High-Resolution Spinal Motor Mapping Using Thoracic Spinal Cord Stimulation in Patients With Chronic Pain

Machine learning predicts spinal cord stimulation surgery outcomes and reveals novel neural markers for chronic pain

Spinal cord stimulation (SCS) is a well-accepted therapy for refractory chronic pain. However, predicting responders remain a challenge due to a lack of objective pain biomarkers. The present study applies machine learning to predict which patients will respond to SCS based on intraoperative electroencephalogram (EEG) data and recognized outcome measures. The study included 20 chronic pain patients who were undergoing SCS surgery. During intraoperative monitoring, EEG signals were recorded under SCS OFF (baseline) and ON conditions, including tonic and high density (HD) stimulation. Once spectral EEG features were extracted during offline analysis, principal component analysis (PCA) and a recursive feature elimination approach were used for feature selection. A subset of EEG features, clinical characteristics of the patients and preoperative patient reported outcome measures (PROMs) were used to build a predictive model. Responders and nonresponders were grouped based on 50% reduction in 3-month postoperative Numeric Rating Scale (NRS) scores. The two groups had no statistically significant differences with respect to demographics (including age, diagnosis, and pain location) or PROMs, except for the postoperative NRS (worst pain: p = 0.028; average pain: p < 0.001) and Oswestry Disability Index scores (ODI, p = 0.030). Alpha-theta peak power ratio differed significantly between CP3-CP4 and T3-T4 (p = 0.019), with the lowest activity in CP3-CP4 during tonic stimulation. The decision tree model performed best, achieving 88.2% accuracy, an F1 score of 0.857, and an area under the curve (AUC) of the receiver operating characteristic (ROC)  of 0.879. Our findings suggest that combination of subjective self-reports, intraoperatively obtained EEGs, and well-designed machine learning algorithms might be potentially used to distinguish responders and nonresponders. Machine and deep learning hold enormous potential to predict patient responses to SCS therapy resulting in refined patient selection and improved patient outcomes.

Correlating Evoked Electromyography and Anatomic Factors During Spinal Cord Stimulation Implantation With Short-Term Outcomes

INTRODUCTION

We examine ways intraoperative neuromonitoring during spinal cord stimulation (SCS) varies between a high-resolution investigational SCS (HR-SCS) paddle and a commercial paddle. Furthermore, the presence of evoked motor responses (eg, electromyography [EMG]) in painful regions during surgery is correlated to outcomes.

MATERIALS AND METHODS

We used HR-SCS to assess EMG response from 18 patients (NCT05459324). Maximum percentage change in root mean squared (maxRMS) EMG values was determined. Correlations were performed with magnetic resonance imaging measurements and patient outcomes collected preoperatively and at three months (numerical rating scale [NRS], McGill Pain, Beck Depression Inventory, Oswestry Disability Index [ODI], and Pain Catastrophizing Score).

RESULTS

Of the 18 patients (12 women to six men; mean age 56 years; eight with neuropathic pain, eight with persistent spinal pain syndrome, two with complex regional pain syndrome), nine had a response at three months based on 50% reduction in NRS, 14 by achieving minimal clinically important difference (MCID) on NRS, and 11 by reaching MCID on ≥three outcome metrics. The anterior posterior diameter (APD) of the spinal column at level of testing correlated with all three responses (p < 0.05). We examined RMS at muscles correlating with individual patient pain distributions and found correlations between RMS and MCID NRS and MCID ODI (p < 0.05). maxRMS in abductor hallucis correlated with improvement in NRS and ODI across the group (p < 0.05).

CONCLUSIONS

We found that eliciting EMGs over the painful areas during surgery caused alleviation of pain intensity and disability. Obtaining stimulation of abductor hallucis (AH) was more predictive of pain improvement than any other muscle group, and APD alone correlated with improvements in pain intensity and holistic outcomes. These pilot data suggest that implanters should consider APD and EMG responses from painful regions and AH during surgery.

Improved Selectivity in Eliciting Evoked Electromyography Responses With High-Resolution Spinal Cord Stimulation

BACKGROUND AND OBJECTIVES

As spinal cord stimulation (SCS) offers a therapy for increasing numbers of patients with chronic pain and spinal cord injury, it becomes increasingly important to better understand its somatotopy. In this prospective study, we investigate whether high-resolution SCS (HR-SCS) offers improved selectivity assessed through elicitation of evoked electromyography (EMG) responses as compared with commercial paddle leads.

METHODS

Vertical tripole configurations were used to elicit EMG responses in both types of paddles placed for standard-of-care indications between T6 and T10. In HR-SCS, evoked EMG responses in lower extremity/abdominal muscle groups were monitored at 6 to 8 mediolateral sites. All commercial paddle columns were tested. Percentage change in the maximum root mean square value was calculated at a group level. Heat maps were generated to identify responders for each muscle group. Responders were considered patients who had a >50% change in root mean square over baseline.

RESULTS

We demonstrated significantly greater motor responses across medial and lateral contacts and greater responder rates consistently at the T6 and T9 levels with HR-SCS as compared with commercial paddles in 18 patients. Distal muscle groups (gastrocnemius and tibialis anterior) and proximal muscle groups (biceps femoris and quadriceps) were selectively activated at both levels.

