Phantom Development

To prepare for the validation of the physical system once built, Imagivo focused on engineering and building tissue phantoms that mimic the optical properties of head and neck tissue. The aim of developing tissue optical phantoms is to mimic the optical properties of human tissue in a stable object that can be repeatedly scanned by an imaging system. This is vital for stability and reproducibility testing, optimization, and calibration which allows for interlaboratory comparison and standardization across multiple units. For this design project, the phantoms created will also directly replicate the Monte Carlo simulation tests (discussed in the prior section) to compare the real system to the ideal imaging environment. Optical properties include scattering and absorbing, which together inform about the amount of light that is transmitted through the material. This phenomenon is described by a derivation of Beer-Lambert’s Law shown in Equation 1 below,

I=I_o * e^-(μ_t * d) (1)

where I is the power of light that is transmitted through the material, Io is the power of the light source, d is the (assumed) path distance through the sample, and μ_t is the total attenuation coefficient. The total attenuation coefficient is further defined as the sum of μ_a and μ_s, the absorption coefficient and scattering coefficient respectively. According to the measurements taken by researchers at UMCG, the ideal values for head and neck tissue are μa: 0.2, μs: 9, and anisotropy coefficient (g): 0.6. To best mimic these coefficients, optical phantom consists of three elements: the main matrix, scatterers, and absorbers. There are a variety of substances that can be used for each of these elements that can be mixed-and-matched to mimic the optical phantoms of a specific tissue type. In Imagivo’s preliminary phantom development efforts, when determining the desired concentration of scatterers and absorbers, agar matrix compositions were used because they set quickly (meaning they could be made and characterized in the same day) and mechanical property longevity was not needed.

Using the work of Ruiz et al., we made a CAD model of their concentration and depth sensitivity calibration phantoms [30]. It’s a 5 by 5 by 2 cm cube with 9 wells that will each hold a solo phantom (Fig. 1). Using this mold, we were able to create an epoxy and ICG fluorescence phantom that was imaged by the Pearl (Fig. 2). The benefit of having a 9 well mold is that we’re able to look at the signal brightness in relation to the other wells, and can eventually create an absolute depth map with colors that are consistent across all images taken.