peripheral sites [1], [4] and 150–300 µm at the vertebral body [5]. While such images may not be sufficient to provide accurate measures of apparent mechanical properties in the context of high-resolution micro-CT-based finite element analysis [6], [7], [8], they may still be able to provide information on bone strength that would be relevant in a clinical-type setting. Furthermore, variations in the strength–density relationships as predicted by these models may enable researchers to detect microarchitecture-related bone quality effects [9] associated with aging, disease, or treatment.

A promising avenue with finite element predictions of bone strength using images at limited spatial resolution is the added benefit of performing nonlinear analyses to obtain measures of failure stress over linear analyses that only supply stiffness or elastic modulus. While elastic modulus is typically well correlated with strength across a range of bone densities [10], [11], [12], estimates of yield stress (that can only be directly obtained from nonlinear finite element modeling) may improve predictions of strength since failure behavior by definition involves nonlinear phenomena such as material yielding and geometrically large deformations. To the best of our knowledge, only one study to date has implemented nonlinear modeling techniques in combination with HR-pQCT-based images and compared strength predictions from linear versus nonlinear finite element models [3], although nonlinear analysis was only performed for a small cohort of specimens. Results indicated that linear models correlated well with strength over a large data range, but that errors in absolute strength prediction were reduced appreciably when nonlinear models were used [3]. However, as imaging technologies improve and are applied to low-density or osteoporotic bone it is unclear whether similar trends will hold true. For example, the relative benefits of performing nonlinear versus linear finite element analysis for predicting bone strength may be influenced by factors such as image resolution (influencing both geometric discretization and finite element model convergence), volume fraction, and anatomic site, since nonlinear phenomena are present to differing degrees in low-density sites and low volume fraction bone [13], [14].

In this context, our overall goal was to assess the ability of both linear and nonlinear high-resolution finite element models to predict bone strength for human trabecular bone from multiple anatomic sites that display a wide range of volume fractions. Our specific objectives were to: 1) determine the predictive capability of coarsened finite element models relative to 20 µm models for apparent elastic and yield properties; 2) assess the added benefits of performing nonlinear versus linear finite element analysis for predicting trabecular bone strength; 3) determine the dependence of convergence behavior of linear and nonlinear models on bone volume fraction; and 4) quantify the differences in strength–density relationships as predicted by finite element models at varying resolutions. This study is, to our knowledge, the first to examine the convergence behavior of nonlinear micro-finite element models of trabecular bone, and to examine the differences in predictive abilities for linear versus nonlinear finite element models across a range of image resolutions, bone volume fractions, and anatomic sites.

Section snippets

Methods

Forty-six cylindrical specimens (∼ 8.1 mm diameter and 20 mm long) of trabecular bone were selected from the human femoral neck (n = 14), greater trochanter (n = 14), and vertebral body (n = 18). All specimens were machined such that the main trabecular orientation was aligned with the axis of the core [11]. None of the donors had a history of metabolic bone disease or cancer and all specimens showed no radiographic evidence of damage or bone pathologies.

Specimens were scanned using micro-computed

Results

The convergence characteristics for yield stress and elastic modulus were qualitatively similar, and showed comparable trends with respect to both bone volume fraction and anatomic site. Errors in predicted apparent yield stress for the 80 and 120 µm models were relatively small (< 15%) for the femoral neck, but could be appreciable (20–70%) for the more porous greater trochanter and vertebral trabecular bone (Fig. 2). Analysis of covariance (including BV/TV and anatomic site as independent

Discussion

The objectives of this study were to determine the convergence behavior of linear versus nonlinear high-resolution finite element models of trabecular bone, and to compare the ability of such coarsened models to predict trabecular bone strength as estimated by very high-resolution models. We found that the convergence behavior for predictions of yield stress and elastic modulus were qualitatively similar, and showed comparable trends with respect to both bone volume fraction and anatomic site.

Acknowledgments

Funding for this work was provided by Merck & Co., Inc., and the National Institutes of Health (AR43784). Cadaveric material was obtained from NDRI and UCSF Willed Body Program. Super-computing resources were obtained from the National Partnership for Advanced Computational Infrastructure (NPACI UCB266). Dr. Keaveny has a financial interest in O.N. Diagnostics and both he and the company may benefit from the results of this research.

References (31)

Cited by (117)

  • MRI-based mechanical competence assessment of bone using micro finite element analysis (micro-FEA): Review

    2022, Magnetic Resonance Imaging
    Citation Excerpt :

    Microcomputed tomography (μCT) has been the primary preferred technique for performing ex vivo micro-FEA in order to assess the mechanical competence of trabecular bone [22,32,33]. Other high-resolution x-ray-based imaging techniques, such as high-resolution peripheral quantitative computed tomography (HR-pQCT), have also been used in FEA studies to predict the mechanical competence of bone structure [34–36]. For example Orwoll et al. [37] studied the correlations between FEA results and the risk of hip fracture in older men and reported that the femoral strength and the load-to-strength ratio obtained from QCT-based FEA were strongly associated with the hip fracture risk.

View all citing articles on Scopus
View full text