Influence of bone lesion location on femoral bone strength assessed by MRI-based finite-element modeling

The effect of deep learning-based lesion segmentation on failure load calculations of metastatic femurs using finite element analysis

Bone ranks as the third most frequent tissue affected by cancer metastases, following the lung and liver. Bone metastases are often painful and may result in pathological fracture, which is a major cause of morbidity and mortality in cancer patients. To quantify fracture risk, finite element (FE) analysis has shown to be a promising tool, but metastatic lesions are typically not specifically segmented and therefore their mechanical properties may not be represented adequately. Deep learning methods potentially provide the opportunity to automatically segment these lesions and change the mechanical properties more adequately. In this study, our primary focus was to gain insight into the performance of an automatic segmentation algorithm for femoral metastatic lesions using deep learning methods and the subsequent effects on FE outcomes. The aims were to determine the similarity between manual segmentation and automatic segmentation; the differences in predicted failure load between FE models with automatically segmented osteolytic and mixed lesions and the models with CT-based lesion values (the gold standard); and the effect on the BOne Strength (BOS) score (failure load adjusted for body weight) and subsequent fracture risk assessments. From two patient cohorts, a total number of 50 femurs with osteolytic and mixed metastatic lesions were included in this study. The femurs were segmented from CT images and transferred into FE meshes. The material behavior was implemented as non-linear isotropic. These FE models were considered as gold standard (Finite Element no Segmented Lesion: FE-no-SL), whereby the local calcium equivalent density of both femur and metastatic lesion was extracted from CT-values. Lesions in the femur were manually segmented by two biomechanical experts after which final lesion segmentation for each femur was obtained based on consensus of opinions between two observers. Subsequently, a self-configuring variant of the popular deep learning model U-Net known as nnU-Net was used to automatically segment metastatic lesions within the femur. For these models with segmented lesions (Finite Element with Segmented Lesion: FE-with-SL), the calcium equivalent density within the metastatic lesions was set to zero after being segmented by the neural network, simulating absence of load-bearing capacity of these lesions. The models (either with or without automatically segmented lesions) were loaded incrementally in axial direction until failure was simulated. Dice coefficient was used to evaluate the similarity of the manual and automatic segmentation. Mean calcium equivalent density values within the automatically segmented lesions were calculated. Failure loads and patterns were determined. Furthermore, sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were calculated for both groups by comparing the predictions to the occurrence or absence of actual fracture within the patient cohorts. The automatic segmentation algorithm performed in a none-robust manner. Dice coefficients describing the similarity between consented manual and automatic segmentations were relatively low (mean 0.45 ± standard deviation 0.33, median 0.54). Failure load difference between the FE-no-SL and FE-with-SL groups varied from 0 % to 48 % (mean 6.6 %). Correlation analysis of failure loads between the two groups showed a strong relationship (R2 > 0.9). From the 50 cases, four cases showed clear deviations for which models with automatic lesion segmentation (FE-with-SL) showed considerably lower failure loads. In the whole database including osteolytic and mixed lesions, sensitivity and NPV remained the same, but specificity and PPV decreased from 94 % to 83 %, and from 78 % to 54 % respectively from FE-no-SL to FE-with-SL. This study indicates that the nnU-Net yielded none-robust outcomes in femoral lesion segmentation and that other segmentation algorithms should be considered. However, the difference in failure pattern and failure load between FE models with automatically segmented osteolytic and mixed lesions were relatively small in most cases with a few exceptions. On the other hand, the accuracy of fracture risk assessment using the BOS score was lower compared to the FE-no-SL. In conclusion, this study showed that automatic lesion segmentation is a none-solved issue and therefore, quantifying lesion characteristics and the subsequent effect on the fracture risk using deep learning will remain challenging.

Predicting Fractures Using Vertebral 18F-NaF Uptake in Prostate Cancer Patients

BACKGROUND

Patients with prostate cancer tend to be at heightened risk for fracture due to bone metastases and treatment with androgen-deprivation therapy. Bone mineral density (BMD) derived from dual energy X-ray absorptiometry (DXA) is the standard for determining fracture risk in this population. However, BMD often fails to predict many osteoporotic fractures. Patients with prostate cancer also undergo 18F-sodium fluoride (18F-NaF)-positron emission tomography/computed tomography (PET/CT) to monitor metastases. The purpose of this study was to assess whether bone deposition, assessed by 18F-NaF uptake in 18F-NaF PET/CT, could predict incident fractures better than DXA- or CT-derived BMD in patients with prostate cancer.

METHODS

This study included 105 males with prostate cancer who had undergone full body 18F-NaF PET/CT. Standardized uptake value (SUVmean and SUVmax) and CT-derived Hounsfield units (HU), a correlate of BMD, were recorded for each vertebral body. The average SUVmean, SUVmax, and HU were calculated for cervical, thoracic, lumbar, and sacral areas. The t-test was used to assess significant differences between fracture and no-fracture groups.

RESULTS

The SUVmean and SUVmax values for the thoracic area were lower in the fracture group than in the no-fracture group. There was no significant difference in cervical, thoracic, lumbar or sacral HU between the 2 groups.

