(D-139) Structural and Material Properties of the Second Metatarsal

Introduction: Bone stress injuries (BSI) are common overuse injuries disproportionately affecting female military recruits and athletes. BSI exist¬ on a continuum due to high magnitude, repetitive stress causing microcracks, temporarily weakening the bone and leading to full fractures if not given adequate time for repair[1]. Bone strength is directly related to structural geometry and material properties[2]. The distal second metatarsal (MET2) is a common site for BSI relative to other metatarsals, in part due to its unique geometric properties, such as smaller cross-sectional area[3]. Smaller bone geometry reduces resistance to deformation and increases risk of BSI. 3D Slicer is a free open-source software that has been developed to visualize, process and segment biomedical 3D images, like those from computed tomography (CT), allowing researchers to quantify tissue characteristics and create 3D models. A 3D Slicer extension called SegmentGeometry was recently developed to compute geometric values over the length of 3D models. To our knowledge, 3D Slicer is the only software with a readily available extension to quantify multi-slice cross-sectional bone geometry. However, application to bone has been limited to non-humans. The purpose of our preliminary investigation was to develop a bone segmentation workflow in 3D slicer and MATLAB to 1) accurately quantify structural and material characteristics of MET2 and 2) segment MET2 regions of interest (ROI) for input to a cross sectional finite element model estimating bone strain during dynamic activity.

Materials and Methods: Two female athletes (23/24yr, 58.5/63.3kg, 160/171.5cm) were recruited to establish an acquisition and analysis workflow for a larger study of 22 participants. Using a CT scanner (Optima CT660, GE MEDICAL SYSTEMS), axial view images were taken of the left foot with an inline phantom (QRM, PTW Company, Model BDC/6). Scanner parameters were 120kVp, 160 mAs, 0.625mm slice thickness, and 512x512 matrix. CT scan DICOM images were uploaded to 3D Slicer (version 5.2.2) which automatically reconstructed coronal and sagittal plane views. Total (cortical + trabecular) and cortical bone of the foot were segmented semi-automatically using threshold values of 150 Hounsfield Units (HU) to max and 650 HU to max, respectively[4]. MET2 was isolated from the rest of the foot using scissors and eraser tools. Using SegmentGeometry[5], geometric properties of MET2 were characterized at 1% intervals of axial bone length. Data in our ROI, distal 33% and 50% cross-sections, were extracted. The inline phantom with six rods of known calcium hydroxyapatite was used to convert image HU to bone mineral density (BMD) using subject-specific linear relationships between segment HU and density (g/cm3)[6]. To assess accuracy of our workflow, geometries and BMD (in both HU and g/cm3) for MET2 and ROI were compared to literature mean and standard deviation (SD) values. Using the “Create DICOM series” tool, reconstructed coronal view images of our ROI were exported into VA-BATTS, a MATLAB based software, for creation of subject-specific cross-sectional finite element meshes with inhomogeneous material properties to estimate stresses in bone sections[7,8].

Results: Within 3D Slicer, 27 structural properties were calculated; most relevant to MET2 BSI are total cross-sectional area (TA), cortical cross-sectional area (CSA), polar stress-strain index (SSIp), and section modulus (Zmajor). At 33% ROI, mean values were 93.78mm2 for TA, 62.43 mm2 for CSA, 246.23 for SSIp, and 132.69mm3 for Zmajor. At 50% ROI, mean TA (72.97mm2), CSA (46.75mm2), SSIp (170.68), and Zmajor (87.08mm3) fell within 1-2.1 SD of reported values[9,2,10]. Total bone HU, prior to conversion to BMD (g/cm3), were within 1 SD of previously reported BMD means[11] for whole bone (612.56 HU) and MET2 shaft (801.55 HU). However, distal 33% ROI BMD (755.07 HU) was 3 SD greater than reported values[11]. After conversion, cortical BMD at 50% ROI averaged 1.691g/cm3, whole bone cortical BMD averaged 1.624g/cm3, and whole bone total BMD averaged 1.373g/cm3. All were greater than 3 SD from mean values in reported literature[10,9]. To address our second aim, the reconstructed cross-sectional DICOM file at 33% ROI, a common site of BSI, was successfully read into VA BATTS, where subject specific inhomogeneous material properties will be incorporated into a 96 element mesh (Figure).

