Multi-volume rendering for three-dimensional power doppler imaging

Multi-volume rendering (MVR) has been used for better localization of tumors by integrating structural, e.g., Computed Tomography (CT), and physiological/functional 3D information, e.g., Positron Emission Tomography (PET) in other medical imaging modalities. Similar fusion in ultrasound imaging between BW-mode and Power Doppler provides improved visualization and better understanding of the correlation between anatomical structures and blood flow, e.g., heart, kidney, liver. In this paper, we present three fusion algorithms for ultrasound imaging: Composite Fusion (CF), Post Fusion (PF) and Progressive Fusion (PGF). In the CF, BW and Power Doppler volumes are fused during volume rendering while in PF both volumes are rendered separately and then fused using alpha blending. In PGF, fusion is performed at several stages and intermixing of volumes occurs both during and after rendering. CF requires the rendering of RGB volume, but in PF and PGF rendering can be performed on luminance (grayscale) values to save computation. We have investigated these three fusion techniques using in vivo data acquired from kidney and liver using a commercial ultrasound machine (i.e., EUB-6500, Hitachi Medical Corporation, Japan). PF displays correlation between tissue and blood flow but loses depth information. The CF provides depth cue information but assigns artificial color to the rendered volume since it changes the ratio between R, G and B values during compositing. PGF retains the assigned RGB color and also provides depth information but renders volume more opaque than PF and CF. Each of these methods has complementary advantages. Thus, evaluating volumes with all three methods could be useful for clinicians to perform diagnostic evaluation. We will present these algorithms along with their relative advantages and disadvantages in detail.