Impact of preoperative computed tomography image slice thickness on the planning of deep brain stimulation surgery: A phantom study

Abstract

Background

Preoperative Magnetic Resonance Imaging (MRI) and Computed Tomography (CT) coregistration is often applied to perform deep brain stimulation (DBS) surgeries. Thin sliced (high dose) CT is used and recommended because of the high reported accuracy but there are no comparative studies. Thin sliced CT could lead to a higher radiation exposure for the patient in contrast to thicker sliced CT (low dose) due to a longer scanning time. In this study we investigated if the preoperative CT slice thickness had an effect on the registration accuracy in a preoperative simulated DBS setting.

Material and methods

An implanted phantom, a Citrullus lanatus (watermelon), was used to acquire an MRI data set (2 millimeter (mm) T2) with a fixed reference point. Two observers targeted the reference point independently and they recorded and compared the coordinates, ring, and arc angles from all coregistered series with different CT slice thicknesses of 0.5 mm, 1 mm, 1.5 mm, 2 mm, 3 mm, 4 mm, and 5 mm. Separately, both MRI and CT were used as reference series for coregistration and analysis. Lastly, inter-observer reliability was calculated with Kendall’s coefficient of concordance (W). With W> 0.9 defined as very good.

Results

Our results show no relevant effect on the preoperative registration accuracies for the different CT’s with all absolute differences in mm for the stereotactic coordinates< 0.5 and angles in degrees< 0.4. Additionally, the inter-observer reliability was high (W 0.991).

Conclusion

We found no relevant effect of increased slice thickness of preoperative CT on the preoperative registration accuracy in a simulated DBS setting.