Discussion: A low-cost workflow that successfully quantified MET2 structural characteristics and segmented a ROI for finite element analysis was developed. Structural properties of MET2 and segment HU were comparable to those reported in the literature, especially in young adults in vivo. Our BMD values in g/cm3, did not compare well with previous reports. Studies reporting BMD in g/cm3 used pQCT and cadaver feet from older adults (24 males, 8 females; 36-93yr). Differences compared to literature could be due to scanner settings, populations of study, or BMD calibration methods (in-line vs phantom-less calibrations).

Conclusion: Structural and material properties are important to understanding physical activity related MET2 BSI. The 3D Slicer SegmentGeometry workflow created with open software enables analysis of BSI risk factors using clinical images. Additionally, the workflow developed can be used to form subject-specific bone models for use in finite element strain analysis during high-risk physical activities, like running and loaded marching, to better understand MET2 BSI etiology and identify targets for intervention.



Acknowledgements (Optional): :

This investigation was funded in part by an East Carolina University Undergraduate Research and Creativity Award. Rachel Golden contributed to preliminary testing, Dr. Ryan Steinbaker with CT acquisition, Dr. Alex Vadati with phantom calibration, and Landon Begin with initial image processing.

This  material  is  based  upon  work  supported  by  the  National  Science Foundation  under  Grant No. 1950507 Any opinions , findings and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily  reflect  the  views  of  the  National  Science  Foundation

 

References (Optional): :

1. Koltun, K. J., Sekel, N. M., Bird, M. B., Lovalekar, M., Mi, Q., Martin, B. J., & Nindl, B. C. (2022). Tibial bone geometry is associated with bone stress injury during military training in men and women. Frontiers in Physiology, 13, 135. 


2. Griffin, N. L., & Richmond, B. G. (2005). Cross-sectional geometry of the human forefoot. Bone, 37(2), 253-260. 


3. Chuckpaiwong, B., Cook, C., Pietrobon, R., & Nunley, J. A. (2007). Second metatarsal stress fracture in sport: comparative risk factors between proximal and non-proximal locations. British journal of sports medicine, 41(8), 510-514. 


4. Neves, C.A., Tran, E.D., Kessler, I.M. et al. Fully automated preoperative segmentation of temporal bone structures from clinical CT scans. Sci Rep 11, 116 (2021). https://doi.org/10.1038/s41598-020-80619-0 


5. Huie, J. M., Summers, A. P., & Kawano, S. M. (2022). Integrative Organismal Biology. 


6. Winsor, C., Li, X., Qasim, M., Henak, C. R., Pickhardt, P. J., Ploeg, H., & Viceconti, M. (2021). Evaluation of patient tissue selection methods for deriving equivalent density calibration for femoral bone quantitative CT analyses. Bone, 143, 115759. 


7. Kourtis LC, Kesari H, Carter DR, Beaupre GS: Transverse and torsional shear stresses in prismatic bodies having inhomogeneous material properties using a new 2D stress function. Journal of Mechanics of Materials and Structures, 4(4):659–674, 2009. 


8. Kourtis LC, Carter DR, Kesari H, Beaupre GS: A new software tool (VA-BATTS) to calculate bending, axial, torsional, and transverse shear stresses within bone cross sections having inhomogeneous material properties. Comput Methods Biomech Biomed Engin, 11(5):463-476, 2008. 


9. Muehleman, C., Lidtke, R., Berzins, A., Becker, J. H., Shott, S., & Sumner, D. R. (2000). Contributions of bone density and geometry to the strength of the human second metatarsal. Bone, 27(5), 709-714. 


10. Chaplais, E., Greene, D., Hood, A. et al. Reproducibility of a peripheral quantitative computed tomography scan protocol to measure the material properties of the second metatarsal. BMC Musculoskelet Disord 15, 242 (2014). https://doi.org/10.1186/1471-2474-15-242 


11. Commean, P. K., Ju, T., Liu, L., Sinacore, D. R., Hastings, M. K., & Mueller, M. J. (2009). Tarsal and metatarsal bone mineral density measurement using volumetric quantitative computed tomography. Journal of Digital Imaging, 22, 492-502.