In vivo quantitative assessment of proximal femoral cortical bone microstructure using double-echo ultrashort echo time magnetic resonance imaging

Introduction

Osteoporosis is a skeletal metabolic disorder characterized by impaired bone strength, predisposing a person to increased risk of fracture due to minor trauma that would not cause fracture in healthy people (1). Cortical bone comprises 80% of bone mass (2,3), which is key to the structural stability of whole bone. Cortical bone porosity is a key determinant of overall bone strength (4) and a quantitative marker of bone loss and fragility (5). Cortical bone thickness (CbTh), which is related to bone stiffness and fracture incidence, can provide at least marginal additional information about bone health not captured by porosity (6,7). Therefore, in vivo assessment of these two cortical bone microstructure measures is essential.

Although bone mineral density (BMD) measured by dual-energy X-ray absorptiometry (DXA) is the standard for the clinical diagnosis of osteoporosis, it is unable to assess cortical bone microstructure due to its nature as the “density” obtained from the two-dimensional projection (3). High-resolution peripheral quantitative computed tomography (HR-pQCT) enables in vivo assessment of cortical bone microstructure in humans (8). However, its limited resolution prevents direct imaging of smaller pores in the Haversian canals or the lacunae and canaliculi, which leads to the underestimation of porosity (7,9,10). Moreover, the limitation of HR-pQCT to distal skeletal sites and the presence of radiation exposure also limit its application (7,10).

With the development of magnetic resonance imaging (MRI) technology, ultrashort echo time (UTE) MRI has overcome the limitation of conventional MRI in capturing bone water proton signals, enabling the in vivo quantitative assessment of cortical bone without radiation (9). Water in cortical bone consists of free water in pores (i.e., pore water, T2 value, >1 ms) and bound water connected to the collagen matrix (T2 value, ~0.3–0.4 ms) (9,11,12). Rajapakse et al. proposed to quantify the pore water fraction using the porosity index (PI) obtained by double-echo UTE, which was highly positively correlated with the cortical porosity measured by microcomputed tomography (micro-CT) and negatively correlated with total and cortical BMD derived by pQCT (9). The PI was calculated as the ratio of image intensities of the second echo (indicating signals mainly from pore water) to the first echo (indicating signals from all water), which saves time by not requiring complex post-processing and has excellent clinical applicability compared to other UTE quantification methods (9,10). Currently, most research on the use of double-echo UTE in cortical bone has focused on distal extremities, which is no substitute for the assessment of cortical microstructure in the proximal femur (10). The proximal femur is at higher risk of osteoporotic fracture compared to the distal extremities, and the consequences are often more severe (13). Previously, Chen et al. achieved in vivo analysis of femoral neck porosity using double-echo UTE (10). However, given the low contrast and signal-to-noise ratio (SNR) of the femoral neck cortical bone due to it being a small and deep structure (7,10), Jones et al. proposed UTE imaging of the proximal femoral shaft cortical bone below the lesser trochanter, which has higher contrast and is easier to reliably contour on UTE images, as a useful surrogate (7). Additionally, it was demonstrated by an in vitro study of cadaveric femoral specimens that the UTE measurements here can be a useful indicator for evaluating the quality of the whole femur (7). However, in vivo application studies of double-echo UTE at this site are lacking, and considering the possible influence of limb dominance, sex, age, and body mass index (BMI) on the microstructure of cortical bone (10,14,15), the effects of these factors on the corresponding UTE measurements at this site remain to be explored.

Therefore, herein, we conducted a preliminary study of in vivo quantitative assessment of proximal femoral cortical bone microarchitecture by double-echo UTE MRI in healthy adults without osteoporosis risk factors, aiming to evaluate the repeatability of this technique and explore the effects of possible influencing factors such as limb dominance, sex, age, and BMI on UTE measurements, so as to provide a reference for the further application of this technique in clinical practice in the future. We present this article in accordance with the STROBE reporting checklist (available at https://qims.amegroups.com/article/view/10.21037/qims-24-1230/rc).

