Traumatic Spinal Injury and Robotic Reconstruction

Introduction of Technology

Traumatic spinal injury (TSI) may involve injury to the osseous structures, discoligamentous components, spinal cord, nerve roots of the spinal column, or a combination thereof. TSI can lead to a wide range of medical conditions, from minimal nondisplaced fractures to mechanically unstable fractures with associated neurological compromise. The clinical sequelae of TSI, which can include pain and/or neurological deficit, may result in severe morbidity and mortality. Although abundant literature exists on spinal cord injury (SCI), a subset of TSI, studies have only recently investigated the epidemiology of TSI. A meta-analysis of 102 studies from 32 countries by Kumar et al. found the global incidence of TSI to be 10.5 cases per 100,000 persons. With the world population exceeding 7 billion people, this translates to over 750,000 new cases of TSI worldwide. Men are more commonly affected than are women, with an average age of injury of 40 years old. The leading causes of TSI worldwide are traffic collisions (40%) and falls (39%). Although multiple older studies have reported approximately 160,000 spine fractures in North America, Kumar et al. found an incidence of 5.1 cases of TSI per 100,000 persons for an estimated 18,220 cases (95% confidence interval: 5002–67,881) annually.

Fractures of the thoracolumbar spine are common. Anatomically, the upper thoracic spine is less mobile due to articulations with the ribs that attach to the sternum. The thoracolumbar junction, located at T10–L2, represents an area of transition from the rigid thoracic spine to the more dynamic lumbar spine. Injury commonly occurs in this region because the thoracic spine lacks the flexibility of the lumbar spine, so sudden and traumatic movements often stress this area at a higher load than it can handle. Katsuura et al. conducted a meta-analysis and found the rate of thoracolumbar fracture to be 6.9% in patients who have sustained blunt trauma. Of patients with thoracolumbar fractures, injury at the thoracolumbar junction can be as high as 41%. Trauma may concurrently occur in other locations, such as the extremities (19%), head (13%), cervical spine (11%), and abdomen (7%).

Fortunately, many fractures are stable and do not involve neurologic compromise; therefore, they can be treated conservatively with pain control, bracing, or both. When TSI causes neurologic deficit or instability of the spinal column, surgery may be needed for fixation, decompression of neural elements, and/or reduction of deformity. With improvements in modern trauma care, surgery can now allow for earlier mobilization, which is thought to be important in reducing morbidity and mortality. Specifically, early surgical intervention can be beneficial in:

•

Improving pulmonary function with decreased rates of pneumonia and decreased time on the ventilator,
•

Diminishing pain from instability,
•

Reducing the incidence and severity of sepsis and respiratory failure, and
•

Decreasing overall intensive care unit and hospital lengths of stay.

In the TSI population, however, surgical intervention does carry certain additional risks. Trauma patients can have inadequate resuscitation, propensity for increased hemorrhage, sensitivity to hypotension, overlooked or underestimated associated injuries, or some combination of these factors. Surgeons may be at an even greater disadvantage by operating in less than ideal conditions. These factors must be weighed by the surgeon when deciding on the need for and timing of surgical intervention.

Pedicle screw instrumentation was first described in the 1960s and 1970s and has since become the gold standard of spinal fusion. As pedicle screws and spinal fixation procedures have become more commonplace, efforts to optimize trajectory while minimizing invasiveness and malpositioning have been heavily invested in technological solutions. Beginning in 1995, computer assisted (i.e., image-guided) navigation was applied to spinal surgical procedures to improve screw placement accuracy and decrease the risk to the patient. Further developments over the past two decades have brought the use of intraoperative navigation tactics (such as the C-arm and O-arm) and robotics into everyday use for spinal fixation.

Robotic navigation combines the precision of computer based navigation with the steadiness of a machine, requiring only trajectory decisions with intraoperative confirmation and screw placement by the surgeon. Several surgical robots to improve pedicle screw placement now exist, including ExcelsiusGPS (Globus Medical, Audubon, PA), Mazor (Mazor Robotics, Caesarea, Israel), ROSA (Zimmer Biomet, Warsaw, IN), TiRobot (TINAVI Medical Technology, Beijing, China), and more. With increasing use of these surgical robots, patient outcomes are now being analyzed more thoroughly. Most patients described in the current robotic literature are those with degenerative spinal disease; indeed, very few studies have included TSI patients.

Han et al. conducted a prospective study of 234 patients with degenerative or traumatic thoracolumbar spinal injury who underwent pedicle screw placement with either the TiRobot system or conventional fluoroscopy. The patients were randomly sorted into the robotic-guided ( n = 115) or the fluoroscopy-assisted ( n = 119) surgical groups and the safety and accuracy of pedicle screw placement in each group was assessed and compared. 76/234 (32%) of participants had a trauma-based pathology. Of the 76 TSI patients, 41 (54%) were assigned to the robotic-guided group and 35 (46%) were assigned to the fluoroscopy-guided group. Screw placement was graded on the Gertzbein-Robbins scale with grades A (i.e., screw completely within the pedicle) and B (i.e., pedicle cortical breach < 2 mm) given to clinically acceptable screw placements. Of the robotic-guided procedures, 507/532 (95.3%) screws were considered grade A and 18/532 (3.4%) grade B, whereas in the fluoroscopy-guided surgeries, 503/584 (86.1%) screws were considered grade A and 43/584 (7.4%), grade B. The proportion of group A and B screws in the robotic-guided group was greater than that in the fluoroscopy-assisted group ( p < .01). The robot-guided group required no revision surgery (two were required in the fluoroscopy-guided group), had significantly less blood loss, and decreased surgeon radiation exposure compared with the fluoroscopy-guided group. Although subgroup analyses between the degenerative versus trauma groups were not carried out, this study demonstrates the successful use of robotic guidance in patients with trauma.

