Refraction corrected specular beamforming applied to cortical bone enhances interface visibility of bone-soft tissues interfaces (2507.08497v1)

Abstract: Ultrasound imaging of the cortex of long bones may enable the measurement of the cortical thickness and the ultrasound wave speed in cortical bone tissue. However, with bone loss, the cortical porosity and the size of the pores increase, resulting in strong ultrasound diffuse scattering whose magnitude can exceed that of the specular reflection from the bone-marrow (endosteal) interface. In this study we adapt to bone a specular beamforming technique proposed to better image a needle in soft tissue. Our approach takes into account both wave refraction and specular reflection physics to enhance the contrast of bone surfaces and reduce speckle from intracortical pores. In vivo ultrasound data were acquired at the center of the human tibia in a plane normal to the bone axis of 11 young healthy volunteeers. Ex vivo ultrasound data were acquired from 16 regions of interest from the femoral diaphysis of three elderly donors (donors 66-98 y.o.) using a 2.5 MHz US transducer. A single-element trans mission synthetic aperture imaging sequence was implemented on a research ultrasound system with a 2.5MHz phased array transducer. Image reconstruction was performed: (A) a delay-and-sum (DAS) algorithm with optimized f-number, correction of refraction at the soft tissue-bone interface and subject-specific ultrasound wave speed and (B) an adaptive algorithm using Snells law of reflection. The improvement of image quality was evaluated with contrast ratios of the average intensities: CEI between the endosteal surface and the center of the cortex. In vivo, specular beamforming improved the visibility of the endosteum (CEI ) by 1 to 13 dB while maintaining the relative contrast between the outer and inner surfaces of the cortex. These results suggest that the visualization of the intra-osseous anatomy can be enhanced if Snells law and wave refraction are taken into account during image reconstruction.