Methods

Study population

This study was conducted in accordance with the Declaration of Helsinki (as revised in 2013). This cross-sectional study was approved by the Institutional Review Board (IRB) of Tongji Hospital of Tongji Medical College of Huazhong University of Science and Technology (No. TJ-IRB20211294), and written informed consent was provided by every participant before the scanning. For sample size calculations, we used G*Power software (version 3.1, Universität Düsseldorf, Düsseldorf, Germany) and determined effect size d with reference to the differences between male and female UTE measurements in Chen et al.’s study (10). The significance level was set at α=0.05, the statistical test was two-tailed, the statistical power 1-β was set to 0.9, the number of groups was two (male and female), the effect size d=1.1, and the least total sample size required was 42.

From November 2021 to May 2022, 57 healthy individuals aged 18–55 years were recruited from staff, students, and their family members in our hospital. The exclusion criteria were as follows: MRI contraindications; failure to complete MR scans; fractures within six months; history of lumbar spine or lower extremity trauma or surgery; histories of diseases or treatments that may affect bone metabolism (e.g., hyperparathyroidism, chronic kidney disease, diabetes, rheumatic diseases, malignant tumors, etc., and use of bisphosphonates, glucocorticoids, thyroid hormone, calcitonin, and estrogens); and postmenopausal females.

UTE MRI scanning

UTE MR imaging was performed on a 3.0-T MRI system (Signa Pioneer; GE Healthcare, Milwaukee, WI, USA) using a body coil and bed spine coil simultaneously. A three-dimensional (3D) double-echo UTE-Cones sequence was adopted for scanning proximal femurs, allowing simultaneous imaging of bilateral proximal femurs in a single scan. The 3D UTE-Cones sequence utilizes a hard radiofrequency (RF) pulse for non-selective excitation, and a center-out 3D spiral k-sampling trajectory, allowing for immediate data acquisition following the RF pulse (7,9,10). The scan parameters were as follows: repetition time (TR)/echo time (TE) =11.3/0.032, 4.4 ms; flip angle =15°; field of view (FOV) =300×300 mm²; in-plane spatial resolution =1.0 mm × 1.0 mm; bandwidth =62.5 kHz; slice thickness =3 mm; number of axial slices =16; and scan time =3 min 6 s. TE (0.032 ms) of the first echo was the shortest TE governed by sequence implementation and hardware constraints, shorter than the shortest TE achieved by some other studies (7,10,16), allowing for the simultaneous acquisition of bound and pore water signals and achieving a higher SNR (17-19). TE (4.4 ms) of the second echo balanced the requirements of scan time and image quality for clinical applicability. The upper boundary of the scan was positioned at the lesser trochanter of the femur, and the proximal femur below it was scanned. Limb dominance was defined as the leg used for climbing stairs, which was chosen in a self-determined way (10).

Image analysis

All images were processed by ImageJ, version 1.8.0 (National Institutes of Health, Bethesda, MD, USA). We obtained ratio images of the second and first echo of proximal femurs, namely, PI maps, according to the equation: PI (%) = (TE_long intensity/TE_short intensity) ×100 (9). The region of interest (ROI) was manually segmented along the bone cortical boundary on the first echo image (TE_short =0.032 ms), which showed the bone cortical boundary more clearly, and then copied to the PI map to measure the PI value. ROIs were limited to the compact-appearing cortex that excluded transitional zones, as most of the UTE signals in transitional zones were derived from adipose tissue, interfering with the cortical water signal and thus with the PI measurement. We selected a 1.2-cm (5 slices × 3 mm) region inferior to the lesser trochanter of femur for ROI outlining as the thicker and denser cortex at this site allows for reliable segmentation of the proximal femoral cortex compared to the femoral neck, which has thin cortex and low contrast with the surrounding trabecularized region (7). Additionally, CbTh was also obtained from the manual segmentations, and the CbTh of each slice was computed by using ImageJ’s Local Thickness function, which could provide a CbTh map (20).

The examples of PI and CbTh map and ROI are presented in Figure 1. The PI and CbTh of each slice were recorded as the average PI and CbTh of the ROI, and then average values of PI and CbTh of the five slices were calculated for subsequent analysis. Two musculoskeletal radiologists (one with 5 years and the other with 3 years of experience) who were blinded to the participants’ information independently drew ROIs and measured PIs and CbThs.

