Development of Machine Learning-Based Models to Predict Treatment Response to Spinal Cord Stimulation

A Semi-structured Interview Predicts Spinal Cord Stimulator Implantation in Patients with Chronic Pain

Pre-surgical psychological evaluations (PSPE) are required during the spinal cord stimulator (SCS) implantation process, but there is no standard protocol for SCS PSPE. In this study, we assessed the concurrent and predictive validity of the Stanford Integrated Pyschosocial Assessment for Transplantation (SIPAT) compared with patient-reported measures and election for SCS implantation. This study used prospectively collected data at the time of PSPE from N = 222 patients at a Midwestern academic medical center. We collected SIPAT scores and Patient-Reported Outcome Measurement Information Systems (PROMIS) scores, and recorded receipt of permanent SCS implantation as a binary (yes/no) outcome. The SIPAT correlated with patient-reported outcomes of Anxiety, Depression, Fatigue, Sleep, and Pain Interference in the expected direction. The SIPAT was a significant predictor of election for permanent SCS implantation when accounting for age and pain diagnosis, such that individuals with higher SIPAT scores were less likely to elect for surgery. Exploratory analyses showed that the SIPAT Patient Readiness subscale and patient-reported Anxiety and Depression PROMIS scales correlated with election for SCS surgery. Results of this study demonstrated validity of the SIPAT in a novel population, patients with chronic pain referred for SCS implantation.

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.

Moving towards the use of artificial intelligence in pain management

BACKGROUND AND OBJECTIVE

While the development of artificial intelligence (AI) technologies in medicine has been significant, their application to acute and chronic pain management has not been well characterized. This systematic review aims to provide an overview of the current state of AI in acute and chronic pain management.

DATABASES AND DATA TREATMENT

This review was registered with PROSPERO (ID# CRD42022307017), the international registry for systematic reviews. The search strategy was prepared by a librarian and run in four electronic databases (Embase, Medline, Central, and Web of Science). Collected articles were screened by two reviewers. Included studies described the use of AI for acute and chronic pain management.

RESULTS

From the 17,601 records identified in the initial search, 197 were included in this review. Identified applications of AI were described for treatment planning as well as treatment delivery. Described uses include prediction of pain, forecasting of individualized responses to treatment, treatment regimen tailoring, image-guidance for procedural interventions and self-management tools. Multiple domains of AI were used including machine learning, computer vision, fuzzy logic, natural language processing and expert systems.

CONCLUSION

There is growing literature regarding applications of AI for pain management, and their clinical use holds potential for improving patient outcomes. However, multiple barriers to their clinical integration remain including lack validation of such applications in diverse patient populations, missing infrastructure to support these tools and limited provider understanding of AI.

SIGNIFICANCE

This review characterizes current applications of AI for pain management and discusses barriers to their clinical integration. Our findings support continuing efforts directed towards establishing comprehensive systems that integrate AI throughout the patient care continuum.

Socioeconomic Determinants of Initiating Neuromodulation for Chronic Pain: A Systematic Review

OBJECTIVES

Neuromodulation is an effective treatment for chronic pain; however, socioeconomic differences may influence decision-making to initiate this therapy. This review investigated potential differences in accessibility of neuromodulation for patients with chronic pain due to socioeconomic determinants.

MATERIALS AND METHODS

Four electronic databases were used for this systematic review: MEDLINE, Embase, Scopus, and Web of Science. Risk of bias was assessed using the modified version of the Downs and Black checklist. The study protocol was prospectively registered on PROSPERO (CRD42023426035).

RESULTS

The initial database search identified a total of 1118 unique studies, of which 36 were eventually included in the systematic review. Of the 36 included studies, six studies reported on education, 24 on employment status, ten on insurance, five on household income, and three on miscellaneous topics. Neuromodulation seems accessible for patients with different education levels and different types of insurance. Additionally, it is not restricted to patients who are (un)employed. When comparing patients who initiated neuromodulation with those who did not, a significantly higher number of patients in the top quartile for education were found in the group without neuromodulation. Regarding insurance, inconclusive evidence was found.

CONCLUSIONS

Although neuromodulation was accessible for patients with varying levels of socioeconomic determinants, disparities were noted. When comparing the socioeconomic profiles of patients who receive neuromodulation and those who do not, education levels differ. Health-related inequality should be carefully monitored in chronic pain management with neuromodulation to ensure that potential disparities do not increase.

The Role and Applications of Artificial Intelligence in the Treatment of Chronic Pain

PURPOSE OF REVIEW

This review aims to explore the interface between artificial intelligence (AI) and chronic pain, seeking to identify areas of focus for enhancing current treatments and yielding novel therapies.

