Automated Image Guided Programming Algorithm Supports Clinicians during DBS Programming for Parkinson's Disease Patients
Jason Aldred1, Corneliu Luca2, Adolfo Ramirez Zamora3, Joshua Wong4, Kristy Wessels1, Taylor Peabody5, Ben Reese6, Beth Farber-Petrey7, Richard Mustakos7, Soroush Niketeghad7, Rajat Shivacharan7, Mahsa Malekmohammadi7
1Selkirk Neurology, 2University of Miami, 3University of Florida - Fixel Neurological Institute, 4University of Florida College of Medicine - Neurology, 5The University of Tennessee Medical Center, 6Boston Scientific, 7Boston Scientific Neuromodulation
Objective:
The clinical utility of an automated image-guided programming (aIGP) algorithm for optimizing deep brain stimulation (DBS) settings in patients with Parkinson’s disease is analyzed, and its performance against standard-of-care (SoC) programming is assessed.
Background:
As a process that can be time-consuming and burdensome for patients, DBS programming traditionally relies on iterative, clinician-led adjustments to balance therapeutic benefit and side effects. Recent advances in imaging and computational modeling have enabled algorithms that rapidly generate anatomically-informed stimulation settings. Automated image-guided programming can use patient-specific imaging to position a stimulation field model (SFM) within seconds, in turn targeting desired regions while minimizing spread to non-target areas.
Design/Methods:
To date, this prospective, multicenter, double-blinded crossover study has enrolled 15 PD patients with bilateral STN or GPi DBS systems. Participants were evaluated acutely and chronically under both SoC and aIGP-guided programming. Motor outcomes were assessed using MDS-UPDRS III (meds off/on), and chronic efficacy was evaluated via motor diaries. Clinicians could optimize aIGP-suggested settings, and changes were quantified using similarity metrics.
Results:
Initial aIGP DBS-on settings improved motor scores by 39% (meds-off) and 54% (meds-on) versus DBS-off (p<0.0001), with no significant difference from optimized SoC programming (p=0.44 and p=0.33, respectively). Chronic assessments showed aIGP-guided programs yielded 25 minutes/day more “ON” time and 58 minutes/day less “ON with dyskinesias” compared to SoC, though group-level differences were not statistically significant. Clinician-optimized aIGP settings remained largely unchanged. Additional data collected in the interim will be reported.
Conclusions:
Automated image-guided programming provides rapid, anatomically guided programming that achieves motor outcomes comparable to traditional methods. Its ability to streamline DBS optimization may reduce clinical burden and enhance personalization, supporting its integration into routine neuromodulation workflows.
Disclaimer: Abstracts were not reviewed by Neurology® and do not reflect the views of Neurology® editors or staff.