Figure 1 Double-echo UTE images and PI and CbTh maps of the proximal femur on the non-dominant side of a 25-year-old healthy male. (A) The first echo image (TE =0.032 ms); (B) the second echo image (TE =4.4 ms); (C) example of ROI outlining on the first echo image; (D) PI map of the compact-appearing cortex; (E) CbTh map of the compact-appearing cortex. When performing ROI outlining, care should be taken to outline along the outer and inner black boundaries of the bone cortex, especially to avoid the gray area near the inner edge, which represents the transition zones between the cortical and cancellous bone. UTE, ultrashort echo time; PI, porosity index; CbTh, cortical bone thickness; TE, echo time; ROI, region of interest.

Statistical analysis

All statistical analyses were performed using the software SPSS 24.0 (IBM Corp., Armonk, NY, USA). The Shapiro-Wilk test was used to determine whether each continuous variable was normally distributed. Differences in age and BMI between males and females were tested by the independent samples t-test (normally distributed variables: BMI) or the Mann-Whitney U test (non-normally distributed variable: age). Inter-reader repeatability for measurements was assessed using intra-class correlation coefficient (ICC) analysis (a two-way random model of absolute agreement). ICC values were interpreted as poor (less than 0.50), moderate (0.50–0.75), good (0.75–0.90), and excellent (greater than 0.90) repeatability (21). When comparing PI and CbTh of proximal femoral cortex on the dominant and non-dominant sides, the paired samples t-test was used for normally distributed data (PI); otherwise, the Wilcoxon signed-rank test was used (CbTh). For comparisons of PI and CbTh between sexes, analysis of covariance (ANCOVA) was used to exclude the influence of confounding factors (e.g., BMI). For all the above tests, the significance level was set at 0.05. The correlations between PI and age, PI and BMI, CbTh and age, and CbTh and BMI were analyzed by Pearson (normally distributed: correlations between PI and BMI, CbTh and BMI) or Spearman correlation coefficient (non-normally distributed: correlations between PI and age, CbTh and age). Meanwhile, the partial Pearson correlation coefficient was also computed for adjustment of age or BMI. Considering that the potential impact of multiple comparisons may arise from separate analyses for males and females, the Bonferroni correction was applied and the adjusted significance level was set at 0.025. Corresponding scatter plots were drawn to visualize the relationships between variables, and curve estimation was used for variables that might exhibit curvilinear correlation rather than linear correlation on the scatter plot.

Results

Participants

A total of 52 healthy cases (33 males and 19 females; mean age, 32.40±8.61 years; age range, 22–55 years) were finally included for analysis, as displayed in the flowchart in Figure 2. The general characteristics of all participants are presented in Table 1. No statistical difference was found in age between males and females (P=0.586), whereas the BMI of males was greater than that of females (P=0.007), which did not allow direct comparison of PI and CbTh between the groups.

Figure 2 The flowchart of 52 healthy individuals included for analysis. MRI, magnetic resonance imaging; MR, magnetic resonance.

Repeatability of measurements

The inter-reader ICCs of PI and CbTh measurements were, respectively, 0.985 [95% confidence interval (CI): 0.979–0.989] and 0.943 (95% CI: 0.856–0.977), which indicated excellent inter-reader repeatability.

Comparison of PI and CbTh by limb dominance and sex

Figure 3 presents the differences in PI and CbTh in the proximal femoral cortex between the dominant and non-dominant sides. The proximal femur on the non-dominant side had greater cortical PI (34.2%±3.4% vs. 32.2%±3.2%, P<0.001), whereas the difference in CbTh between the non-dominant and dominant sides was not statistically significant (6.60±0.84 vs. 6.49±1.01 mm, P=0.470). Differences in PI and CbTh between males and females are shown in Table 1, which were analyzed by ANCOVA to exclude the effect of BMI, and the dominant and non-dominant sides were analyzed separately. Males had greater PIs than females in both the non-dominant and dominant proximal femurs (non-dominant side: F=22.044, P<0.001; dominant side: F=4.877, P=0.032). CbTh of the non-dominant side proximal femur was greater in males than in females (F=4.639, P=0.036), whereas no significant difference was noted between different sexes in CbTh of the dominant proximal femur (F=4.000, P=0.051).

