Histotripsy is a non-invasive ablation surgery which uses high intensity acoustic pulses to stimulate a cavitation cloud and homogenize a tissue target¹. Magnetic resonance imaging (MRI) is a useful tool for assessing lesions made by this therapy. MR contrast parameters such as R2, and the apparent diffusion coefficient (ADC) are sensitive to structural properties of cellular tissue. Changes to these properties, such as homogenization by histotripsy, should induce changes in MR image contrast. In this study we estimate R2 and the ADC of histotripsy lesions made in ex vivo brain, liver, kidney, blood clot, and various red blood cell (RBC) phantoms commonly used in histotripsy studies.

A 500 kHz, electronically steered, focused transducer (256 elements, f#: 0.5, focal distance: 15 cm, focal length: 3.5 mm, focal width: 1.5mm, PRF: 10 Hz) generated histotripsy bubble clouds in in vitro samples of porcine liver, kidney, bovine liver, brain, and blood clot as well as agar gels mixed with 3, 6, and 16%/wt red blood cells (RBC’s). Lesions were made by electronically steering the focal zone through a grid of points spaced 0.7 mm apart. For each steering point, the transducer emitted a single, two-cycle acoustic pulse (~ 5 us long) with a peak negative pressure that exceeded 35 MPa. This treatment pattern was repeated until 3, 30, or 300 pulses were deposited per grid point. A total of 4 lesions for each pulse number were formed in each tissue sample.

After treatment, each sample was placed in the bore of a 7-Tesla small animal MR scanner (Agilent Technologies, Walnut Creek, CA) and imaged using spin-echo sequences with various echo-times and diffusion-weighting b-values. Contrast parameters R2, and ADC were estimated from the resulting images by selecting a region of interest (ROI) within each lesion and performing a non-linear, least-squares fit of the mean signal within the ROI.

Example R2 and diffusion-weighted (DW) images of lesions made in porcine liver and bovine brain are displayed in Fig 1. The R2 and ADC parameters measured in each tissue type are plotted as a function of pulse number in Fig 2. In liver, kidney, blood clot, and the red blood cell phantoms, homogenization induced measurable decreases in the R2 rate. However, R2 did not change appreciably in brain with treatment. In the red blood cell phantoms, the magnitude of change of the R2 rate decreases with decreasing RBC concentration. The ADC for all samples increased measurably with treatment.

For most samples, the R2 relaxation rate and the ADC changed appreciably with increasing pulse numbers. Both contrast parameters asymptotically approach a final value such that further pulses cause marginal changes in the contrast parameters. R2 weighted imaging may be a good indicator of homogenization in samples with high iron content such as liver and RBC phantoms. However, for samples with low iron content such as brain, R2 changes little with increased pulse number.

These results match well with theory by Brooks et al.², Gillis et al.³, and Ye et al.⁴: R2 relaxation enhancement occurs when water is able to drift through magnetic field gradients induced by aggregations of paramagnetic materials such as hemoglobin or ferritin packed into a cell. Homogenization by histotripsy breaks up these aggregations and thus reduces the field gradients they impose. This effectively removes the enhancement effect and R2 appears to decrease. Because there are very few iron products in brain tissue, these field perturbations do not exist and homogenization does not result in a more uniform magnetic field.

The ADC appears to change with pulse number in all materials reported here, likely because homogenization removes cellular membranes. These results suggest that diffusion-weighted imaging is a good assessment tool for histotripsy therapy. However, R2-weighted imaging may suffice for histotripsy therapy in the body, where tissues contain more iron content and respiratory motion makes diffusion-weighted imaging difficult.