Estimation of in vivo inter-vertebral loading during motion using fluoroscopic and magnetic resonance image informed finite element models

An artificial neural network for full-body posture prediction in dynamic lifting activities and effects of its prediction errors on model-estimated spinal loads

Musculoskeletal models have indispensable applications in occupational risk assessment/management and clinical treatment/rehabilitation programs. To estimate muscle forces and joint loads, these models require body posture during the activity under consideration. Posture is usually measured via video-camera motion tracking approaches that are time-consuming, costly, and/or limited to laboratories. Alternatively, posture-prediction tools based on artificial intelligence can be trained using measured postures of several subjects performing many activities. We aimed to use our previous posture-prediction artificial neural network (ANN), developed based on many measured static postures, to predict posture during dynamic lifting activities. Moreover, effects of the ANN posture-prediction errors on dynamic spinal loads were investigated using subject-specific musculoskeletal models. Seven individuals each performed twenty-five lifting tasks while their full-body three-dimensional posture was measured by a 10-camera Vicon system and also predicted by the ANN as functions of the hand-load positions during the lifting activities. The measured and predicted postures (i.e., coordinates of 39 skin markers) and their model-estimated L5-S1 loads were compared. The overall root-mean-squared-error (RMSE) and normalized (by the range of measured values) RMSE (nRMSE) between the predicted and measured postures for all markers/tasks/subjects was equal to 7.4 cm and 4.1 %, respectively (R2 = 0.98 and p < 0.05). The model-estimated L5-S1 loads based on the predicted and measured postures were generally in close agreements as also confirmed by the Bland-Altman analyses; the nRMSE for all subjects/tasks was < 10 % (R2 > 0.7 and p > 0.05). In conclusion, the easy-to-use ANN can accurately predict posture in dynamic lifting activities and its predicted posture can drive musculoskeletal models.

Multi-task artificial neural networks and their extrapolation capabilities to predict full-body 3D human posture during one- and two-handed load-handling activities

Machine-learning based human posture-prediction tools can potentially be robust alternatives to motion capture measurements. Existing posture-prediction approaches are confined to two-handed load-handling activities performed at heights below 120 cm from the floor and to predicting a limited number of body-joint coordinates/angles. Moreover, the extrapolating power of these tools beyond the range of the input dataset they were trained for (e.g., for underweight, overweight, or left-handed individuals) has not been investigated. In this study, we trained/validated/tested two posture-prediction (for full-body joint coordinates and angles) artificial neural networks (ANNs) using both 70%/15%/15% random-hold-out and leave-one-subject-out methods, based on a comprehensive kinematic dataset of forty-one full-body skin markers collected from twenty right-handed normal-weight (BMI = 18-26 kg/m2) subjects. Subjects performed 204 one- and two-handed unloaded activities at different vertical (0 to 180 cm from the floor) and horizontal (up to 60 cm lateral and/or anterior) destinations. Subsequently, the extrapolation capability of the trained/validated/tested ANNs was evaluated using data collected from fifteen additional subjects (unseen by the ANNs); three individuals in five groups: underweight, overweight, obese, left-handed individuals, and subjects with a hand-load. Results indicated that the ANNs predicted body joint coordinates and angles during various activities with errors of ∼ 25 mm and ∼ 10°, respectively; considerable improvements when compared to previous posture-prediction ANNs. Extrapolation errors of our ANNs generally remained within the error range of existing ANNs with obesity and being left-handed having, respectively, the most and least compromising effects on their accuracy. These easy-to-use ANNs appear, therefore, to be robust alternatives to common posture-measurement approaches.

Novel force-displacement control passive finite element models of the spine to simulate intact and pathological conditions; comparisons with traditional passive and detailed musculoskeletal models

