Validation of automated image-guided programming in STN-DBS for Parkinson's disease

Current steering with multiple independent current control (MICC) and directional leads has expanded programming options in subthalamic nucleus deep brain stimulation (STN-DBS), but has also increased programming complexity. Manual image-guided programming (mIGP) using patient-specific anatomical information facilitates optimization of stimulation settings. More recently, algorithm-based platforms for automated image-guided programming (aIGP) have been developed, although their clinical validity remains unclear.

Thirteen patients with Parkinson's disease who had undergone STN-DBS using a MICC directional lead system and had stable motor symptoms under mIGP established with StimviewTM XT were enrolled. aIGP was performed using IlluminaTM 3D under two conditions. In Part 1, the target region was the STN and the avoid region was the internal capsule. In Part 2, the target region was the STN, whereas the red nucleus and substantia nigra were designated as avoid regions to direct stimulation toward the dorsolateral STN. Stimulation amplitude, pulse width, and frequency were unchanged; only the spatial distribution of stimulation was modified. Motor symptoms were assessed using the Unified Parkinson's Disease Rating Scale part III (UPDRS III) immediately before and 1 hour after switching from mIGP to aIGP-derived settings.

In Part 1, 4 of 13 patients showed obvious worsening shortly after switching and were immediately returned to mIGP. Of the remaining 9, 8 reverted to mIGP within 1 week, whereas 1 continued aIGP because of improved speech difficulty. Mean UPDRS III scores changed from 16.4 ± 1.7 to 19.6 ± 2.5 (P = 0.16). In Part 2, 3 of 13 patients developed tremor shortly after switching and were immediately returned to mIGP. The remaining 10 all reverted to mIGP within 2 weeks. Mean UPDRS III scores changed from 15.5 ± 1.5 to 18.3 ± 1.8 (P = 0.07).

Although aIGP rapidly generated anatomically plausible current-steering configurations, most patients did not maintain these settings, indicating limited clinical acceptability compared with expert-guided mIGP. Further refinement, particularly with more selective emphasis on the dorsolateral STN, will be necessary before aIGP can be considered a reliable alternative.