Electromyography stimulation compared with intraoperative O-arm imaging for evaluating pedicle screw breaches in lumbar spine surgery: a prospective analysis of 1006 screws in 164 patients

Abstract

Background contextLumbar pedicle screw placement can be technically challenging. Malpositioned screws occur in up to 15% of patients and could result in radiculopathy or instrumentation failure. PurposeTo compare intraoperative electromyography (EMG) and image guidance using an O-arm for identifying pedicle breach during elective lumbar fusion. Study designProspective observational study. Patient sampleAll adult patients undergoing elective lumbar spinal fusion operations for degenerative spine disorders (including adjacent segment degeneration, degenerative scoliosis, and symptomatic spondylosis and spondylolisthesis) at a single institution from July 1, 2014, to December 1, 2015, were prospectively tracked. Outcome measuresPedicle breach. MethodsPedicle screws from L2–S1 were placed using C-arm assisted freehand technique. All screws were stimulated with EMG and evaluated using the O-arm intraoperative imaging system. Electromyography data were compared with intraoperative images to assess the accuracy of identifying pedicle breaches. No funding was received for this work. ResultsOne thousand six lumbar pedicles screws were placed from L2 to S1 in 164 consecutive cases. The mean patient age was 59.2 years. Thirty-five breaches (15 lateral and 20 medial) were visualized with O-arm imaging and confirmed by palpation (3.5% of screws placed). Of the breaches, 14 screws stimulated below the 12-mA threshold, nine screws stimulated between 12 and 20 mA, and 12 screws did not generate an EMG response. Forty screws stimulated below a 12-mA threshold but showed no breach on imaging. Using the 12-mA threshold, the sensitivity of EMG was 40%, specificity was 96%, positive predictive value was 26%, and negative predictive value was 98%. All 35 breached screws were corrected during surgery. There were no postoperative symptoms caused by breached screws and no patients required reoperation. ConclusionsOur findings indicate that EMG may not be a highly reliable tool in determining an anatomical breach during placement of lumbar pedicle screws. O-arm may be better for detecting either medial or lateral breaches than EMG stimulation if there are concerns about screw placement or for confirmation of placement before leaving the operating room.