CONCLUSIONS

Our study reports that lower PET-derived non-metastatic bone deposition in the thoracic spine is correlated with incidence of fractures in patients with prostate cancer. CT-derived HU, a correlate of DXA-derived BMD, was not predictive of fracture risk. 18F-NaF PET/CT may provide important insight into bone quality and fracture risk.

Imaging techniques to study diabetic bone disease

PURPOSE OF REVIEW

This review article presents the most recent research on bone fragility in individuals with diabetes from a medical imaging perspective.

RECENT FINDINGS

The widespread availability of dual-energy X-ray absorptiometry (DXA) and trabecular bone score (TBS) software has led to recent assessments of bone fragility with this texture parameter in several studies of type 2 diabetes mellitus (T2D), but in few of type 1 diabetes mellitus (T1D). Although most studies show a trend of reduced TBS values in T2D independent of areal bone mineral density (aBMD) of the lumbar spine, some studies also show the limitations of TBS in both T2D and T1D. Given the limitations of DXA to assess bone strength and investigate the etiology of bone fragility in diabetes, more investigators are incorporating three-dimensional (3D) medical imaging techniques in their studies. Recent use of 3D medical imaging to assess bone fragility in the setting of diabetes has been mostly limited to a few cross-sectional studies predominantly incorporating high-resolution peripheral quantitative computed tomography (HR-pQCT). Although HR-pQCT studies indicate higher tibial cortical porosity in subjects with T2D, results are inconsistent in T1D due to differences in study designs, sample sizes, and subject characteristics, among other factors. With respect to central CT, recent studies support a previous finding in the literature indicating femoral neck geometrical impairments in subjects with T2D and provide encouraging results for the incorporation of finite element analysis (FEA) to assess bone strength in studies of T2D. In the recent literature, there are no studies assessing bone fragility in T1D with QCT, and only two studies used pQCT reporting tibial and radial impairments in young women and children with T1D, respectively. Magnetic resonance imaging (MRI) has not been recently used in diabetic studies of bone fragility.

SUMMARY

As bone fragility in diabetes is not explained by DXA-derived aBMD and given the limitations of cross-sectional studies, it is imperative to use 3D imaging techniques for longitudinal assessments of the density, quality, and microenvironment of bone to improve our understanding of the effects of diabetes on bone and reduce the risk of fracture in this large and vulnerable population of subjects with diabetes.

Two Cannulated Screws Provide Sufficient Biomechanical Strength for Prophylactic Fixation in Adult Patients With an Aggressive Benign Femoral Neck Lesion

Background: Two cannulated screws were proposed for prophylactic fixation in adult patients with an aggressive benign femoral neck lesion in recent literature. However, the biomechanical properties of this intervention have not yet been investigated.

Methods: After the evaluation of the heterogeneity of bone mineral density and geometry via quantitative computed tomography, 24 embalmed adult human cadaver femurs were randomized into the control, inferior half of the anterior cortical (25%) bone defect, entire anterior cortical (50%) bone defect, and the 50% bone defect and two cannulated screw group. Biomechanical analysis was conducted to compare the stiffness and failure load among the four groups when mimicking a one-legged stance. A CT-based finite element analysis (FEA) was performed to mimic the cortical and cancellous bone defect and the implantation of two cannulated screws of the four groups. Measurements of the maximal displacement and von Mises stress were conducted with the longitudinal load force and boundary conditions being established for a one-leg-standing status.

Results: We noted a significant improvement in the failure load after the insertion of two 6.5 mm cannulated screws in femurs with 50% bone defect (+95%, p = 0.048), and no significant difference was found between the screw group and the intact femur. Similar trends were also found in the measurements of stiffness (+23%, p > 0.05) via biomechanical testing and the von Mises stresses (−71%, p = 0.043) by FEA when comparing the screw group and the 50% bone defect group.

Conclusion: Our findings suggest that two cannulated screws provided sufficient biomechanical strength for prophylactic fixation in adult patients with an aggressive benign femoral neck lesion even when the entire anterior cortical bone is involved.

The effect of variations in CT scan protocol on femoral finite element failure load assessment using phantomless calibration

Recently, it was shown that fracture risk assessment in patients with femoral bone metastases using Finite Element (FE) modeling can be performed using a calibration phantom or air-fat-muscle calibration and that non-patient-specific calibration was less favorable. The purpose of this study was to investigate if phantomless calibration can be used instead of phantom calibration when different CT protocols are used. Differences in effect of CT protocols on Hounsfield units (HU), calculated bone mineral density (BMD) and FE failure loads between phantom and two methods of phantomless calibrations were studied. Five human cadaver lower limbs were scanned atop a calibration phantom according to a standard scanning protocol and seven additional commonly deviating protocols including current, peak kilovoltage (kVp), slice thickness, rotation time, field of view, reconstruction kernel, and reconstruction algorithm. The HUs of the scans were calibrated to BMD (in mg/cm³) using the calibration phantom as well as using air-fat-muscle and non-patient-specific calibration, resulting in three models for each scan. FE models were created, and failure loads were calculated by simulating an axial load on the femur. HU, calculated BMD and failure load of all protocols were compared between the three calibration methods. The different protocols showed little variation in HU, BMD and failure load. However, compared to phantom calibration, changing the kVp resulted in a relatively large decrease of approximately 10% in mean HU and BMD of the trabecular and cortical region of interest (ROI), resulting in a 13.8% and 13.4% lower failure load when air-fat-muscle and non-patient-specific calibrations were used, respectively. In conclusion, while we observed significant correlations between air-fat-muscle calibration and phantom calibration as well as between non-patient-specific calibration and phantom calibration, our sample size was too small to prove that either of these calibration approaches was superior. Further studies are necessary to test whether air-fat-muscle or non-patient-specific calibration could replace phantom calibration in case of different scanning protocols.