Figure 3 Comparisons of PI (A) and CbTh (B) in the proximal femoral cortex between the non-dominant and dominant sides. The box represents the interquartile range, whereas the solid line within it represents the median and the whiskers indicate 1.5 times the interquartile range. P<0.05 are highlighted in boldface font. PI, porosity index; CbTh, cortical bone thickness.

Correlations of PI and CbTh with age and BMI

The correlation and partial correlation coefficients between UTE measurement parameters and participants’ age and BMI are presented in Table 2. There was no statistically significant linear correlation with age for both dominant and non-dominant proximal femoral cortical PI in both males and females at the Bonferroni correction level of P<0.025. Similarly, no statistically significant correlation was found between proximal femoral CbTh and age in all groups. For BMI, proximal femoral cortical PI on the dominant side demonstrated a positive correlation with BMI in males (r=0.535, P=0.001, partial r=0.489, P=0.004 after adjustment for age); proximal femoral cortical CbTh on the non-dominant side was positively correlated with BMI in males (r=0.482, P=0.005, partial r=0.477, P=0.006 after adjustment for age). No statistically significant correlation was found between the UTE measurements of proximal femoral cortical bone and BMI in the other groups at the Bonferroni correction level of P<0.025. The scatter plots of all the associations are shown in Figure 4. Interestingly, there was a U-shaped curve trend in the relationship between dominant side cortical PI and age in females (y = 65.32 – 1.88x + 0.02x², R²=0.348, P=0.033), which initially tended to decrease with age but tended to increase with age after approximately 42 years of age, although there was no statistical significance at the Bonferroni correction level of P<0.025.

Figure 4 Scatter plots presenting correlations of PI and CbTh with age and BMI. Solid dots represent male data and hollow squares represent female data. The gray solid and dashed lines in (B) represent the curve fitting and 95% confidence interval for the female dominant side PI and age, respectively. Significant linear fittings are marked with black solid lines for male UTE measurements and BMI, and the black dashed lines represent 95% confidence intervals. * indicates significant at Bonferroni corrected value of P<0.025. PI, porosity index; CbTh, cortical bone thickness; BMI, body mass index.

Discussion

This study was the first attempt at in vivo investigation of cortical bone microarchitecture in the proximal femur by double-echo UTE technique. Cortical PI and CbTh of proximal femur were measured with excellent repeatability. Our results showed that the non-dominant side had greater PI; males had greater PI and CbTh compared to females; PI and CbTh were positively correlated with BMI in males; and the relationship between PI and female age showed a U-shaped trend. This study provides a new viable option for clinical application, and the results provide a reference for subsequent in vivo studies.

The double-echo UTE-based metrics have been validated in previous studies (7,9). Rajapakse et al., the proposers of the double-echo UTE method, have validated the correlation of PI with metrics obtained from micro-CT, which is the gold standard (22), in human tibia specimens, as well as its correlation with the total and cortical BMD obtained from pQCT (9). Jones et al. obtained PI and CbTh of the proximal femur consistent with our analysis site, and validated the correlations between cortical PI and BMD obtained from pQCT as well as whole-bone stiffness by in vitro studies (7). The double-echo UTE sequence used in this study has good clinical applicability, as it took only three minutes to scan and could be performed on the 3.0 T clinical scanner, which is the ideal field strength for pore water quantification (7). Furthermore, the cortical PI obtained from this sequence has self-normalizing properties, which minimizes the effect of variations in both RF and static field inhomogeneity (9), and does not require specialized adiabatic inversion pulses, customized scripts, or programmatic knowledge. The inter-reader agreement of proximal femoral PI and CbTh measurements in the present study was excellent and similar to that of Rajapakse et al. in tibial PI measurements (9), which reflects the excellent repeatability of the measurements.

In the present study, we found differences in cortical PI in different sides of the limbs, which may be related to differences in exercise loads in the dominant and non-dominant limbs. Bone structure adapts to habitual mechanical loading (23,24), especially the cortical bone, which accounts for 80% of the skeletal mass (25,26). With prolonged exposure to greater mechanical loads, the cortical structure of the proximal femur on the dominant side would be denser than that on the contralateral side, and the corresponding cortical porosity would be reduced compared with the contralateral side.

