(50) Personalizing Spinal Cord Stimulation Therapy with Artificial Intelligence for Optimal Patient Care

Introduction: In healthcare, artificial intelligence (AI), has been leveraged as a predictive tool for patient selection, and therapy customization for improved outcomes.[1] In the field of neuromodulation, where personalization and optimization of therapy is needed, data from robust data pipelines can be leveraged to feed decision making algorithms to customize therapy for improved chronic pain management. The use of AI-enabled systems for therapy optimization without clinician intervention is limited to date [2]. We report a case study of a patient implanted with the HFX iQ Spinal Cord Stimulation (SCS) system (Nevro Corp, Redwood City), a system that uses patient inputs from a mobile phone application to provide customized therapy recommendations [3].

Methods: This system collects 187 unique variables for each patient. The system integrates patient reported outcomes and device function basic (e.g., low battery, failing to charge, not using devices) and advanced monitoring (e.g., relief status, lead impedance). The mobile app prompts the patient daily on their pain relief, sleep, activity and pain medication use. The patient status informs the expert recommendation algorithm (HFX iQ) that may then recommend a therapy change. The therapy parameters that can be customized by HFX iQ include therapy target bipole(s), amplitude, and waveform, while any adjustments to therapy, including returning to a previously favorited setting, remain in the patient’s control.

Results: The case study presented is an 80-year-old male with chronic bilateral leg pain, who was implanted with the HFX iQ SCS system on 2/2024. The app-based assessments show that the patient finds pain relief early but experiences wide fluctuations in the first 90 days (Figure 1). As the level of relief falls below a threshold and other factors are met, HFX iQ sends the patient a therapy change recommendation, leading to more stable pain relief. Patients’ pain relief continues to be optimized iteratively and sustained through HFX iQ’s recommendations. Other factors of device integrity, such as lead impedances are continuously monitored (Figure 2) and an alert is sent to a care team member to resolve the issue.

Conclusion: Leveraging the evolving potential of AI in SCS, patients can achieve and return to pain relief more efficiently than previously, as well as be more adherent to therapy through stronger engagement and notifications related to device functioning. AI-driven shifts toward pulse dosing can also lower charging burden and more variation in therapy targets and parameters may reduce habituation, all leading to higher therapy satisfaction and improved outcomes for patients.