Modeling of Intraoperative Patient Positioning to Facilitate and Improve Corrective Spine Surgeries

Severe spinal deformities requiring orthopedic surgeries such as scoliosis, spondylolisthesis or spinal trauma vary in etiology, but generally for their treatment patients need to be positioned on a specialized operating table to facilitate the various stages of surgery. Patients are secured on this operating table via a set of positioning components such as cushions, pads, and straps to ensure spine stability, provide surgical and radiological imaging access, minimize bleeding, stabilize and preserve vital structures (e.g., organs, nerves), provide proper ventilation and facilitate instrumentation procedures. The simple act of positioning the patient has been shown to have an influence on spinal correction and offers a demonstrated and documented potential to leverage the patient's body kinematics to segmentally adjust the shape of the spine prior to any surgical correction. During surgery, corrective maneuvers are performed by the clinician which impose important loads on the corrective instrumentation, but the potential of targeted position induced spine manipulations on reducing these intraoperative loads, benefiting both the clinician and patient, has never been documented and must be further researched. This project looked to answer the following question: when compared to the current practice, can position induced spine manipulations further improve and facilitate the correction of spinal deformities, as well as reduce the forces required to perform corrective maneuvers? It was hypothesized that: 1) a biomechanical model of the human spine calibrated with preoperative clinical data can simulate the changes in spinal curvatures resulting from being prone on a typical 4-post adjustable surgical frame with sagittal adjustment of the pelvis and torso, to within 5 degrees; 2) the previously validated biomechanical model of the human spine calibrated with preoperative clinical data can simulate the changes in spinal curvatures and the forces exerted on the vertebrae resulting from posterior instrumentation with rods and pedicle screws to within 5 degrees of accuracy for curvatures and within the measured range of comparable studies for forces; 3) perioperative patient positioning with pelvic and thoracic adjustments on a 4-post operating frame can reduce interface loads between the vertebral bone and the pedicle screws by up to 25% during instrumentation procedures. From there three specific objectives guided the strategy to answer this question and test these hypotheses: 1) develop a biomechanical finite element model (FEM) of the human spine capable of simulating the intraoperative prone position and posterior spinal instrumentation surgery; 2) verify and validate the model with existing clinical cases and quantify the effects of model uncertainties; 3) characterize the influence of patient positioning on the correction and forces resulting from instrumentation surgery, which was the principal objective of this project. To achieve objective 1, a personalized finite-element model of the human spine which included the thoracic and lumbar vertebrae, intervertebral discs, ribs, sternum, costal cartilages, abdominal soft tissues, and pelvis was adapted to include a surgical frame and spine deformity correction instrumentation. The correction instrumentation consisted in contoured rods and pedicle screws. Model simulation procedures with associated boundary conditions were developed to simulate the patient prone position and the instrumentation surgery via a translation technique. The model allowed to measure the resulting spine curvatures and axial forces at the bone-screw interface. To achieve objective 2, a verification, validation and uncertainty quantification plan was developed and followed. Five patients with adolescent idiopathic scoliosis from the Sainte-Justine University Hospital patient database were used to compare the simulation results with intraoperative clinical data measurements. Published in vivo and in silico studies were also used to compare results. Model uncertainty was quantified via a series of parameter sensitivity studies with a varied set of positioning, patient-related and instrumentation parameters. To achieve objective 3, the model was used to measure the influence of raising the surgical frame's thoracic cushions on the intraoperative post-instrumented pedicle screws loads. Our simulations yielded the following results: • The model allowed to simulate the changes in spinal curvatures resulting from the intraoperative prone position within clinical significance (≤ 5⁰) on average. • The model allowed to simulate a 50-mm vertical displacement of the thoracic cushions with comparable results as previous experiments with healthy patients. The thoracic cushions vertical displacement had a significant influence on thoracic kyphosis with an average change of 9⁰ corresponding to a 67% increase in kyphosis. • The model allowed to simulate the changes in spinal curvatures and forces exerted on the vertebrae resulting from posterior instrumentation with rods and pedicle screws within clinical significance (≤ 5⁰) on average for curvatures and within the measured range of comparable in silico studies for screw pull-out forces. • The studied standard 4-post surgical frame favored spine mobility in the sagittal plane and restricted mobility in the coronal plane. Pelvic rotation was effective at manipulating the sagittal curves with an average reduction of 32⁰ (at maximum flexion of lower limbs) to an increase of 14⁰ (at maximum extension) in lordosis. Longitudinal positioning of the thoracic cushions had a significant influence on thoracic kyphosis with an average 6⁰ reduction for a 50-mm cephalad displacement to a 4⁰ increase for a 50-mm caudal displacement. The combination of pelvic rotation in the sagittal plane and thoracic cushions placement was shown as pertinent for manipulating the sagittal curves prior to instrumentation, in accord with published studies. • Small variations (3 mm) in screw-to-rod connection modeling assumptions had a significant impact on screw pull-out forces average (up to 30%) and maximum (up to 69%) values. This finding confirmed how millimeter level adjustments in the instrumentation play a significant role in measured load levels. • The instrumented simulations using the model revealed a varying effect on intraoperative post-instrumented screw pull-out forces resulting from an increase in thoracic kyphosis using patient positioning. Results ranged from an average reduction of 23 N or 28% to an increase of 13 N or 17% in average forces and a reduction of 80 N or 37% to an increase of 52 N or 27% in maximum forces. It was observed that the sagittal correction objective imposed by the contoured rods shape may work in combination with or against the position induced correction. From this finding it was proposed to further investigate how the correction strategy could integrate the position induced correction to deduce contoured rod shapes that would minimize intraoperative forces. It was noted that further model developments would be required to also measure screw loads at intermediate stages of the instrumentation surgery which would involve a refinement of the screw-to-rod connection modeling.The proposed numerical model is the first to combine intraoperative position-induced spine manipulations with the capability to simulate instrumentation surgery and measure resulting implant forces. Our FEM has a great potential in the rapid development and testing of novel positioning modalities which aim to advantageously manipulate the spine to further facilitate and improve the correction of spinal deformities.