CONCLUSION

We demonstrate that HR-SCS has greater selectivity in eliciting evoked EMG responses in an intraoperative setting. HR-SCS offers recruitment of muscle groups at lateral contacts concurrently with medial contacts. We provide data that HR-SCS may provide higher spatial resolution, which has the potential to allow for personalization of care and treatment of pain syndromes/symptoms which to date have not been effectively treated.

Investigation of the intraoperative cortical responses to spinal motor mapping in a patient with chronic pain

Intraoperative neurophysiological monitoring (IONM) in spinal cord stimulation (SCS) surgery for chronic pain is shown to provide effective guidance during device placement. Electromyography (EMG) is used to determine the laterality of the paddle. In some SCS cases, laterality cannot be obtained via EMG due to patient physiology. Electroencephalography (EEG) is already used in IONM to monitor cortical responses. Here, we show proof-of-concept of assessing the responses of epidurally evoked EMGs simultaneously with EEGs to determine laterality during IONM using a high-resolution (HR) SCS paddle. An 8-column HR-SCS paddle was acutely placed at T9-T10 interspace in patients with failed back surgery syndrome. EMG signals from 18 muscle groups were recorded simultaneously with 60-channel EEG signals at various stimulation amplitudes (0-10 mA). Particular attention was paid to regions associated with pain including the somatosensory cortex (S1), prefrontal cortex (PFC), and motor cortex (M1). When left and right lateral contacts were stimulated at low amplitudes (1-2 mA), significant changes were seen in θ, α, and β powers in the contralateral PFC but not in M1 or S1. There was a significant correlation between M1 and contralateral contacts in α power. At higher currents (7-8 mA), right-sided contacts resulted in α power change. We found significant differences in α, θ, and β powers in PFC for contralateral stimulation of the lateral SCS contacts at low amplitudes and in α power at higher amplitudes. The changes in PFC suggest the potential of EEG for understanding a cortical mechanism of action of SCS and provide insight into the pathophysiology of chronic pain.NEW & NOTEWORTHY Here, we present proof of concept of assessing the responses of epidurally evoked electromyography simultaneously with scalp electroencephalography to determine whether both laterality and insights into pain mechanisms can be elucidated. With stimulation, significant changes were seen in θ, α, and β band power in the contralateral prefrontal cortex and in α power in the motor cortex. We provide insight into the mechanism of action of SCS in preventing pain in this patient.

Neurotechnology for Pain

Neurotechnologies for treating pain rely on electrical stimulation of the central or peripheral nervous system to disrupt or block pain signaling and have been commercialized to treat a variety of pain conditions. While their adoption is accelerating, neurotechnologies are still frequently viewed as a last resort, after many other treatment options have been explored. We review the pain conditions commonly treated with electrical stimulation, as well as the specific neurotechnologies used for treating those conditions. We identify barriers to adoption, including a limited understanding of mechanisms of action, inconsistent efficacy across patients, and challenges related to selectivity of stimulation and off-target side effects. We describe design improvements that have recently been implemented, as well as some cutting-edge technologies that may address the limitations of existing neurotechnologies. Addressing these challenges will accelerate adoption and change neurotechnologies from last-line to first-line treatments for people living with chronic pain.

Intraoperative electrical stimulation of the human dorsal spinal cord reveals a map of arm and hand muscle responses

Although epidural stimulation of the lumbar spinal cord has emerged as a powerful modality for recovery of movement, how it should be targeted to the cervical spinal cord to activate arm and hand muscles is not well understood, particularly in humans. We sought to map muscle responses to posterior epidural cervical spinal cord stimulation in humans. We hypothesized that lateral stimulation over the dorsal root entry zone would be most effective and responses would be strongest in the muscles innervated by the stimulated segment. Twenty-six people undergoing clinically indicated cervical spine surgery consented to mapping of motor responses. During surgery, stimulation was performed in midline and lateral positions at multiple exposed segments; six arm and three leg muscles were recorded on each side of the body. Across all segments and muscles tested, lateral stimulation produced stronger muscle responses than midline despite similar latency and shape of responses. Muscles innervated at a cervical segment had the largest responses from stimulation at that segment, but responses were also observed in muscles innervated at other cervical segments and in leg muscles. The cervical responses were clustered in rostral (C4-C6) and caudal (C7-T1) cervical segments. Strong responses to lateral stimulation are likely due to the proximity of stimulation to afferent axons. Small changes in response sizes to stimulation of adjacent cervical segments argue for local circuit integration, and distant muscle responses suggest activation of long propriospinal connections. This map can help guide cervical stimulation to improve arm and hand function.NEW & NOTEWORTHY A map of muscle responses to cervical epidural stimulation during clinically indicated surgery revealed strongest activation when stimulating laterally compared to midline and revealed differences to be weaker than expected across different segments. In contrast, waveform shapes and latencies were most similar when stimulating midline and laterally, indicating activation of overlapping circuitry. Thus, a map of the cervical spinal cord reveals organization and may help guide stimulation to activate arm and hand muscles strongly and selectively.