RECENT FINDINGS

In the United States, the prevalence of chronic pain is estimated to be upwards of 40%. Its impact extends to increased healthcare costs, reduced economic productivity, and strain on healthcare resources. Addressing this condition is particularly challenging due to its complexity and the significant variability in how patients respond to treatment. Current options often struggle to provide long-term relief, with their benefits rarely outweighing the risks, such as dependency or other side effects. Currently, AI has impacted four key areas of chronic pain treatment and research: (1) predicting outcomes based on clinical information; (2) extracting features from text, specifically clinical notes; (3) modeling 'omic data to identify meaningful patient subgroups with potential for personalized treatments and improved understanding of disease processes; and (4) disentangling complex neuronal signals responsible for pain, which current therapies attempt to modulate. As AI advances, leveraging state-of-the-art architectures will be essential for improving chronic pain treatment. Current efforts aim to extract meaningful representations from complex data, paving the way for personalized medicine. The identification of unique patient subgroups should reveal targets for tailored chronic pain treatments. Moreover, enhancing current treatment approaches is achievable by gaining a more profound understanding of patient physiology and responses. This can be realized by leveraging AI on the increasing volume of data linked to chronic pain.

Spinal Cord Stimulator Inequities Within the US Military Health System

OBJECTIVES

Although studies have described inequities in spinal cord stimulation (SCS) receipt, there is a lack of information to inform system-level changes to support health care equity. This study evaluated whether Black patients exhaust more treatment options than do White patients, before receiving SCS.

MATERIALS AND METHODS

This retrospective cohort study included claims data of Black and non-Latinx White patients who were active-duty service members or military retirees who received a persistent spinal pain syndrome (PSPS) diagnosis associated with back surgery within the US Military Health System, January 2017 to January 2020 (N = 8753). A generalized linear model examined predictors of SCS receipt within two years of diagnosis, including the interaction between race and number of pain-treatment types received.

RESULTS

In the generalized linear model, Black patients (10.3% [8.7%, 12.0%]) were less likely to receive SCS than were White patients (13.6% [12.7%, 14.6%]) The interaction term was significant; White patients who received zero to three different types of treatments were more likely to receive SCS than were Black patients who received zero to three treatments, whereas Black and White patients who received >three treatments had similar likelihoods of receiving a SCS.

CONCLUSIONS

In a health care system with intended universal access, White patients diagnosed with PSPS tried fewer treatment types before receiving SCS, whereas the number of treatment types tried was not significantly related to SCS receipt in Black patients. Overall, Black patients received SCS less often than did White patients. Findings indicate the need for structured referral pathways, provider evaluation on equity metrics, and top-down support.

Variables associated with nonresponders to high-frequency (10 kHz) spinal cord stimulation

INTRODUCTION

The use of spinal cord stimulation (SCS) therapy to treat chronic pain continues to rise. Optimal patient selection remains one of the most important factors for SCS success. However, despite increased utilization and the existence of general indications, predicting which patients will benefit from neuromodulation remains one of the main challenges for this therapy. Therefore, this study aims to identify the variables that may correlate with nonresponders to high-frequency (10 kHz) SCS to distinguish the subset of patients less likely to benefit from this intervention.

MATERIALS AND METHODS

This was a retrospective single-center observational study of patients who underwent 10 kHz SCS implant. Patients were divided into nonresponders and responders groups. Demographic data and clinical outcomes were collected at baseline and statistical analysis was performed for all continuous and categorical variables between the two groups to calculate statistically significant differences.

RESULTS

The study population comprised of 237 patients, of which 67.51% were responders and 32.49% were nonresponders. There was a statistically significant difference of high levels of kinesiophobia, high self-perceived disability, greater pain intensity, and clinically relevant pain catastrophizing at baseline in the nonresponders compared to the responders. A few variables deemed potentially relevant, such as age, gender, history of spinal surgery, diabetes, alcohol use, tobacco use, psychiatric illness, and opioid utilization at baseline were not statistically significant.

CONCLUSION

Our study is the first in the neuromodulation literature to raise awareness to the association of high levels of kinesiophobia preoperatively in nonresponders to 10 kHz SCS therapy. We also found statistically significant differences with greater pain intensity, higher self-perceived disability, and clinically relevant pain catastrophizing at baseline in the nonresponders relative to responders. It may be appropriate to screen for these factors preoperatively to identify patients who are less likely to respond to SCS. If these modifiable risk factors are present, it might be prudent to consider a pre-rehabilitation program with pain neuroscience education to address these factors prior to SCS therapy, to enhance successful outcomes in neuromodulation.