Le et al., from the same research group, performed a similar study with the TiRobot to assess the safety and accuracy of cortical bone trajectory screw placement in the lumbar spine with robot guidance compared to fluoroscopy guidance. Of 58 patients included, only 4 were TSI patients, and all were sorted into the fluoroscopy-guided group. Another study by Kam et al. examined the learning curve of robotic assisted screw placement in 80 patients, in which only 4 patients underwent surgery for a traumatic etiology. Neither study specifically analyzed the TSI patients.

A true paucity of literature exists pertaining to robotics and TSI; however, as the use of robotics increases in spine surgery, robotic reconstruction may become more valuable in minimally invasive procedures for patients who have sustained a traumatic injury. One of the main advantages of robotic assisted spine surgery is the potential to minimize the physiologic burden associated with open surgery. Prior research has shown that minimally invasive spine surgery lessens the extent of muscle and soft tissue dissection, resulting in decreased blood loss, length of stay, and complication rates, all while producing equivalent overall postoperative outcomes. Robotic assistance allows the surgeon to plan the screw trajectories and the place pedicle screws without the intensive dissection needed in open surgery. In trauma patients already prone to hemorrhage and who may be hemodynamically complicated due to other injuries, minimizing blood loss and invasiveness of surgery may help minimize morbidity. Additionally, robot assistance can be useful in patients with challenging anatomy (e.g., small pedicles) or abnormal anatomy by utilizing preoperative or real-time intraoperative computed tomography (CT) or fluoroscopy.

Surgical Treatment for Spinal Pathology

Clinical Case #1

History/Clinical Presentation

A 73-year-old male taxicab driver with a past medical history of coronary artery disease, diabetes, hypertension, horseshoe kidney, obstructive sleep apnea, and gout presented to the emergency department (ED) after being dragged underneath his taxi for 40 feet. Upon presentation to the ED, the patient described midline back pain and abdominal pain, but denied radicular pain in his arms or legs, numbness or tingling, saddle anesthesia, and bowel/bladder dysfunction. He was a former smoker and was on aspirin 81 mg daily.

The patient was hemodynamically stable with a temperature of 36.2ºC, heart rate of 74 bpm, blood pressure of 144/52, and SpO ₂ of 97% on room air. The patient was morbidly obese with a body mass index of 38.6 kg/m ² . He was awake, alert, and oriented. He had baseline weakness 4/5 strength in his left deltoid from a prior injury of his left shoulder but otherwise had full strength in all muscle groups. The examination was negative for the Hoffman sign and clonus.

CT of the spine showed an unstable burst fracture of the T8 vertebral body, with up to 6 mm of anteropulsion and approximately 25% height loss of the anterior superior endplate. The scan also demonstrated multilevel diffuse bridging ossification of the anterior vertebral bodies from T3 to T11, consistent with diffuse idiopathic skeletal hyperostosis. His labs were notable for platelets of 176, 000, creatinine of 0.98 mg/dL, INR 1.1, and aPTT 23.2.

After discussion with the attending neurosurgeon, the patient was admitted to neurosurgery floor care, with regular neurological exam checks and strict spinal precautions at all times. The patient was unable to undergo magnetic resonance imaging due to metal in his left shoulder. Given his age and medical comorbidities, medicine was consulted for preoperative risk assessment and the patient was medically optimized for surgery.

Preoperative Imaging

Unstable burst fracture of the T8 vertebral body with up to 6 mm of anteropulsion and approximately 25% height loss of the anterior superior endplate ( Fig. 11.1 ).

Surgical Plan

The patient underwent T6–T10 percutaneous, minimally invasive robotic assisted fusion. Following the induction of general anesthesia, baseline intraoperative monitoring with somatosensory-evoked potentials was obtained. The patient was carefully positioned prone on the surgical table with intraoperative neuromonitoring performed throughout the entirety of the procedure. The patient was prepared and draped in the usual sterile technique. An intraoperative radiograph was used to identify T11 and a midline incision was made over the T11 spinous process. An image-guided spinal tracker and a surveillance marker were placed on the T11 spinous process. Intraoperative anteroposterior and lateral radiographs were obtained and transferred to the ExcelsiusGPS planning station. His preoperative CT was already evaluated and the screws were planned using the robot software ( Fig. 11.2A and B ). Once the intraoperative fluoroscopic images were merged with the preoperative CT, real-time image guidance and robotics could be used.