Leandro Von Krannichfeldt

Semester Work
Supervisors: Dr. Richard Rau

Speed-of-Sound Reconstruction Using a Curved Array Ultrasound Transducer

In medical ultrasound imaging, some pathologies such as breast cancer can still not be diagnosed in a reliable way. A promising candidate for on-site diagnosis is a novel speed-of-sound (SoS) imaging method using a spatial domain reconstruction (SDR). This method reconstructs a SoS map from ultrasound plane wave echoes in order to improve tissue differentiation [1]. In this context, it leverages the sensitivity of speed-of-sound to structural changes induced by pathologies. Until now, SoS imaging for conventional ultrasound systems has solely been implemented for a linear array transducer with a rectangular field-of-view. In this work, the field of application is expanded for a curved array transducer. This probe geometry is usually employed for pathologies requiring a more extended field of view in width. For this end, the imaging pipeline given by the SDR method is extended, with main contributions to numerical wave simulation and beamforming. Furthermore, an angle dependent displacement correction was incorporated to investigate reconstruction behaviour in polar direction. The relevant data is simulated for a curved array transducer with a multi-static imaging approach. Different simulation scenarios are used to assess pipeline functionality and SoS image reconstruction quality. The beamformed results imply patterns which are mainly due to the lower lateral resolution in the beamforming grid. In turn, this also affects the SoS reconstruction. Reconstruction results indicate an inability to reconstruct inclusions for strongly convex transducers. Moreover, a constant SoS offset arises displaying a proportionality to grid resolution. This offset can be traced back to a propagated discretization error. Ultimately, the displacement correction showed no significant effect on the reconstructions.