The difference in cortical bone PI between males and females found in our study may also have resulted from the difference in mechanical loading of the lower limbs due to the difference in gait between the two sexes. Kerrigan et al. found that females walked at greater cadences and greater hip flexion, and also perform greater mechanical work per unit time and distance (27). Several studies of femoral structure in participants of different sexes have demonstrated that this gait difference resulted in smaller cortical porosity in females compared to males (10,28). The difference in proximal femoral CbTh between sexes in the present study can be explained by differences in sex hormones: in females, estrogens inhibit periosteal apposition but stimulate endocortical apposition; in males, both androgens and estrogens stimulate periosteal bone expansion, resulting in a thicker cortex in males than in females (29,30). Comparative results consistent with ours have been reported in other studies on the cortical microstructure of lower limb load-bearing bones (10,31).

The effect of BMI on cortical bone microarchitecture has been unclear to date. BMI affects cortical bone microarchitecture through its influence on mechanical loading and adipose tissue (10,32): on the one hand, those with greater BMI tend to be subjected to greater mechanical load, which contributes to adaptive changes in cortical microarchitecture (23,24); on the other hand, increased fat mass also contributes to increased BMI, and adiposity may impair bone metabolic processes by promoting the production of inflammatory cytokines and increasing bone resorption, thereby affecting the cortical microstructure (33,34). In this study, we found a positive correlation between proximal femoral cortical PI and BMI in males. On the one hand, since studies have shown a causal association between higher levels of physical activity and lower BMI, and that higher BMI reduces physical activity (35,36), although a higher BMI leads to a greater load per physical activity, the associated lower levels of physical activity represent lower loading frequency, and ultimately the total load may be reduced. On the other hand, as mentioned above, the effect of fat on bone metabolism may also contribute to the increase in PI. As for CbTh, we found a positive correlation with BMI in males. The findings of Ng et al. (37) that a positive correlation between total body fat mass and proximal femur CbTh in young and middle-aged males may partially explain our results, but further studies on the relationship between BMI and CbTh are required. Correlations of proximal femoral cortical UTE measurements and BMI in females were not found in this study. Our study results were not entirely consistent with some other in vivo studies that involved BMI and cortical bone microstructure (10,34,37,38), primarily because the study populations and analysis sites were all different, making it hard to directly compare the results of the various studies. Thus, more in vivo research on the microstructure of the proximal femoral cortical bone is needed to further explore its relationship with BMI, and this study has provided a reference for subsequent research.

In this study, there was no linear correlation found between proximal femoral cortical UTE measurements and age. Some previous studies found that bone cortical porosity increased with age (39,40), which does not align with our results. These studies all included elderly, postmenopausal females, some of which focused only on middle-aged and older adults, and the analysis sites varied, which may partly explain the differences in results. Moreover, it was suggested that there were individual variations in cortical porosity, of which age explained only 7.1%, especially for males (41). For females, changes in cortical porosity with age are more pronounced in the postmenopausal than in the premenopausal state (39,40). Our study, which was conducted in young and middle-aged healthy people and excluded postmenopausal females, detected a possible U-shaped trend in changes in proximal femoral porosity with age in premenopausal healthy adult females. Although not statistically significant at the Bonferroni correction level of P<0.025, given that Bonferroni correction is so restrictive that the tests may be relatively conservative, we believe that this result is suggestive of a possible trend in cortical PI in females. We hypothesized that this might be related to changes in female hormone levels with age and so on, and future studies with larger sample sizes are needed to further validate and explore it.

There are still some limitations of this study yet to be addressed. First, this study cohort was small and only focused on young and middle-aged healthy people, as it was a preliminary study of double-echo UTE in vivo imaging of the proximal femur, which needs to be performed firstly in a healthy population without risk factors for osteoporosis. Despite our best efforts to ensure the reliability and validity of the tests during statistical analyses, the relatively limited sample size has constrained the validity of some results. Therefore, subsequent studies with a larger cohort and wider age range are needed for validation and further exploration. Second, the ROI was manually segmented in this study, and some automatic segmentation methods should be developed in subsequent studies to save image processing time. Third, the participants in this study did not undergo DXA or QCT, which are currently clinically available for evaluating bone health. DXA is unable to differentiate between cortical and trabecular bone, and the femur and spine BMD measured by it has been found to have no significant correlation with cortical porosity (9,42); as for the cortical BMD obtained by QCT, in vitro studies have validated its correlation with double-echo UTE-based metrics (7,9). Finally, micro-CT, considered the gold standard (22), was not performed in this study. Due to its high exposure to ionizing radiation and small scanning FOV, its use is limited to small animals and cadaveric bone specimens (22), making it infeasible and unethical to obtain micro-CT images of the proximal femur in humans in vivo. Additionally, previous studies have validated the correlation between parameters measured by double-echo UTE and the cortical bone microstructure obtained by micro-CT (7,9).

