Dennis E. Anderson

Purpose: This study investigated the effect of bone metastasis on the biomechanical environment of human vertebrae in patients with metastatic spine disease through the metric of load‐to‐strength ratio (LSR). Specifically, we compared the patients' LSRs to age and sex‐similar noncancer controls from the Framingham Heart Study. Methods: Derived from clinical CT data of 135 metastatic spine disease patients planned for radiotherapy and 246 normative controls from the Framingham Heart Study, individualized spinal musculoskeletal models and vertebral strength estimates were used to compute level‐specific LSR under natural standing and three weight‐holding conditions (standing + weight, flexion + weight, and lateral bending + weight). Results: Adjusted for age, BMI, and spinal region, osteosclerotic and mixed lesion vertebrae had higher strength than osteolytic and control vertebrae. The musculoskeletal models suggested breast, prostate, and male lung cancer patients had higher compressive vertebral loading, and female lung cancer patients had lower compressive vertebral loading than controls. Male patients had higher standardized LSRs in natural standing, while female patients had lower LSRs for all activities than controls. Independent of sex, vertebrae with osteosclerotic and mixed bone metastasis had lower LSRs than controls, while, for osteolytic bone lesions, males had higher and females lower LSRs than controls. Vertebrae with no observed lesion on CT had higher LSRs than controls in males and lower LSRs in females. Discussion: Our findings highlighted that primary cancer and lesion type differentially affected task‐specific vertebral loading and strength, thus modifying the vertebral LSRs. Sex‐mediated differences in LSRs between FHS controls and vertebrae with no observed metastatic lesions suggest that considering the latter as “normal” should be taken with care. Our initial assessment supports further examination of whether vertebral LSR measurements are associated with vertebral risk and, if so, what threshold values indicate risk. Level of Evidence: 3.

Background: Adult spinal deformity (ASD) is an increasingly prevalent disorder in the aging population. Surgical intervention is a common and generally effective treatment for severe cases. However, it is associated with relatively high rates of complications, one of the most common, and devastating of which is proximal junctional failure (PJF). PJF is characterized by symptomatic mechanical failure at the junction of the spinal fusion construct and the adjacent proximal mobile spinal segments, leading to a kyphotic deformity. Current Limitations: The etiology of PJF remains a topic of ongoing investigation, with uncertainty surrounding the specific factors that predispose individual patients to this complication. Current predictive models primarily rely on radiographic parameters on standing X‐rays to assess PJF risk, but their clinical utility remains limited. We contend that these models universally fail to adequately account for the role of paraspinal muscle function and dysfunction, iatrogenic surgical muscle injury, bone quality, integrity of the discoligamentous elements, and spinal kinetics. Proposed Approach: Musculoskeletal modeling offers a powerful tool to enhance our understanding of human body kinetics and kinematics, including the complex biomechanical interactions in the spine. By integrating the biomechanical characteristics of bone and soft tissue into surgical treatment planning, we contend that subject‐specific musculoskeletal modeling will improve PJF predictability, enable the explanation and interpretation of PJF, and ultimately optimize outcomes for patients undergoing surgery for ASD. Conclusion: Subject‐specific musculoskeletal modeling represents a critical opportunity to address the limitations of existing predictive systems and advance the field of ASD management.