Passive finite element (FE) models of the spine are commonly used to simulate intact and various pre- and postoperative pathological conditions. Being devoid of muscles, these traditional models are driven by simplistic loading scenarios, e.g., a constant moment and compressive follower load (FL) that do not properly mimic the complex in vivo loading condition under muscle exertions. We aim to develop novel passive FE models that are driven by more realistic yet simple loading scenarios, i.e., in vivo vertebral rotations and pathological-condition dependent FLs (estimated based on detailed musculoskeletal finite element (MS-FE) models). In these novel force-displacement control FE models, unlike the traditional passive FE models, FLs vary not only at different spine segments (T12-S1) but between intact, pre- and postoperative conditions. Intact, preoperative degenerated, and postoperative fused conditions at the L4-L5 segment for five static in vivo activities in upright and flexed postures were simulated by the traditional passive FE, novel force-displacement control FE, and gold-standard detailed MS-FE spine models. Our findings indicate that, when compared to the MS-FE models, the force-displacement control passive FE models could accurately predict the magnitude of disc compression force, intradiscal pressure, annulus maximal von Mises stress, and vector sum of all ligament forces at adjacent segments (L3-L4 and L5-S1) but failed to predict disc shear and facet joint forces. In this regard, the force-displacement control passive FE models were much more accurate than the traditional passive FE models. Clinical recommendations made based on traditional passive FE models should, therefore, be interpreted with caution.

A Reference Database of Standardised Continuous Lumbar Intervertebral Motion Analysis for Conducting Patient-Specific Comparisons

Lumbar instability has long been thought of as the failure of lumbar vertebrae to maintain their normal patterns of displacement. However, it is unknown what these patterns consist of. Research using quantitative fluoroscopy (QF) has shown that continuous lumbar intervertebral patterns of rotational displacement can be reliably measured during standing flexion and return motion using standardised protocols and can be used to assess patients with suspected lumbar spine motion disorders. However, normative values are needed to make individualised comparisons. One hundred and thirty-one healthy asymptomatic participants were recruited and performed guided flexion and return motion by following the rotating arm of an upright motion frame. Fluoroscopic image acquisition at 15fps was performed and individual intervertebral levels from L2-3 to L5-S1 were tracked and analysed during separate outward flexion and return phases. Results were presented as proportional intervertebral motion representing these phases using continuous means and 95%CIs, followed by verification of the differences between levels using Statistical Parametric Mapping (SPM). A secondary analysis of 8 control participants matched to 8 patients with chronic, non-specific low back pain (CNSLBP) was performed for comparison. One hundred and twenty-seven asymptomatic participants’ data were analysed. Their ages ranged from 18 to 70 years (mean 38.6) with mean body mass index 23.8 kg/m² 48.8% were female. Both the flexion and return phases for each level evidenced continuous change in mean proportional motion share, with narrow confidence intervals, highly significant differences and discrete motion paths between levels as confirmed by SPM. Patients in the secondary analysis evidenced significantly less L5-S1 motion than controls (p < 0.05). A reference database of spinal displacement patterns during lumbar (L2-S1) intersegmental flexion and return motion using a standardised motion protocol using fluoroscopy is presented. Spinal displacement patterns in asymptomatic individuals were found to be distinctive and consistent for each intervertebral level, and to continuously change during bending and return. This database may be used to allow continuous intervertebral kinematics to drive dynamic models of joint and muscular forces as well as reference values against which to make patient-specific comparisons in suspected cases of lumbar spine motion disorders.

Computational modelling of the scoliotic spine: A literature review

Scoliosis is a deformity of the spine that in severe cases requires surgical treatment. There is still disagreement among clinicians as to what the aim of such treatment is as well as the optimal surgical technique. Numerical models can aid clinical decision-making by estimating the outcome of a given surgical intervention. This paper provided some background information on the modelling of the healthy spine and a review of the literature on scoliotic spine models, their validation, and their application. An overview of the methods and techniques used to construct scoliotic finite element and multibody models was given as well as the boundary conditions used in the simulations. The current limitations of the models were discussed as well as how such limitations are addressed in non-scoliotic spine models. Finally, future directions for the numerical modelling of scoliosis were addressed.

Distribution of Young's modulus at various sampling points in a human lumbar spine vertebral body

BACKGROUND CONTEXT

Mathematical modeling for creating computer spine models is one of the basic methods underlying many scientific publications. The accuracy of strength parameters of tissues introduced into such models translates directly into the reliability of obtained results. Experimental determination of Young's modulus (E) in various areas of spongy bone tissue seems to be crucial for creating a reliable spine model without excessive simplifications in the form of a single E value for the whole vertebral body.

PURPOSE

The aim of the study was to determine Young's modulus in different parts of the lumbar vertebral column for samples subjected to compression and bending.

STUDY DESIGN

Cylindrical spongy bone tissue samples were subjected to bending and compression strength tests.

