Determining the relationship between three-dimensional power Doppler data and true blood flow characteristics: an in-vitro flow phantom experiment.

OBJECTIVES: Three-dimensional (3D) ultrasound can be used to acquire power Doppler data which can be quantified to give an objective impression about blood flow within a tissue or organ. Proprietary software can be used to calculate three indices of vascularity: vascularization index (VI), flow index (FI) and vascularization flow index (VFI). Although these indices appear to have a predictive value in the clinical setting and can be shown to vary between different patient populations and over time within the same population, their relationship with true in-vivo blood flow characteristics has not been established. The objective was to examine the effect of flow rate, vessel number, attenuation and erythrocyte density on these indices.
METHODS: A computer-driven flow phantom was used to continuously pump a nylon particle-based blood mimic (Orgasol(trade mark)) around a closed system through three different ultrasound test tanks. These tanks were designed specifically for these experiments and contained C-Flex(trade mark) tubing, in a variety of arrangements, encased in an agar-based tissue mimic. The test tanks were insonated with a modified 3D transvaginal 4-8-MHz ultrasound transducer and 3D power Doppler data were then acquired over a graduated series of flow rates, depths and blood mimic concentrations. Regression analysis was used to determine the resulting relationships.
RESULTS: The VI increased linearly with an increase in flow rate (P < 0.05), whereas the FI increased in a cubic manner with a more rapid initial increase (P < 0.05). The VI demonstrated a similar linear increase with an increase in the erythrocyte mimic density (P < 0.05), whereas the FI increased markedly with a small change in erythrocyte mimic density and then plateaued (P < 0.01). There was a significant reduction in each index as the distance between the transducer and vessel increased (P < 0.05). Patterns similar to those seen in relation to the change in flow rate were evident, with a more linear relationship between depth and the VI and VFI than between depth and the FI, although the FI remained relatively constant and was not significantly affected by distance from the transducer until a depth of 55 mm was reached. Although a positive linear relationship was seen between vessel number and VI and VFI (P < 0.05) the FI demonstrated a very different and complex, cubic relationship (P < 0.001), increasing linearly until a maximum of three vessels were present when it decreased, and no overall correlation was seen (P > 0.05).
CONCLUSIONS: The VI, FI and VFI are all significantly affected by volume flow, attenuation, vessel number and erythrocyte density, but in different ways. The VI and VFI seem to have a more predictable relationship, whereas the FI often demonstrates a more complex cubic relationship that is not always logical. Further work is required to establish the effect of other confounding parameters before valid conclusions may be made and a better understanding of 3D power Doppler ultrasound imaging achieved.