Machine Learning in Pain Neuromodulation

This chapter highlights the intersection of pain neuromodulation and machine learning (ML), exploring current limitations in pain management and how ML techniques can address these challenges. Neuromodulation technologies, such as spinal cord stimulation (SCS), have emerged as promising interventions for chronic pain, but limitations such as patient selection have resulted in high rates of failure and costly removal of these devices. ML offers a powerful approach to augment pain management outcomes by leveraging predictive modeling for enhanced patient selection, adaptive algorithms for programming optimization, and identification of objective biomarkers for improved outcome assessment. This chapter discusses various ML applications in pain neuromodulation and how we can expect it to shape the future of the field. While ML holds great promise, challenges such as algorithm transparency, data quality, and generalizability must be addressed to fully realize its potential in revolutionizing pain management.

Dorsal Root Entry Zone Lesioning Following Unresponsive Spinal Cord Stimulation for Post-Traumatic Neuropathic Pain

OBJECTIVE

Spinal cord stimulation (SCS) and dorsal root entry zone (DREZ) lesioning are important therapeutic options for intractable post-traumatic neuropathic pain (PNP). However, surgical choice is controversial due to the need to maximize pain relief and reduce complications. This study aims to retrospectively analyze the effect and complications of DREZ lesioning for patients with PNP who were unresponsive to SCS and provide a surgical reference.

METHODS

Demographic data and surgical characteristics of patients with PNP who underwent DREZ lesioning after an unresponsive SCS were reviewed. Long-term outcomes including numeric rating scale, global impression of change, and long-term complications were assessed. Kaplan-Meier analysis was used to evaluate pain-free survival.

RESULTS

Of 19 patients with PNP, 8 had brachial plexus injury (BPI), 7 had spinal cord injury, 2 had cauda equina injury, 1 had intercostal nerve injury, and 1 had lumbosacral plexus injury. All patients were unresponsive or had a recurrence of pain after SCS, with an average pain-relief rate of 9.3%. After DREZ lesioning, the mean numeric rating scale scores significantly decreased from 7.6 ± 1.5 to 1.8 ± 1.7, with an average pain-relief rate of 75.3%. Seven patients (36.8%) experienced worsened neurologic dysfunction at the last follow-up. Patients with BPI had a significantly better outcome than other pathologies (P < 0.001) after DREZ lesioning.

CONCLUSIONS

DREZ lesioning is an effective alternative procedure to SCS for patients with PNP who have lost limb function. Particularly for those with BPI, DREZ lesioning has shown good efficacy and can be considered a preferred surgical option.

Daring discourse: artificial intelligence in pain medicine, opportunities and challenges

Artificial intelligence (AI) tools are currently expanding their influence within healthcare. For pain clinics, unfettered introduction of AI may cause concern in both patients and healthcare teams. Much of the concern stems from the lack of community standards and understanding of how the tools and algorithms function. Data literacy and understanding can be challenging even for experienced healthcare providers as these topics are not incorporated into standard clinical education pathways. Another reasonable concern involves the potential for encoding bias in healthcare screening and treatment using faulty algorithms. And yet, the massive volume of data generated by healthcare encounters is increasingly challenging for healthcare teams to navigate and will require an intervention to make the medical record manageable in the future. AI approaches that lighten the workload and support clinical decision-making may provide a solution to the ever-increasing menial tasks involved in clinical care. The potential for pain providers to have higher-quality connections with their patients and manage multiple complex data sources might balance the understandable concerns around data quality and decision-making that accompany introduction of AI. As a specialty, pain medicine will need to establish thoughtful and intentionally integrated AI tools to help clinicians navigate the changing landscape of patient care.

Machine Learning in Spinal Cord Stimulation for Chronic Pain

Spinal cord stimulation (SCS) is an effective treatment for chronic neuropathic pain. The success of SCS is dependent on candidate selection, response to trialing, and programming optimization. Owing to the subjective nature of these variables, machine learning (ML) offers a powerful tool to augment these processes. Here we explore what work has been done using data analytics and applications of ML in SCS. In addition, we discuss aspects of SCS which have narrowly been influenced by ML and propose the need for further exploration. ML has demonstrated a potential to complement SCS to an extent ranging from assistance with candidate selection to replacing invasive and costly aspects of the surgery. The clinical application of ML in SCS shows promise for improving patient outcomes, reducing costs of treatment, limiting invasiveness, and resulting in a better quality of life for the patient.