METHODS

The study included 975 pathologically unchanged samples of spongy bone tissue harvested from the lumbar vertebrae of 15 male donors. The samples were subjected to compression or bending strength tests and then Young's modulus was determined for each sample depending on its location in the vertebral body. The samples were tested differently between given locations within one vertebra as well as between vertebrae.

RESULTS

Compressed specimens are characterized by highly significantly different Young's modulus values depending on the location in the vertebral body. Samples No. 7 and No. 9 in the anterior part of the vertebral body have highly significantly higher Young's modulus values compared to those in the posterior part of the vertebral body for all lumbar vertebrae. Samples subjected to bending showed significant differences (p<.05) between samples located closer to the vertebral canal (No.16, No.17) and samples located further away (No.14, No.15) with higher values for the samples located in the posterior part of the vertebral body.

CONCLUSIONS

Accommodating the anisotropic structure of spongy bone in computer models and the application of different Young's module values for areas within one vertebral body will allow one to obtain realistic results of computer simulations used.

CLINICAL SIGNIFICANCE

Determining the exact strength parameters of spongy bone tissue within one vertebra and changes in these properties in subsequent vertebrae will allow to create more accurate computer models of the lumbar spine and the whole spine. This, in turn, will translate into more reliable computer simulations used, among others, to determine the risk of fractures or osteoporotic changes, or simulation of the procedure of spinal fusion.

Development of a multiscale model of the human lumbar spine for investigation of tissue loads in people with and without a transtibial amputation during sit-to-stand

Quantification of lumbar spine load transfer is important for understanding low back pain, especially among persons with a lower limb amputation. Computational modeling provides a helpful solution for obtaining estimates of in vivo loads. A multiscale model was constructed by combining musculoskeletal and finite element (FE) models of the lumbar spine to determine tissue loading during daily activities. Three-dimensional kinematic and ground reaction force data were collected from participants with ([Formula: see text]) and without ([Formula: see text]) a unilateral transtibial amputation (TTA) during 5 sit-to-stand trials. We estimated tissue-level load transfer from the multiscale model by controlling the FE model with intervertebral kinematics and muscle forces predicted by the musculoskeletal model. Annulus fibrosis stress, intradiscal pressure (IDP), and facet contact forces were calculated using the FE model. Differences in whole-body kinematics, muscle forces, and tissue-level loads were found between participant groups. Notably, participants with TTA had greater axial rotation toward their intact limb ([Formula: see text]), greater abdominal muscle activity ([Formula: see text]), and greater overall tissue loading throughout sit-to-stand ([Formula: see text]) compared to able-bodied participants. Both normalized (to upright standing) and absolute estimates of L4-L5 IDP were close to in vivo values reported in the literature. The multiscale model can be used to estimate the distribution of loads within different lumbar spine tissue structures and can be adapted for use with different activities, populations, and spinal geometries.

An in vivo study exploring correlations between early-to-moderate disc degeneration and flexion mobility in the lumbar spine

PURPOSE

Early disc degeneration (DD) has been thought to be associated with loss of spine stability. However, before this can be understood in relation to back pain, it is necessary to know the relationship between DD and intervertebral motion in people without pain. This study aimed to find out if early-to-moderate DD is associated with intervertebral motion in people without back pain.

METHODS

Ten pain-free adults, aged 51-71, received recumbent and weight bearing MRI scans and quantitative fluoroscopy (QF) screenings during recumbent and upright lumbar flexion. Forty individual level and 10 composite (L2-S1) radiographic and MRI DD gradings were recorded and correlated with intervertebral flexion ROM, translation, laxity and motion sharing inequality and variability for both positions.

RESULTS

Kinematic values were similar to previous control studies. DD was evidenced up to moderate levels by both radiographic and MRI grading. Disc height loss correlated slightly, but negatively with flexion during weight bearing flexion (R =  - 0.356, p = 0.0.025). Composite MRI DD and T2 signal loss evidenced similar relationships (R =  - 0.305, R =  - 0.267) but did not reach statistical significance (p = 0.056, p = 0.096). No significant relationships between any other kinematic variables and DD were found.

CONCLUSION

This study found only small, indefinite associations between early-to-moderate DD and intervertebral motion in healthy controls. Motion sharing in the absence of pain was also not related to early DD, consistent with previous control studies. Further research is needed to investigate these relationships in patients.