Fracture Risk Evaluation of Bone Metastases: A Burning Issue

Major progress has been achieved to treat cancer patients and survival has improved considerably, even for stage-IV bone metastatic patients. Locomotive health has become a crucial issue for patient autonomy and quality of life. The centerpiece of the reflection lies in the fracture risk evaluation of bone metastasis to guide physician decision regarding physical activity, antiresorptive agent prescription, and local intervention by radiotherapy, surgery, and interventional radiology. A key mandatory step, since bone metastases may be asymptomatic and disseminated throughout the skeleton, is to identify the bone metastasis location by cartography, especially within weight-bearing bones. For every location, the fracture risk evaluation relies on qualitative approaches using imagery and scores such as Mirels and spinal instability neoplastic score (SINS). This approach, however, has important limitations and there is a need to develop new tools for bone metastatic and myeloma fracture risk evaluation. Personalized numerical simulation qCT-based imaging constitutes one of these emerging tools to assess bone tumoral strength and estimate the femoral and vertebral fracture risk. The next generation of numerical simulation and artificial intelligence will take into account multiple loadings to integrate movement and obtain conditions even closer to real-life, in order to guide patient rehabilitation and activity within a personalized-medicine approach.

Patient-specific finite element computer models improve fracture risk assessments in cancer patients with femoral bone metastases compared to clinical guidelines

PURPOSE

To determine whether patient-specific finite element (FE) computer models are better at assessing fracture risk for femoral bone metastases compared to clinical assessments based on axial cortical involvement on conventional radiographs, as described in current clinical guidelines.

METHODS

Forty-five patients with 50 femoral bone metastases, who were treated with palliative radiotherapy for pain, were included (64% single fraction (8Gy), 36% multiple fractions (5 or 6x4Gy)) and were followed for six months to determine whether they developed a pathological femoral fracture. All plain radiographs available within a two month period prior to radiotherapy were obtained. Patient-specific FE models were constructed based on the geometry and bone density obtained from the baseline quantitative CT scans used for radiotherapy planning. Femoral failure loads normalized for body weight (BW) were calculated. Patients with a failure load of 7.5 x BW or lower were identified as having high fracture risk, whereas patients with a failure load higher than 7.5 x BW were classified as low fracture risk. Experienced assessors measured axial cortical involvement on conventional radiographs. Following clinical guidelines, patients with lesions larger than 30mm were identified as having a high fracture risk. FE predictions were compared to clinical assessments by means of diagnostic accuracy values (sensitivity, specificity and positive (PPV) and negative predictive values (NPV)).

RESULTS

Seven femurs (14%) fractured during follow-up. Median time to fracture was 8 weeks. FE models were better at assessing fracture risk in comparison to axial cortical involvement (sensitivity 100% vs. 86%, specificity 74% vs. 42%, PPV 39% vs. 19%, and NPV 100% vs. 95%, for the FE computer model vs. axial cortical involvement, respectively).

CONCLUSIONS

Patient-specific FE computer models improve fracture risk assessments of femoral bone metastases in advanced cancer patients compared to clinical assessments based on axial cortical involvement, which is currently used in clinical guidelines.

Patient-Specific Bone Multiscale Modelling, Fracture Simulation and Risk Analysis-A Survey

This paper provides a starting point for researchers and practitioners from biology, medicine, physics and engineering who can benefit from an up-to-date literature survey on patient-specific bone fracture modelling, simulation and risk analysis. This survey hints at a framework for devising realistic patient-specific bone fracture simulations. This paper has 18 sections: Section 1 presents the main interested parties; Section 2 explains the organzation of the text; Section 3 motivates further work on patient-specific bone fracture simulation; Section 4 motivates this survey; Section 5 concerns the collection of bibliographical references; Section 6 motivates the physico-mathematical approach to bone fracture; Section 7 presents the modelling of bone as a continuum; Section 8 categorizes the surveyed literature into a continuum mechanics framework; Section 9 concerns the computational modelling of bone geometry; Section 10 concerns the estimation of bone mechanical properties; Section 11 concerns the selection of boundary conditions representative of bone trauma; Section 12 concerns bone fracture simulation; Section 13 presents the multiscale structure of bone; Section 14 concerns the multiscale mathematical modelling of bone; Section 15 concerns the experimental validation of bone fracture simulations; Section 16 concerns bone fracture risk assessment. Lastly, glossaries for symbols, acronyms, and physico-mathematical terms are provided.