Conclusions

This study used double-echo UTE to perform in vivo quantitative evaluation of the proximal femoral cortical microstructure in healthy people, the excellent repeatability of each measurement parameter was verified, and differences in cortical porosity and thickness between different limbs and sexes were determined, as well as their correlation with BMI and age. It provides a reference for the subsequent in vivo quantitative studies of proximal femoral cortical bone, and represents a necessary preliminary study before applying this technique to clinical practice.

Acknowledgments

We express sincere gratitude to the people who have helped in this work. We acknowledge Weiyin Vivian Liu from GE Healthcare for the technical support. We also thank Donglin Wen for helping with the MR scans.

Funding: This research was supported by the National Natural Science Foundation of China (NSFC) (Nos. 81930045 and 82371938).

Footnote

Reporting Checklist: The authors have completed the STROBE reporting checklist. Available at https://qims.amegroups.com/article/view/10.21037/qims-24-1230/rc

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://qims.amegroups.com/article/view/10.21037/qims-24-1230/coif). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. This study was conducted in accordance with the Declaration of Helsinki (as revised in 2013). This cross-sectional study was approved by the Institutional Review Board (IRB) of Tongji Hospital of Tongji Medical College of Huazhong University of Science and Technology (No. TJ-IRB20211294), and written informed consent was provided by every participant before the scanning.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Sözen T, Özışık L, Başaran NÇ. An overview and management of osteoporosis. Eur J Rheumatol 2017;4:46-56. [Crossref] [PubMed]
2. Zebaze RM, Ghasem-Zadeh A, Bohte A, Iuliano-Burns S, Mirams M, Price RI, Mackie EJ, Seeman E. Intracortical remodelling and porosity in the distal radius and post-mortem femurs of women: a cross-sectional study. Lancet 2010;375:1729-36. [Crossref] [PubMed]
3. Johannesdottir F, Turmezei T, Poole KE. Cortical bone assessed with clinical computed tomography at the proximal femur. J Bone Miner Res 2014;29:771-83. [Crossref] [PubMed]
4. Morgan EF, Unnikrisnan GU, Hussein AI. Bone Mechanical Properties in Healthy and Diseased States. Annu Rev Biomed Eng 2018;20:119-43. [Crossref] [PubMed]
5. Bala Y, Zebaze R, Seeman E. Role of cortical bone in bone fragility. Curr Opin Rheumatol 2015;27:406-13. [Crossref] [PubMed]
6. Kral R, Osima M, Borgen TT, Vestgaard R, Richardsen E, Bjørnerem Å. Increased cortical porosity and reduced cortical thickness of the proximal femur are associated with nonvertebral fracture independent of Fracture Risk Assessment Tool and Garvan estimates in postmenopausal women. PLoS One 2017;12:e0185363. [Crossref] [PubMed]
7. Jones BC, Jia S, Lee H, Feng A, Shetye SS, Batzdorf A, Shapira N, Noël PB, Pleshko N, Rajapakse CS. MRI-derived porosity index is associated with whole-bone stiffness and mineral density in human cadaveric femora. Bone 2021;143:115774. [Crossref] [PubMed]
8. Chen H, Zhou X, Shoumura S, Emura S, Bunai Y. Age- and gender-dependent changes in three-dimensional microstructure of cortical and trabecular bone at the human femoral neck. Osteoporos Int 2010;21:627-36. [Crossref] [PubMed]
9. Rajapakse CS, Bashoor-Zadeh M, Li C, Sun W, Wright AC, Wehrli FW. Volumetric Cortical Bone Porosity Assessment with MR Imaging: Validation and Clinical Feasibility. Radiology 2015;276:526-35. [Crossref] [PubMed]
