Phantoms containing uniform media are frequently used to mimic human tissue, but are an oversimplification for some applications. For example, heterogeneous tissue properties are important when determining the spatial profile of the radiofrequency (RF) field, of the specific absorption rate (SAR) and of tissue heating¹. Determination of SAR is closely tied with computational simulations, and, as such, there is a need to validate simulation results in close association with 3D heterogeneous phantoms. Such phantoms are now easily achieved via 3D printing technology. In this work, we propose a heterogeneous phantom created as part of investigating, monitoring, and ultimately suppressing heating effects resulting from implanted deep brain stimulators (DBS)².

The tissue compartments of interest are the basal ganglia (common DBS targets for treating Parkinson’s Disease), the surrounding sub-cortical white matter (represented by the “inner brain”), cortical grey matter (represented by the “outer brain”), and the skull. Soft tissues within this set were represented by oil-in-gelatin dispersions^3,4 of varying concentrations. Three different gelatin dispersions were created to simulate grey matter (gelatin #1), white matter (gelatin #2) and CSF (gelatin #3). Bovine skin gelatin (225 bloom) was used to create a stiff gelatin, ideal for maintaining its shape through the multi-step setting process outlined below. To vary the permittivity, various concentrations of canola oil were emulsified in the gelatin mixture prior to setting. Although canola oil affected the conductivity as well as the permittivity, the effect was not large enough at the frequency of interest (128 MHz) to produce substantial systematic errors. Thus, the desired conductivities were achieved simply by adding sodium chloride to distilled water used in the gelatin mixture. The permittivity and conductivity of each tissue-equivalent material was measured with a dielectric probe (KT-85070E, Keysight Technologies, Inc.). To represent implanted DBS leads, two long copper wires were inserted within the gelatin brain along with two temperature probes (OTG-MPK5, Opsens) located at the tips.

The shapes of each soft tissue structure were created using 3D printed moulds designed to approximate the shape of human brain (Figure 1(a)). An acrylic casing in the approximate shape of a human head was also 3D-printed (Figure 1(b)). The 3D printed moulds of the basal ganglia, white matter and grey matter and the acrylic shell were created via a digital mesh file. Thus, the physical phantom was related to a mesh-based digital model compatible with most electromagnetic simulation software. A CT scan of the human skull was obtained and converted to a mesh file using Amira (FEI Co). The mesh files were assembled with relative positions that were matched to those created in the gelatin setting method described below.

The gelatin dispersions mimicking grey and white matter were set in a concentric layered fashion to emulate a simplified version of the human brain. First, gelatin #1 was poured into moulds to create the basal ganglia. The copper wire was then inserted after gelatin #1 was set. Gelatin #2 was then filled around the basal ganglia and the wires to maintain positioning. Afterwards, the basal ganglia complex was placed in the mould for the inner brain structure which was filled with gelatin #2 to a precise depth to ensure appropriate location of the basal ganglia relative to the inner brain. The wires and temperature probes exited the mould through drill holes oriented at a 60º angle to the central axis of the mould. The mould was then filled with gelatin #2 and allowed to set. The outer brain, using gelatin #1, was set via a similar method. The acrylic shell was then filled with gelatin #3 to create a base for the skull. Gelatin #3 was poured in two additional steps to allow for accurate placement of the brain.

Oil and sodium concentrations, permittivity and conductivity for each medium are listed in Table 1. The permittivity and conductivity values approximate the desired structure to within 10% of values found in literature⁵. The assembled digital model is shown in Figure 2. The brain at two stages of development (white matter only and with the external grey matter) is shown in Figure 3. A procedure for creating a complex, heterogeneous gelatin-based phantom with corresponding 3D mesh files has been developed. This phantom will improve evaluation of heating in DBS implants over what can be deduced using a uniform phantom.