Background: Spine kinematics assessment is crucial for understanding intervertebral joint motion, particularly in conditions like spinal deformity, which alters and reduces spinal motion. Estimating spine kinematics in vivo usually relies on kinematic constraints to reduce the degrees of freedom in musculoskeletal models, but they lack standardization and fail to generalize across populations. This study proposes a novel method utilizing coordinate optimization instead of kinematic constraints, aiming to improve the generalizability and accuracy of spine kinematics estimation across different populations and marker protocols. Methods: This study used two retrospective datasets: 13 subjects with spinal deformities and 11 healthy individuals. Spine kinematics were estimated by minimizing errors between simulated and experimental marker positions and penalizing large intervertebral joint angles. 3D orientation and position errors against image‐based ground truth vertebral orientations and positions and experimental marker positions were calculated and compared for eight different weight settings. The accuracy was further assessed using standard error of measurements (SEM) compared to kinematic constraint methods. Results: The best‐performing optimization settings resulted in average vertebral orientation errors of 5.1°, 3.2°, and 3.2° for axial rotation, lateral bending, and flexion‐extension, respectively, and 3D position errors of 7.7 mm. These values reflect the average of vertebra‐specific errors within each subject, further averaged across all subjects in the deformity dataset. Similarly, in the healthy dataset, average 3D marker errors remained below 1 cm, and SEM values remained below 1.3°. Conclusions: The coordinate optimization method showed robust performance, achieving high accuracy in vertebral orientation and position (deformity) and marker tracking (healthy). This method consistently matched or surpassed state‐of‐the‐art kinematic constraints methods while introducing generalizability across different populations and marker protocols.

Background: Subject‐specific musculoskeletal models may be used to estimate spine loads that cannot be measured in vivo. Model generation methods may use detailed measurements extracted from medical imaging, but it may be possible to create accurate models without these measurements. We aimed to determine which physiological and anthropometric factors are associated with spine loading and should be accounted for in model creation.Methods: We created models of 440 subjects from the Framingham Heart Study Multi‐detector CT Study, extracting muscle morphology and spine profile information from CT scans of the trunk. Five lifting activities were simulated, and compressive and shear loading estimates were produced. We performed principal component analysis on the loading data from three locations in the spine, as well as univariate correlations between predictor variables and each principal component (PC). We identified multivariate predictive regression models for each PC and individual loading estimate.Results: A single PC explained 90% of the variability in compressive loading, while four PCs were identified that explained 10%–37% individually, 86% in total, of the variability in shear loading. Univariate analysis showed that body weight, BMI, lean mass, and waist circumference were most associated with the compression PC and first shear PC. Multivariate regression modeling showed predictor variables predicted 94% of the variability in the compression PC, but only 54% in the first shear PC, with body weight having the highest contribution. Additional shear PCs were less predictable. Level‐ and activity‐specific compressive loading was predicted using a limited set of physiological and anthropometric factors.Conclusions: This work identifies easily measured characteristics, particularly weight and height, along with sex, associated with subject‐specific loading estimates. It suggests that compressive loading, or models to evaluate compressive loading, may be based on a limited set of anthropometric attributes. Shear loading appears more complex and may require additional information not captured in the set of factors we examined.

Purpose Given the high prevalence of vertebral fractures post-radiotherapy in patients with metastatic spine disease, accurate and rapid muscle segmentation could support efforts to quantify muscular changes due to disease or treatment and enable biomechanical modeling for assessments of vertebral loading to improve personalized evaluation of vertebral fracture risk. This study presents a deep-learning approach for segmenting the complete volume of the trunk muscles from clinical CT images trained using sparsely annotated data. Materials and Methods we extracted 2,009 axial CT images at the midpoint of each vertebral level (T4 to L4) from clinical CT of 148 cancer patients. The key extensor and flexor muscles (up to 8 muscles per side) were manually contoured and labeled per image in the thoracic and lumbar regions. We first trained a 2D nnU-Net deep-learning model on these labels to segment key extensor and flexor muscles. Using these sparse annotations per spine, we trained the model to segment each muscle’s entire 3D volume Results The proposed method achieved comparable performance to manual segmentations, as assessed by expert radiologists, with a mean Dice score above 0.769. Significantly, the model drastically reduced segmentation time, from 4.3-6.5 hours for manual segmentation of 14 single axial CT images to approximately 1 minute for segmenting the complete thoracic-abdominal 3D volume. Conclusion The approach demonstrates high potential for automating 3D muscle segmentation, significantly reducing the manual intervention required for generating musculoskeletal models, and could be instrumental in enhancing clinical decision-making and patient care in radiation oncology. Summary A deep learning 2D nnU-Net model, trained on a sparse set of 2D muscle annotations, successfully segmented the entire volume of 20 thoracolumbar muscles from cancer patients’ clinical CT data. The model showed a remarkable increase in segmentation efficacy and generalizability, achieving comparable performance to manual segmentations in delineating each muscle anatomy. Key Points ▪ A deep learning model (2D nnU-Net), developed using a sparse set of single axial CT-slice at each mid-per vertebral level, containing manual image annotation of 20 thoracic and lumbar muscles, achieved comparable performance to manual segmentations, as assessed by expert radiologists, with a mean Dice score above 0.769. ▪ The model drastically reduced segmentation time, from 4.3-6.5 hours for manual segmentation of 14 single axial CT images to approximately 1 minute for segmenting the complete thoracic-abdominal 3D volume. ▪ Radiologist assessment based on a Likert scale (0-5) for clinical acceptability of the muscle anatomical segmentation showed strong model performance for a representative sample of clinical CT data (a (mean(SD) of 4.66 (0.73)) and external data (4.66 (0.73).

