Influence of bone microstructure on ultrasound loss through skull-mimicking digital phantoms

Abstract

Transcranial focused ultrasound applications often use simulations that require accurate acoustic properties, which can be related to computed tomography (CT) Hounsfield Units (HU). However, clinical CT is insensitive to microstructure. This study examines how bone/marrow microstructure introduces variations in acoustic property relationships to CT HUs. The insertion loss was found through skull-mimicking digital phantoms with two materials (bone/marrow) from 0 to 75% porosity. The phantoms had one of six pore diameters ranging from 0.2 mm to 1.0 mm. k-Wave simulations were computed with a continuous 230 kHz or 650 kHz uniform pressure source. The insertion loss was defined as the transmitted mean pressure relative to a water-only reference. The 230 kHz loss changed with porosity, but microstructure had little effect. However, the 650 kHz loss in both non-absorbing and absorbing simulations depended on porosity and pore diameter. Larger pore phantoms generally had higher losses than smaller pore diameter phantoms at the same porosity. In the non-absorbing phantoms, the maximum range in insertion loss was 2% to 52%, which occurred at 20% porosity. Absorption increased the loss by 8.2% on average, with the greatest increase of 13% in the smallest pore (0.2 mm) and 2.5% porosity phantom. Coherent multiple reflections from the phantom's planar interfaces influenced the loss from smaller pores, while larger pores disrupt this phase coherence. The insertion loss dependence on porosity and pore diameter shows that the attenuation-HU relationship is ill-determined at 650 kHz. This uncertainty has implications for CT-derived acoustic models, as no single attenuation corresponds to HUs with variable microstructures.