Robot-assisted stereoelectroencephalography electrode placement in twenty-three pediatric patients: a high-resolution analysis of individual lead placement time and accuracy at a single institution.

ABSTRACT

PURPOSE: We describe a detailed evaluation of predictors associated with individual lead placement efficiency and accuracy for 261 stereoelectroencephalography (sEEG) electrodes placed for epilepsy monitoring in twenty-three children at our institution. METHODS: Intra- and post-operative data was used to generate a linear mixed model to investigate predictors associated with three outcomes (lead placement time, lead entry error, lead target error) while accounting for correlated observations from the same patients. Lead placement time was measured using electronic time-stamp records stored by the ROSA software for each individual electrode; entry and target site accuracy was measured using postoperative stereotactic CT images fused with preoperative electrode trajectory planning images on the ROSA computer software. Predictors were selected from a list of variables that included patient demographics, laterality of leads, anatomic location of lead, skull thickness, bolt cap device used, and lead sequence number. RESULTS: Twenty-three patients (11 female, 48%) of mean age 11.7 (± 6.1) years underwent placement of intracranial sEEG electrodes (median 11 electrodes) at our institution over a period of 1 year. There were no associated infections, hemorrhages, or other adverse events, and successful seizure capture was obtained in all monitored patients. The mean placement time for individual electrodes across all patients was 6.56 (± 3.5) min; mean target accuracy was 4.5 (± 3.5) mm. Lesional electrodes were associated with 25.7% (95% CI 6.7-40.9%, p = 0.02) smaller target point errors. Larger skull thickness was associated with larger error for every 1-mm increase in skull thickness, there was a 4.3% (95% CI 1.2-7.5%, p = 0.007) increase in target error. Bilateral lead placement was associated with 26.0% (95% CI 9.9-44.5%, p = 0.002) longer lead placement time. The relationship between placement time and lead sequence number was nonlinear it decreased consistently for the first 4 electrodes, and became less pronounced thereafter. CONCLUSIONS: Variation in sEEG electrode placement efficiency and accuracy can be explained by phenomena both within and outside of operator control. It is important to keep in mind the factors that can lead to better or worse lead placement efficiency and/or accuracy in order to maximize patient safety while maintaining the standard of care.