Comparison between MR and CT imaging used to correct for skull-induced phase aberrations during transcranial focused ultrasound

Abstract

Transcranial focused ultrasound with the InSightec Exablate system uses thermal ablation for the treatment of movement and mood disorders and blood brain barrier disruption for tumor therapy. The system uses computed tomography (CT) images to calculate phase corrections that account for aberrations caused by the human skull. This work investigates whether magnetic resonance (MR) images can be used as an alternative to CT images to calculate phase corrections. Phase corrections were calculated using the gold standard hydrophone method and the standard of care InSightec ray tracing method. MR binary image mask, MR-simulated-CT (MRsimCT), and CT images of three ex vivo human skulls were supplied as inputs to the InSightec ray tracing method. The degassed ex vivo human skulls were sonicated with a 670 kHz hemispherical phased array transducer (InSightec Exablate 4000). 3D raster scans of the beam profiles were acquired using a hydrophone mounted on a 3-axis positioner system. Focal spots were evaluated using six metrics: pressure at the target, peak pressure, intensity at the target, peak intensity, positioning error, and focal spot volume. Targets at the geometric focus and 5 mm lateral to the geometric focus were investigated. There was no statistical difference between any of the metrics at either target using either MRsimCT or CT for phase aberration correction. As opposed to the MRsimCT, the use of CT images for aberration correction requires registration to the treatment day MR images; CT misregistration within a range of ± 2 degrees of rotation error along three dimensions was shown to reduce focal spot intensity by up to 9.4%. MRsimCT images used for phase aberration correction for the skull produce similar results as CT-based correction, while avoiding both CT to MR registration errors and unnecessary patient exposure to ionizing radiation.

Figure 1

CT contrast of the skull compared to various MR contrasts (skull A). Cortical and trabecular bone contrast is clearly depicted by CT and may be preserved by two of the three MR post-processing methods. The units for MRsimCT, –log(short TE), and short TE–long TE are not the same, thus different windowing and leveling were used. Scale bars and skull density ratios (SDRs) are shown at the bottom of each image. MRsimCT was the preferred choice based on bone contrast, minimal SDR change compared to CT, minimal background signal bias, and generalizable post-processing. Therefore, it was used for the remainder of this study.

Figure 2

Relationship between CT and MRsimCT values. MRsimCT values prior to HU_bone scaling in (2) span a range from 0 to 1 and can be used as an estimate for bone fraction. For the CT parameters used in this study, pure bone was calculated to have a value of 2000 HU. The black line depicts the nominal relationship between CT and MRsimCT values. Because only three ex vivo skulls were used in this study, using a linear regression to predict CT HU from MRsimCT values may overfit to the data. Therefore, the nominal linear relationship was used to predict CT HU instead.

Figure 3

Experimental setup used in this study. Three ex vivo skulls were used for experimentation. Each of the three skulls was fixed to a head frame to ensure consistent positioning. The head frame was secured in an InSightec 670 kHz hemispherical phased array transducer. Data were acquired with a needle hydrophone and 3-axis positioner system. Illustrations were drawn by Sarah Hwang.

Figure 4

2D cross section raster scans of the refocused focal spot (skull A). The sagittal view is shown. These cross sections are a subset of the 3D volume scans acquired. A red x marks the position of the target, which was placed at the geometric focus. The focal spot maximum may be out of plane. Figure 5 shows quantitative metrics for all three skulls.

Figure 5

Beamforming performance of different image types with the target placed at the geometric focus. (a) Target and peak pressure (shown with darker and lighter colors, respectively) when operating the transducer at 20 W of electrical power. (b) Target and peak intensity normalized to the hydrophone method. (c) Focal spot positioning error. (d) Focal spot volume. (e) Dice similarity coefficient. Standard deviation bars are shown.

Figure 6

Effects of CT misregistration on normalized target intensity. (a) Displacement and (b) rotation errors were applied along one dimension only. Axes are colored according to their corresponding dimensions.