10. Chen M, Yuan H. Assessment of porosity index of the femoral neck and tibia by 3D ultra-short echo-time MRI. J Magn Reson Imaging 2018;47:820-8. [Crossref] [PubMed]
11. Horch RA, Nyman JS, Gochberg DF, Dortch RD, Does MD. Characterization of 1H NMR signal in human cortical bone for magnetic resonance imaging. Magn Reson Med 2010;64:680-7. [Crossref] [PubMed]
12. Bae WC, Chen PC, Chung CB, Masuda K, D'Lima D, Du J. Quantitative ultrashort echo time (UTE) MRI of human cortical bone: correlation with porosity and biomechanical properties. J Bone Miner Res 2012;27:848-57. [Crossref] [PubMed]
13. Cummings SR, Melton LJ. Epidemiology and outcomes of osteoporotic fractures. Lancet 2002;359:1761-7. [Crossref] [PubMed]
14. Yu F, Xu Y, Hou Y, Lin Y, Jiajue R, Jiang Y, Wang O, Li M, Xing X, Zhang L, Qin L, Hsieh E, Xia W. Age-, Site-, and Sex-Specific Normative Centile Curves for HR-pQCT-Derived Microarchitectural and Bone Strength Parameters in a Chinese Mainland Population. J Bone Miner Res 2020;35:2159-70. [Crossref] [PubMed]
15. Faulkner RA, Houston CS, Bailey DA, Drinkwater DT, McKay HA, Wilkinson AA. Comparison of bone mineral content and bone mineral density between dominant and non-dominant limbs in children 8-16 years of age. Am J Hum Biol 1993;5:491-9. [Crossref] [PubMed]
16. Manhard MK, Horch RA, Gochberg DF, Nyman JS, Does MD. In Vivo Quantitative MR Imaging of Bound and Pore Water in Cortical Bone. Radiology 2015;277:221-9. [Crossref] [PubMed]
17. Xiong Y, He T, Wang Y, Liu WV, Hu S, Zhang Y, Wen D, Hou B, Li Y, Zhang P, Liu J, He F, Li X. CKD Stages, Bone Metabolism Markers, and Cortical Porosity Index: Associations and Mediation Effects Analysis. Front Endocrinol (Lausanne) 2021;12:775066. [Crossref] [PubMed]
18. Weiger M, Pruessmann KP. Short-T(2) MRI: Principles and recent advances. Prog Nucl Magn Reson Spectrosc 2019;114-115:237-70. [Crossref] [PubMed]
19. Carl M, Bydder GM, Du J. UTE imaging with simultaneous water and fat signal suppression using a time-efficient multispoke inversion recovery pulse sequence. Magn Reson Med 2016;76:577-82. [Crossref] [PubMed]
20. Dougherty R, Kunzelmann K-H. Computing Local Thickness of 3D Structures with ImageJ. Microscopy and Microanalysis 2007;13:1678-9.
21. Koo TK, Li MY. A Guideline of Selecting and Reporting Intraclass Correlation Coefficients for Reliability Research. J Chiropr Med 2016;15:155-63. [Crossref] [PubMed]
22. Rajapakse CS, Magland JF, Wald MJ, Liu XS, Zhang XH, Guo XE, Wehrli FW. Computational biomechanics of the distal tibia from high-resolution MR and micro-CT images. Bone 2010;47:556-63. [Crossref] [PubMed]
23. Ruff C, Holt B, Trinkaus E. Who's afraid of the big bad Wolff?: "Wolff's law" and bone functional adaptation. Am J Phys Anthropol 2006;129:484-98. [Crossref] [PubMed]
24. Abe S, Narra N, Nikander R, Hyttinen J, Kouhia R, Sievänen H. Exercise loading history and femoral neck strength in a sideways fall: A three-dimensional finite element modeling study. Bone 2016;92:9-17. [Crossref] [PubMed]
25. Seeman E. Age- and menopause-related bone loss compromise cortical and trabecular microstructure. J Gerontol A Biol Sci Med Sci 2013;68:1218-25. [Crossref] [PubMed]
26. Bala Y, Zebaze R, Ghasem-Zadeh A, Atkinson EJ, Iuliano S, Peterson JM, Amin S, Bjørnerem Å, Melton LJ 3rd, Johansson H, Kanis JA, Khosla S, Seeman E. Cortical porosity identifies women with osteopenia at increased risk for forearm fractures. J Bone Miner Res 2014;29:1356-62. [Crossref] [PubMed]
