Cavitation dose painting for focused ultrasound-induced blood-brain barrier disruption

Abstract

Focused ultrasound combined with microbubble for blood-brain barrier disruption (FUS-BBBD) is a promising technique for noninvasive and localized brain drug delivery. This study demonstrates that passive cavitation imaging (PCI) is capable of predicting the location and concentration of nanoclusters delivered by FUS-BBBD. During FUS-BBBD treatment of mice, the acoustic emissions from FUS-activated microbubbles were passively detected by an ultrasound imaging system and processed offline using a frequency-domain PCI algorithm. After the FUS treatment, radiolabeled gold nanoclusters, ⁶⁴Cu-AuNCs, were intravenously injected into the mice and imaged by positron emission tomography/computed tomography (PET/CT). The centers of the stable cavitation dose (SCD) maps obtained by PCI and the corresponding centers of the ⁶⁴Cu-AuNCs concentration maps obtained by PET coincided within 0.3 ± 0.4 mm and 1.6 ± 1.1 mm in the transverse and axial directions of the FUS beam, respectively. The SCD maps were found to be linearly correlated with the ⁶⁴Cu-AuNCs concentration maps on a pixel-by-pixel level. These findings suggest that SCD maps can spatially "paint" the delivered nanocluster concentration, a technique that we named as cavitation dose painting. This PCI-based cavitation dose painting technique in combination with FUS-BBBD opens new horizons in spatially targeted and modulated brain drug delivery.

Figure 1

Illustration of the FUS system, as well as images acquisition and registration methods. PCI was acquired during FUS treatment, and PET/CT image was acquired 24 h after FUS treatment. The registration between PCI and PET was performed based on the shared anatomic feature of the skull (indicated by the asterisks) in both B-mode and CT images.

Figure 2

Representative SC and IC maps. (A) SC level maps obtained during FUS treatment at two different time points (1 s and 60 s). (B) SCD map acquired by integrating the SC level for reach pixel through the total FUS treatment time. (C) Representative SC level-time curve for the pixel identified by the cross in B. (D–F) Corresponding IC level maps, ICD map, and IC level-time curve.

Figure 3

Representative PET images of a mouse in the coronal, sagittal, and horizontal planes.

Figure 4

PCI/B-mode images (left column), PET/CT images (middle column), and PCI and PET overlaid images (right column) for three representative cases shown in A, B, and C, respectively.

Figure 5

Quantitative analysis of the spatial and dose correlation between SCD maps and ⁶⁴Cu-AuNC concentration maps. (B) The offsets between the pixel with maximum SCD and the pixel with the maximum ⁶⁴Cu-AuNC concentration (indicated by crosses in A). (C) A pixel-by-pixel correlation between SCD maps and ⁶⁴Cu-AuNC concentration maps within the 2D region in the brain (indicated by the rectangular box in A). Error bar in (B) represents standard deviation. Shaded blue region in (C) represents the 95% confidence level.

Figure 6

(A) The ex vivo mouse brains were sliced coronally into 2-mm slices, and the slice containing the targeted brainstem was cut into two halves for quantifying the radioactivity in each half (illustrated by the rectangle boxes). The SCD was also averaged within each half of the brain for identifying the correlation between SCD and radioactivity. (B) The radioactivity of ⁶⁴Cu-AuNCs quantified using gamma counting and (C) the gold concentration of ⁶⁴Cu-AuNCs quantified using ICP-MS in the FUS-treated halves and contralateral non-treated halves were well correlated with the corresponding spatial-averaged SCD within the same brain regions using the segmented linear regression (R² = 0.62 for gamma counting and R² = 0.53 for ICP-MS).