Dynamic interactions between lumbar intervertebral motion segments during forward bending and return

Continuous dynamic multi-segmental studies of lumbar motion have added depth to our understanding of the biomechanics of back pain, but few have attempted to continuously measure the proportions of motion accepted by individual levels. This study attempted to compare the motion contributions of adjacent lumbar levels during an active weight bearing flexion and return protocol in chronic, non-specific low back pain (CNSLBP) patients and controls using quantitative fluoroscopy (QF). Eight CNSLBP patients received QF during guided standing lumbar flexion. Dynamic motion sharing of segments from L2 to S1 were calculated and analysed for interactions between levels. Eight asymptomatic controls were then matched to the 8 patients for age and sex and their motion sharing patterns compared. Share of intersegmental motion was found to be consistently highest at L2-L3 and L3-L4 and lowest at L5-S1 throughout the motion in both groups, with the exception of maximum flexion where L4-L5 received the greatest share. Change in motion sharing occurred throughout the flexion and return motion paths in both participant groups but tended to vary more at L4-L5 in patients (p < 0.05). In patients, L5-S1 provided less angular range (p < 0.05) and contributed less at maximum bend (p < 0.05), while L3-L4, on average over the bending sequence, provided a greater share of motion (p < 0.05). Intervertebral motion sharing inequality is therefore a normal feature during lumbar flexion. However, in patients, inequality was more pronounced, and variability of motion share at some levels increased. These effects may result from differences in muscular contraction or in the mechanical properties of the disc.

Coupled artificial neural networks to estimate 3D whole-body posture, lumbosacral moments, and spinal loads during load-handling activities

Biomechanical modeling approaches require body posture to evaluate the risk of spine injury during manual material handling. The procedure to measure body posture via motion-analysis techniques as well as the subsequent calculations of lumbosacral moments and spine loads by, respectively, inverse-dynamic and musculoskeletal models are complex and time-consuming. We aim to develop easy-to-use yet accurate artificial neural networks (ANNs) that predict 3D whole-body posture (ANN_posture), segmental orientations (ANN_angle), and lumbosacral moments (ANN_moment) based on our measurements during load-handling activities. Fifteen individuals each performed 135 load-handling activities by reaching (0 kg) or handling (5 and 10 kg) weights located at nine different horizontal and five vertical (0, 30, 60, 90, and 120 cm from the floor) locations. Whole-body posture was measured via a motion capture system and lumbosacral moments were calculated via a 3D top-down eight link-segment inverse-dynamic model. ANN_posture, ANN_angle, and ANN_moment were trained (RMSEs = 6.7 cm, 29.8°, and 16.2 Nm, respectively) and their generalization capability was tested (RMSE = 7.0 cm and R² = 0.97, RMSE = 29.9° and R² = 0.85, and RMSE = 16.5 Nm and R² = 0.97, respectively). These ANNs were subsequently coupled to our previously-developed/validated ANN_load, which predicts spinal loads during 3D load-handling activities. The results showed outputs of the coupled ANNs for L4-L5 intradiscal pressure (IDPs) during a number of activities were in agreement with measured IDPs (RMSE = 0.37 MPa and R² = 0.89). Hence, coupled ANNs were found to be robust tools to evaluate posture, lumbosacral moments, spinal loads, and thus risk of injury during load-handling activities.

Can Biomechanics Research Lead to More Effective Treatment of Low Back Pain? A Point-Counterpoint Debate

SYNOPSIS

Although biomechanics plays a role in the development and perhaps the persistent or recurrent nature of low back pain (LBP), whether biomechanics alone can provide the basis for intervention is debated. Biomechanics, which refers to the mechanics of the body, including its neuromuscular control, has been studied extensively in LBP. But, can gains be made in understanding LBP by research focused on this component of biology in the multifactorial biopsychosocial problem of LBP? This commentary considers whether biomechanics research has the potential to advance treatment of LBP, and how likely it is that this research will lead to better treatment strategies. A point-counterpoint format is taken to present both sides of the argument. First, the challenges faced by an approach that considers biomechanics in isolation are presented. Next, we describe 3 models that place substantial emphasis on biomechanical factors. Finally, reactions to each point are presented as a foundation for further research and clinical practice to progress understanding of the place for biomechanics in guiding treatment of LBP. J Orthop Sports Phys Ther 2019;49(6):425-436. Epub 15 May 2019. doi:10.2519/jospt.2019.8825.