27. Kerrigan DC, Todd MK, Della Croce U. Gender differences in joint biomechanics during walking: normative study in young adults. Am J Phys Med Rehabil 1998;77:2-7. [Crossref] [PubMed]
28. Bell KL, Loveridge N, Power J, Garrahan N, Meggitt BF, Reeve J. Regional differences in cortical porosity in the fractured femoral neck. Bone 1999;24:57-64. [Crossref] [PubMed]
29. Seeman E. Pathogenesis of bone fragility in women and men. Lancet 2002;359:1841-50. [Crossref] [PubMed]
30. Nieves JW. Sex-Differences in Skeletal Growth and Aging. Curr Osteoporos Rep 2017;15:70-5. [Crossref] [PubMed]
31. Hansen S, Shanbhogue V, Folkestad L, Nielsen MM, Brixen K. Bone microarchitecture and estimated strength in 499 adult Danish women and men: a cross-sectional, population-based high-resolution peripheral quantitative computed tomographic study on peak bone structure. Calcif Tissue Int 2014;94:269-81. [Crossref] [PubMed]
32. Deminger A, Klingberg E, Lorentzon M, Hedberg M, Carlsten H, Jacobsson LTH, Forsblad-d'Elia H. Factors associated with changes in volumetric bone mineral density and cortical area in men with ankylosing spondylitis: a 5-year prospective study using HRpQCT. Osteoporos Int 2022;33:205-16. [Crossref] [PubMed]
33. Cao JJ. Effects of obesity on bone metabolism. J Orthop Surg Res 2011;6:30. [Crossref] [PubMed]
34. Sundh D, Rudäng R, Zoulakis M, Nilsson AG, Darelid A, Lorentzon M. A High Amount of Local Adipose Tissue Is Associated With High Cortical Porosity and Low Bone Material Strength in Older Women. J Bone Miner Res 2016;31:749-57. [Crossref] [PubMed]
35. Richmond RC, Davey Smith G, Ness AR, den Hoed M, McMahon G, Timpson NJ. Assessing causality in the association between child adiposity and physical activity levels: a Mendelian randomization analysis. PLoS Med 2014;11:e1001618. [Crossref] [PubMed]
36. Carrasquilla GD, García-Ureña M, Fall T, Sørensen TIA, Kilpeläinen TO. Mendelian randomization suggests a bidirectional, causal relationship between physical inactivity and adiposity. Elife 2022;11:e70386. [Crossref] [PubMed]
37. Ng AC, Melton LJ 3rd, Atkinson EJ, Achenbach SJ, Holets MF, Peterson JM, Khosla S, Drake MT. Relationship of adiposity to bone volumetric density and microstructure in men and women across the adult lifespan. Bone 2013;55:119-25. [Crossref] [PubMed]
38. Evans AL, Paggiosi MA, Eastell R, Walsh JS. Bone density, microstructure and strength in obese and normal weight men and women in younger and older adulthood. J Bone Miner Res 2015;30:920-8. [Crossref] [PubMed]
39. Burghardt AJ, Kazakia GJ, Ramachandran S, Link TM, Majumdar S. Age- and gender-related differences in the geometric properties and biomechanical significance of intracortical porosity in the distal radius and tibia. J Bone Miner Res 2010;25:983-93. [Crossref] [PubMed]
40. Vilayphiou N, Boutroy S, Sornay-Rendu E, Van Rietbergen B, Chapurlat R. Age-related changes in bone strength from HR-pQCT derived microarchitectural parameters with an emphasis on the role of cortical porosity. Bone 2016;83:233-40. [Crossref] [PubMed]
41. Stein MS, Feik SA, Thomas CD, Clement JG, Wark JD. An automated analysis of intracortical porosity in human femoral bone across age. J Bone Miner Res 1999;14:624-32. [Crossref] [PubMed]
42. Carvalho C, Magalhães J, Neto R, Pereira L, Branco P, Adragão T, Frazão JM. Cortical bone analysis in a predialysis population: a comparison with a dialysis population. J Bone Miner Metab 2017;35:513-21. [Crossref] [PubMed]