Extracting Quantitative Information From MRI Bound- and Pore-Water Maps of Cortical Bone
Mary Kate Manhard1, Sasidhar Uppuganti2, Mathilde C Granke2, Daniel F Gochberg3, Jeffry S Nyman2, and Mark D Does1

1Biomedical Engineering, Vanderbilt University, Nashville, TN, United States, 2Department of Orthopaedics & Rehabilitation, Vanderbilt University, Nashville, TN, United States, 3Vanderbilt University Institute of Imaging Science, Vanderbilt University, Nashville, TN, United States

Synopsis

Bound and pore water concentration measures of cortical bone found from MRI have been shown to correlate with material properties of bone, but the ideal way to analyze and draw information from 3D quantitative maps remains unclear. Material properties of cadaver radii found from a 3-point bend test were correlated with characteristics of the distribution of bound and pore water concentrations (e.g. mean, skewness) in ROIs found from different segmentations. Results highlighted the importance of segmentation method as well as quantitative measures drawn from the maps.

Introduction

Material properties of cortical bone such as strength and toughness correlate with bound and pore water concentration as measured from UTE MRI,1–3 and bound and pore water mapping methods have been implemented in vivo in clinically practical scans.4,5 However, the best way to analyze and draw information from these quantitative bound and pore water maps currently remains unclear. For example, small ROIs of bound and pore water maps may miss changes in porosity relevant for predicting fracture.6 In addition, the segmentation of the whole bone can greatly affect results, particularly the inclusion or exclusion of the transition zone between cortical and trabecular bone near the endosteal surface. This area of bone has been shown to be the first to deteriorate when bone loss occurs.7 The purpose of this work was to investigate 1) how segmenting the whole bone affects overall measurements of bound and pore water concentration and 2) what quantitative information from these ROIs are important for evaluating fracture risk.

Methods

The Adiabatic Inversion Recovery (AIR) and Double Adiabatic Full Passage (DAFP) sequences were used with 3D UTE MRI to create bound and pore water concentration maps in 40 cadaveric forearms (mean age 80yrs, 20male/20female)4,5. Scans were acquired at 1 mm isotropic resolution with a 250 mm FOV. Bones were segmented by slice using a polar transformation method described by Rad et. al.8 through approximately 13 mm in the axial direction at the distal third location. The segmentation was repeated after increasing the amount of endosteal region that was classified as cortical bone. These results were then compared with biomechanical measurements acquired from a three-point bend test on the same bones. Spearman correlation coefficients (ρ) were used to analyze how the characteristics of the distribution of bound and pore water (e.g. mean, skewness) in a given ROI relate to fracture risk and how these relationships depend on the segmentation.

Results

Figure 1 shows bound and pore water concentration maps and corresponding histograms from bones with high and low bending strength. The changes in the histograms between segmentation methods are apparent across both strong and weak bones, though are more significant with weaker bones. Across all 40 bones, the segmentation that excludes more of the endosteal region (without zone 2) resulted in an increase in mean bound water concentration per voxel (i.e. a shift to the right in the histogram) by an average of 8% (up to 17%) and a decrease in mean pore water per voxel (i.e. a shift to the left) by 49% (up to 60%). This highlights the sensitivity of these measures to ROI placement. Figure 2 shows correlations of bending strength with the mean bound and pore water concentrations and the skewness of the bound and pore water distributions. Correlations were significant with both large and small endosteal regions. The skewness of both bound and pore water data gave high correlations to strength, especially when the larger endosteal region (zone 1+2) was used. While the difference in correlation coefficients for the mean values are moderate, the correlation coefficients for skewness change markedly depending on whether or not the endosteal region is included. Multivariate analysis may give further insight into predicting fracture risk with these measurements of bound and pore water concentrations.

Discussion

These results emphasize the importance of the segmentation method on assessing bound and pore water maps of bone. Specifically, the same image has the potential to give drastically different results depending on the extent to which the segmentation method includes bone within the endosteal region. While this may not severely affect correlations with the mean bound and pore water concentrations, it is important that segmentation is not user dependent and is done the same way every time to get reliable results. It is also important to look beyond the mean concentrations, especially given the skewness of the distribution of concentrations within an ROI. Bones with a high positive skewness of pore water (shift towards lower concentrations) show higher strength, as well as bones with a high negative skewness of bound water (shift towards higher concentrations). Though including more of the endosteal region risks over-/under-estimating the mean pore/bound water concentrations by including voxels that may include marrow space, it also is capturing essential information that is sensitive to the degradation of bone. Dependable analyses of these maps are critical for future studies with bound and pore water MRI.

Acknowledgements

No acknowledgement found.

References

1. Nyman JS, Ni Q, Nicolella DP, et al. Measurements of mobile and bound water by nuclear magnetic resonance correlate with mechanical properties of bone. Bone 2008;42:193–9.

2. Horch RA, Gochberg DF, Nyman JS, et al. Non-invasive predictors of human cortical bone mechanical properties: T(2)-discriminated H NMR compared with high resolution X-ray. PLoS One 2011;6:e16359.

3. Bae WC, Chen PC, Chung CB, et al. Quantitative ultrashort echo time (UTE) MRI of human cortical bone: correlation with porosity and biomechanical properties. J Bone Min Res 2012;27:848–57.

4. Manhard MK, Horch RA, Harkins KD, et al. Validation of quantitative bound- and pore-water imaging in cortical bone. Magn Reson Med 2014;71:2166–2171.

5. Manhard MK, Horch RA, Gochberg DF, et al. In vivo quantitative MR imaging of bound and pore water in cortical bone. Radiology 2015;277:221–229.

6. Rajapakse CS, Bashoor-Zadeh M, Li C, et al. Volumetric cortical bone porosity assessment with MR imaging: Validation and clinical feasibility. Radiology 2015;276:526–535.

7. Lauretani F, Bandinelli S, Griswold ME, et al. Longitudinal changes in BMD and bone geometry in a population-based study. J Bone Min Res 2008;23:400–408.

8. Rad HS, Lam SCB, Magland JF, et al. Quantifying cortical bone water in vivo by three-dimensional ultra-short echo-time MRI. NMR Biomed 2011;24:855–64.

Figures

UTE images of a radius with high (top) and low (bottom) bending strength and the two lines for different segmentations shown. The plots show histograms (in % of total) for the bound and pore water concentrations found from the two segmentations. The blue shows histograms for the total bone area including the larger endosteal region (zone 1 + zone 2) and the red shows the smaller endosteal region only (zone 1).

Correlations of mean bound and pore water (left) and skewness of bound and pore water distributions (right) with bending strength. Blue lines show total bone area including the larger endosteal region and red lines shows the smaller endosteal region.



Proc. Intl. Soc. Mag. Reson. Med. 24 (2016)
0262