Pathologic vertebral fractures (PVF) are common and serious complications in patients with metastatic lesions affecting the spine. Accurate assessment of cancer patients’ PVF risk is an unmet clinical need. Load-to-strength ratios (LSRs) evaluated in vivo by estimating vertebral loading from biomechanical modeling and strength from computed tomography imaging (CT) have been associated with osteoporotic vertebral fractures in older adults. Here, for the first time, we investigate LSRs of thoracic and lumbar vertebrae of 135 spine metastases patients compared to LSRs of 246 healthy adults, comparable by age and sex, from the Framingham Heart Study under four loading tasks. Findings include: (1) Osteolytic vertebrae have higher LSRs than osteosclerotic and mixed vertebrae; (2). In patients’ vertebrae without CT observed metastases, LSRs were greater than healthy controls. (3) LSRs depend on the spinal region (Thoracic, Thoracolumbar, Lumbar). These findings suggest that LSRs may contribute to identifying patients at risk of incident PVF in metastatic spine disease patients. The lesion-mediated difference suggests that risk thresholds should be established based on spinal region, simulated task, and metastatic lesion type.

Costal cartilage plays an important functional role in the rib cage, but its mechanical properties have not been well characterized. The objective of this study is to characterize the properties of human costal cartilage and examine the effects of age, sex, rib level, and degree of calcification. We obtained cadaveric costal cartilage samples of ribs 3-6 with intact perichondrium from 24 donors (12 females and 12 males) evenly distributed by age (range 47-94 yr). Peripheral QCT scans were used to quantify geometric properties (area moments) and tissue calcification (as volume, length, and classified as central, peripheral, and mixed). Four-point bending tests were performed on each sample, and bending stiffness and modulus outcomes were evaluated by fitting data from mechanical testing with non-linear pseudo-elastic models (composed of linear and cubic components, separated into loading and unloading regimes). Effects of sex, age, rib level, and cartilage calcification on bending stiffness and modulus outcomes were assessed with mixed-effects regression models. Cartilage size (area moment) was larger in males than females and positively associated with age, while there was more calcification volume in cartilage of females than males. During loading, stiffness (linear and cubic) was larger in males, while modulus (linear and cubic) was larger in females. Linear stiffness and modulus were both negatively associated with age, positively associated with calcification, and varied between rib levels. Cubic (nonlinear) components of stiffness and modulus were positively associated with calcification and varied by rib, while modulus (but not stiffness) was negatively associated with age. During unloading, the linear stiffness and modulus values were much lower, though some similar associations were found. Overall, this study adds to our understanding of the behavior of costal cartilage as a nonlinear visco-elastic material, and the effects of sex, aging, and calcification on mechanical behavior.