Fig 1.
Preoperative volumetric analyses.
(A) T1 MP-RAGE coronal cut through the hippocampal body for a representative patient. (B) Same image showing manually-traced borders of CSFAbove (light blue) and CSFLateral (dark blue) and automatically segmented hippocampal borders (green). (C) 3D model for the same patient showing relationship of CSF structures to amygdala (orange) and hippocampus (green). The red arrow shows the imaging cut used for A & B. (D-F) Same conventions as in A—C for a patient with larger CSF spaces. The white asterisks in E shows the vertical digitation of the hippocampus at its posterior end, which was used as a consistent anatomic landmark for laser position calculations. Scale bar– 5 mm. AC, anterior commissure; MP, midline point; PC, posterior commissure.
Fig 2.
Overlap of TDE maps with immediate and delayed ablation zone boundaries.
(A-B) Axial and sagittal TDE maps at the end of the LITT procedure for sample Patient 1. The white dashed boxes show the regions highlighted in the panels below. (C-D) Axial and sagittal post ablation contrasted T1 images showing the manually-traced immediate ablation zone (blue) and the boundary of the TDE (yellow) from the same imaging plane. The numbers below the image panels show the DSC values. (E-F) Axial and sagittal T1 MP-RAGE images 6-months after surgery demonstrating the manually-traced delayed ablations (red) and the TDE (yellow). (G-L) Same conventions as A—F but for Patient 2. A, anterior; Del, delayed; Imm, immediate; L, lateral; S, superior; TDE, thermal damage estimate.
Table 1.
Regression coefficients for independent variables with the DSC(DelayedAx, TDEAx).
Fig 3.
Comparison of averaged TDE maps and ablation zone boundaries for entire patient cohort.
(A) Composite maps showing averaged axial TDEs (yellow) and immediate ablation boundaries through the same imaging plane (blue) for all patients in the study (N = 28). Thick lines show mean distance from the laser and thin lines mean ± standard error. Vertical black lines show positions along the laser where TDE maps differed from the ablation zones with p ≤ 0.05, determined using bootstrap analysis. (B) Axial TDE maps (yellow) and delayed ablation boundaries (red). The axial MRI above estimates the mesiotemporal location of data used for A and B. (C) Sagittal TDE maps (yellow) and immediate ablation boundaries (blue) (N = 16). (D) Sagittal TDE maps (yellow) and delayed ablation boundaries (red). The sagittal MRI above estimates the mesiotemporal location of data used for C and D. A, anterior; Del, delayed; Imm, immediate; L, lateral; S, superior; TDE, thermal damage estimate.
Fig 4.
Dynamics of irreversible ablation during mesiotemporal LITT.
(A) Different frames of a sample axial TDE video. The larger image on the right shows a lower magnification view, with the white dashed box highlighting the region used in the left panels. Trajectories were traced from the first frame in each video and the appearance of yellow pixels indicating irreversible ablation was quantified relative to the laser using Matlab Image Processing Toolbox. The blue (lateral) and yellow (mesial) rectangles demonstrate regions of interest used to measure pixel counts on each side of the laser. (B) Time course of total (black), lateral (blue), mesial (yellow), and lateral—mesial (red) pixel counts from the same patient as in G. A, anterior; L, lateral.
Table 2.
Relationship of ablation dynamics measures to the independent variables.
Table 3.
Regression coefficients for independent variables with the thermal time constant.
Fig 5.
Mesiotemporal LITT ablation asymmetry.
(A) Time course of pixel counts lateral (blue) and mesial (yellow) to the laser averaged from all axial TDE maps (N = 29). Pixel counts were greater lateral vs. mesial to the laser past 60 seconds, with statistically significant differences between the curves shown with a horizontal line above the graph. At each time point, significance was evaluated with bootstrap analysis using p ≤ 0.05 as the criterion. (B) Pixel counts measured from the axial TDE videos fit with exponential curves for each patient (N = 28). (C) Difference in mesial and lateral pixel counts modeled with best-fit lines for each patient, done separately for early and late differences, distinguished by the 60 second time point (dotted grey line). (D-F) Same conventions as A—C but calculated from sagittal TDE maps (N = 16). Unlike for the axial maps, no differences were found for pixel counts measured superior (green) vs. inferior (red) to the laser.
Table 4.
Evaluation of ablation asymmetry.