AIS is a complex three-dimensional deformity associated with rotation and structural abnormalities of the vertebrae, making treatment technically challenging. Implants for pedicle fixation have been widely used in surgical treatments of the thoracic and lumbar spine, with better results in arthrodesis rate, correction power and early mobilization of the patient compared with fixation systems that employ hooks or mixed systems [17, 18]. Neurological complications due to incorrect positioning of pedicle screws is rare, comprising 0.9% of cases as described in the literature [19, 20]. In this study, we analyzed the occurrence of pedicle screw misplacement in a specific pathology, AIS. Additionally, we evaluated the accuracy of EMG as a diagnostic tool to predict screws ARNI in AIS surgery.
Pedicle screw misplacement is detected in 3–44% of cases in the literature [20], and in this study, the rate was 40.9%. Screws with LCP are associated with risk of vascular or visceral damage [2, 9, 21, 22]. In the present series, 14.6% of screws had LCP, in line with the literature [2, 21, 22], and no cases were associated with complications. For MCP, misplacement rates of 1.4–14% have been reported in the literature [2, 9, 22], reaching 28% in one series [23]. In the present study, 16% of screws had MCP according to postoperative CT. There are limited data in the literature describing the misplacement of pedicle screws in the sagittal plane with superior or inferior cortical perforation.
Intraoperative neurophysiology evaluation can allow the early detection and correction of possible lesions during spinal surgery. Such techniques include the evaluation of motor evoked potentials, somatosensory evoked potentials and EMG [10, 11]. The role of the EMG stimulus in the early identification of pedicle cortical perforation has been established for lumbar pedicles. Thresholds below 4 or 5 mA are suggestive of perforation [10], while thresholds above 15 mA indicate correct positioning of screws [24]. However, the correlation between EMG thresholds and screw positioning in thoracic pedicles has not yet been well established.
In evaluating the accuracy of EMG as an intraoperative diagnostic method to detect misplaced screws ARNI, there was a statistically significant association between EMG responses and the positioning of screws associated with risk for nerve injury. A decreased EMG threshold was associated with an increased odds of screw position ARNI among thoracic and lumbar screws. The association between EMG threshold and screw misplacement in thoracic pedicles was recently shown using pulse-train stimulation [25]. The previous study only evaluated the association between EMG stimulation and screw position in the axial plane with MCP, while the present study evaluated both the axial and sagittal planes.
Despite the association found between EMG and the position of screws, the ability of EMG to intraoperatively discriminate between screws ARNI and NRNI was poor to moderate. NPV and PPV are the most meaningful measures of diagnostic accuracy in terms of making clinical decisions based on a test result. NPV expresses the probability of not having the condition under study given a negative test outcome, and PPV expresses the probability of having the condition given a positive test outcome. EMG showed a very low PPV (< 18%) at every threshold cutoff evaluated, meaning that less than 1 out of 5 screws that test positive (achieving an electrical response at a threshold lower than the cutoff) would actually be positioned ARNI.
Considering the risk for neurological deficit or stenosis of the spinal canal if a screw breaches the medial wall during thoracic pedicle screw instrumentation in AIS surgery, the consequences of a false-negative result of a diagnostic test for screw malposition can be severe [25]. Therefore, it is imperative that the diagnostic test detect true positives and minimize false negatives, as represented by a high NPV. The present study revealed a moderate to high NPV of EMG as diagnostic test for every cutoff analyzed, and thus, EMG may be considered an accurate test to minimize false-negative screws ARNI.
The main limitation of the present study is that despite the aim to analyze the ability of EMG to intraoperatively predict screws ARNI, the sample was composed exclusively of patients with no nerve injury, spinal cord injury, or nerve root injury. Some false-positive and false-negative cases were found among the EMG responses of the diagnostic test, but these cases did not result in any clinical consequence. Therefore, it is not possible to assume that EMG is not an accurate tool for predicting screws ARNI. Furthermore, we consider intraoperative neurophysiologic monitoring, particularly motor evoked potentials, as extremely important during thoracic screw insertion for the early detection and prevention of severe neurological complications.