Siril Dukkipati

A rigid body musculoskeletal model of the lumbar spine along with intra-abdominal pressure (IAP) was presented in this research. Traditional spinal biomechanical models in literature often neglect the load-sharing effect of IAP on the spine and can be computationally intensive. These limitations limit them to be effectively used for computationally intensive applications like muscle recruitment simulations. As such, there remains a need for high-fidelity validated musculoskeletal spine models, hence developed herein. The skeletal components were adapted from 3D MRI scans and the intervertebral discs were modeled as 3 degrees-of-freedom (DOF) gimbal joints with a custom joint torque feedback for their control. Supraspinous, interspinous, and intertransverse ligaments were modeled as tension-only springs. Further, two methods of modeling IAP were discussed and implemented. The resultant model consisted a total of 15 DOFs, constrained by 279 independent force-generating elements. The lumbar segmental stiffness profiles in all three directions were found to be within one standard deviation of the literature data. A linear increase in the extensor torque about L3 was observed with an increase in IAP emphasizing the IAP’s load-sharing effect during flexion. The 6 sec compilation and 1.4 sec run times suggest that this model is suitable for muscle recruitment optimization problems. This MATLAB based simulation platform offers a quick and intuitive tool to visualize physiological loading for the biomechanics community.

Spine biomechanical testing methods in the past few decades have not evolved beyond employing either cadaveric studies or finite element modeling techniques. However, both these approaches may have inherent cost and time limitations. Cadaveric studies are the present gold standard for spinal implant design and regulatory approval, but they introduce significant variability in measurements across patients, often requiring large sample sizes. Finite element modeling demands considerable expertise and can be computationally expensive when complex geometry and material nonlinearity are introduced. Validated analogue spine models could complement these traditional methods as a low-cost and high-fidelity alternative. A fully 3D printable L-S1 analogue spine model with ligaments is developed and validated in this research. Rotational stiffness of the model under pure bending loading in flexion-extension, Lateral Bending (LB) and Axial Rotation (AR) is evaluated and compared against historical ex vivo and in silico models. Additionally, the effect of interspinous, intertransverse ligaments and the Thoracolumbar Fascia (TLF) on spinal stiffness is evaluated by systematic construction of the model. In flexion, model Range of Motion (ROM) was 12.92 ± 0.11° (ex vivo: 16.58°, in silico: 12.96°) at 7.5Nm. In LB, average ROM was 13.67 ± 0.12° at 7.5 Nm (ex vivo: 15.21 ± 1.89°, in silico: 15.49 ± 0.23°). Similarly, in AR, average ROM was 17.69 ± 2.12° at 7.5Nm (ex vivo: 14.12 ± 0.31°, in silico: 15.91 ± 0.28°). The addition of interspinous and intertransverse ligaments increased both flexion and LB stiffnesses by approximately 5%. Addition of TLF showed increase in flexion and AR stiffnesses by 29% and 24%, respectively. This novel model can reproduce physiological ROMs with high repeatability and could be a useful open-source tool in spine biomechanics.

There exists a need for validated lumbar spine models in spine biomechanics research. Although cadaveric testing is the current gold standard for spinal implant development, it poses significant issues related to reliability, repeatability, and ethics. Analogue or synthetic models can act as cost saving alternatives to human tissue. This study presents a first reproducible 3D printable L1-S1 lumbar spine motion segment and validated it in flexion-extension (F-E), lateral bending (LB), and axial rotation (AR) in the range of 15° against in vitro data and in silico models. The model consisted of L1 to S1 vertebrae, corresponding intervertebral discs, intertransverse, interspinous, anterior and posterior longitudinal ligaments, and facet joints. Displacement controlled pure moments were applied in the range of 15° in all three bending modes and the resisting moment was recorded. Rotational stiffness was calculated by plotting the resisting moment against rotation. The model reached a maximum of 5.66Nm and 3.53Nm at 15° flexion and rotation, 3.84Nm and 3.93Nm at 15° right and left lateral bending, and 2.45Nm and 2.59Nm at 15° right and left axial rotations respectively. Scaling factors b=1.6 and g=3.0 were applied in LB and AR to calibrate the model response. Post scaling, the model response in all axes correlated well with in vitro and in silico literature data. An RMS error of 1.57°, 1.64°, 0.82° in F-E, LB and AR respectively was estimated in the model when compared to in vitro human response. The reproduceable 3D printed analogue model described herein can be a valid substitute for cadaveric lumbar motion segments within the explored context of use.