gms | German Medical Science

Objective: Intracerebral hemorrhage (ICH) poses a devastating complication of DBS surgery associated with high age and the use of microelectrodes. In preOP MRIs we encountered wide perivascular spaces (PVS) in cases with ICH during electrode implantation. PVS enlargement is associated with occurrence of spontaneous intracerebral hemorrhage (Raposo et al. 2019, Duperron et al. 2019) and may reflect a general vascular vulnerability. Here we aimed to explore the role of enlarged PVS as a risk indicator for ICH during DBS electrode implantation.

Methods: The data of 362 consecutive patients (01/2013-09/2020) were analyzed. Unsuitable MR images were excluded (movement artefacts, <3 Tesla, WMH of non-vascular origin e.g. multiple sclerosis, preexisting DBS, etc.), resulting in 270 MRIs. Prior to PVS burden quantification, 1mm isotropic T1 und T2 images were coregistered and T1 images segmented using SPM12’s segmentation algorithm in WM/GM/CSF. For normalization of the T2 contrast a smooth T2 CSF reference image was computed by 1) masking out the all signal non-CSF in the T2 image (using SPM’s CSF-T1 segmentation), and 2) extending the CSF signal to adjacent regions by a broad Gaussian smooth. The raw T2 is then divided by this reference image to obtain comparable contrasts. PVS burden was computed by the commonly used Frangi-filter for tubular structures (smin=0,4mm, smax=2,0mm, scale ratio=2, α=0,1, β=1, c=0,01) on the normalized T2 (Frangi et al. 1998). All responses in non-WM and in cerebellum, pons and medulla oblongata were excluded. Finally, PVS index was computed by a simple summation of all non-negative filter responses within white matter. A Mann-Whitney-U-Test across all 206 operations with microelectrodes (encompassing all cases with subcortical ICH) was calculated to analyze age and PVS-Index in cases with and without ICH.

Results: 10 cases of subcortical ICH occurred during implantation of 722 electrodes in 374 operations (2.7 %) in 362 patients. Relevant neurological deficits resolved in all but one of the affected patients. ICHs occurred more often in patients with PVS enlargement (p = .006, 2-sided) and higher age (p =.022, 2-sided).

Conclusion: Enlarged PVS may serve as an additional risk indicator for ICH in DBS surgery and help to mitigate invasivity (e.g. avoid microelectrodes) in patients at risk. The presented results are preliminary and further analyses are in preparation to address the sparse data bias (Greenland et al. 2016) due to low incidence of ICH.

Figure 1 [Fig. 1]