Subject-specific loads on the lumbar spine in detailed finite element models scaled geometrically and kinematic-driven by radiography images

Traditional load-control musculoskeletal and finite element (FE) models of the spine fail to accurately predict in vivo intervertebral joint loads due mainly to the simplifications and assumptions when estimating redundant trunk muscle forces. An alternative powerful protocol that bypasses the calculation of muscle forces is to drive the detailed FE models by image-based in vivo displacements. Development of subject-specific models, however, both involves the risk of extensive radiation exposures while imaging in supine and upright postures and is time consuming in terms of the reconstruction of the vertebrae, discs, ligaments, and facets geometries. This study therefore aimed to introduce a remedy for the development of subject-specific FE models by scaling the geometry of an existing detailed FE model of the T12-S1 lumbar spine. Five subject-specific scaled models were driven by their own radiography image-based displacements in order to predict joint loads, ligament forces, facet joint forces, and disc fiber strains during relaxed upright as well as moderate flexion and extension tasks. The predicted intradiscal pressures were found in adequate agreement with in vivo data for upright, flexion, and extension tasks. There were however large intersubject variations in the estimated joint loads and facet forces.

Variation in human vertebral body strength for vertebral body samples from different locations in segments L1-L5

BACKGROUND

The human spine, in particular the lumbar spine, is subject to significant compressive and bending stresses, which affect the structure of the bone tissue of the vertebrae. The more heterogeneous the structure of the spongy bone tissue, the less resistant the whole vertebral body. It is therefore necessary to establish variations in bone strength parameters within one particular vertebral body.

METHODS

The research material comprised human L1-L5 lumbar vertebrae sampled from 15 donors aged 29-35. A total of 975 samples prepared from the collected material were subjected to compressive and bending strength tests. The samples for the tests were collected from carefully selected locations in order to discover the strength properties of various parts of the vertebral body.

FINDINGS

In the case of sample 2 (located in the posterior part of the vertebra, at mid-height) the stress values were the lowest and there were statistically significant differences compared to other samples. Moreover the value of compressive force in this case was lower for vertebrae with higher numbers. Top and bottom samples demonstrated statistically significant higher mean values of destructive stress. In terms of the bending strength test, the mean value of destructive stress in all lumbar vertebrae for all samples increased for vertebrae with higher numbers.

INTERPRETATION

The spongy tissue in healthy vertebral bodies has a very heterogeneous structure. This may be due to the presence of the nutrient canal and the arc structure allowing more springy movement and improved transfer of loads by the vertebral body.

Aberrant intervertebral motion in patients with treatment-resistant nonspecific low back pain: a retrospective cohort study and control comparison

PURPOSE

Intervertebral kinematic assessments have been used to investigate mechanical causes when back pain is resistant to treatment, and recent studies have identified intervertebral motion markers that discriminate patients from controls. However, such patients are a heterogeneous group, some of whom have structural disruption, but the effects of this on intervertebral kinematics are unknown.

METHODS

Thirty-seven patients with treatment-resistant back pain referred for quantitative fluoroscopy were matched to an equal number of pain-free controls for age and sex. All received passive recumbent flexion assessments for intervertebral motion sharing inequality (MSI), variability (MSV), laxity and translation. Comparisons were made between patient subgroups, between patients and controls and against normative levels from a separate group of controls.

RESULTS

Eleven patients had had surgical or interventional procedures, and ten had spondylolisthesis or pars defects. Sixteen had no disruption. Patients had significantly higher median MSI values (0.30) than controls (0.27, p = 0.010), but not MSV (patients 0.08 vs controls 0.08, p = 0.791). Patients who received invasive procedures had higher median MSI values (0.37) than those with bony defects (0.30, p = 0.018) or no disruption (0.28, p = 0.0007). Laxity and translation above reference limits were not more prevalent in patients.

CONCLUSION

Patients with treatment-resistant nonspecific back pain have greater MSI values than controls, especially if the former have received spinal surgery. However, excessive laxity, translation and MSV are not more prevalent in these patients. Thus, MSI should be investigated as a pain mechanism and for its possible value as a prognostic factor and/or target for treatment in larger patient populations. These slides can be retrieved under Electronic Supplementary Material.