MRgHIFU safety issue: Validation of targeting accuracy using an MR compatible ballistic model

Introduction Volumetric ablation has been investigated in vivo demonstrating that a large and uniform ablation zone can be obtained rapidly (1-2). However, there exists no experimental demonstration to show that such volumetric lesion is indeed centered on a specific predefined target in 3D, i.e. no “gold standard” proof of absence of thermal buildup drift during the sonication (3). The effective spatial control of the induced thermal lesion during fast volumetric ablation should be considered a major safety issue. Therefore before clinical trials, it is highly desirable to evaluate in animal models, the accuracy of the spatial control of ablation for a given volumetric HIFU sonication paradigm. We describe here a method to create a user-defined ballistic target to be used as an absolute reference marker that is both MR-compatible and MR-detectable, while also being a well-established histology staining method. Material and Methods A quality assurance method for pre-clinical studies is described here, based on the trapping of a mixture of MR contrast agent (Gadolinium) and histology stain (methylene blue) using localized energy deposition causing cell death and coagulation. As a result, the contrast agent/stain is encapsulated in the intracellular space. The marker fixation paradigm consisted of three steps: 1) RF coagulation (Fig1a): a dedicated RF electrode was designed to operate in monopolar mode using a clinical radio opaque intravenous (i.v.)-dedicated catheter as an active electrical source and a square 6cm copper sheet in contact with the skin (via saline serum) as ground contact (RF clinical generator, Celon, Germany)), 2) injection via the needle of 0.2mL mixture of methylene blue (Patent Blue V Sodium 2.5%, Guerbet SA, France) doped with 0.75% gadolinium-DTPA (Dotarem, Guerbet, France), and 3) a second RF coagulation, identical to the first. This ballistic model was used here as a mimicking target for a “virtual” tumor-center to assess the spatial control of HIFU ablation in vivo on four New Zealand rabbit thighs (female, weight=3.5±0.5 kg). A high resolution T1w 3D gradient-echo (VIBE) acquisition (0.8x0.8x0.8mm, TE/TR/TA/FA/ BW =1.6ms, 4ms, 2.55min, 10°, 650Hz/Px, 50 sec scan time) was used to predefine the position of the ballistic target in the rabbit leg, far from bone/fascia. Fig1. (a) Macroscopic RF thermal lesion appearing as a white spheroid area of cooked tissue ex vivo in the pink meat sample background. (b) Macroscopic thermal lesion obtained by MRgHIFU (large white spheroid area of cooked tissue in the pink ex-vivo meat sample) centered on the ballistic target (high density blue spot). Mild blue density resulted from contrast agent dispersion while slicing the meat to cross the central plane of the lesion.