In vitro dual perfusion of human placental lobules as a flow phantom to investigate the relationship between fetoplacental flow and quantitative 3D power doppler angiography.

Flow phantoms have been used to investigate and quantify three-dimensional power Doppler data but this is the first study to use the in vitro, dual perfused, placental perfusion model. We used this model to investigate and quantify the effect of variation in fetal-side flow rates and attenuation on 3D power Doppler angiography. Perfusion of a placental lobule was commenced within 30 min of delivery and experimentation was successful in 8 of the 18 placenta obtained. Fetal and maternal perfusate was modified Earle's bicarbonate buffer which, following equilibration, was supplemented on the fetal side with whole heparinised cord blood. Imaging was performed with a Voluson-i ultrasound machine. A 'vascular biopsy' the thickness of the placental lobule was defined and signal quantified within using VOCAL (GE Medical Systems, Zipf, Austria). Three vascular indices are generated: vascularisation index (VI) defined as the percentage of power Doppler data within a volume of interest; flow index (FI), the mean signal intensity of the power Doppler information; and vascularisation flow index (VFI), a combination of both factors derived through their multiplication. Attenuation was investigated in this model with the addition of tissue mimic blocks. Our results showed a predictable relationship between flow rates and the vascular indices VI and VFI. However the FI was a less reliable predictor of flow; thus it should be interpreted with caution. The power Doppler signal was markedly affected by attenuation leading to a complete loss of information at a depth of 6 cm in the model used. In conclusion this model can be adapted to provide a phantom to analyse and quantify 3D power Doppler signals and demonstrates that vascular indices within a tissue remain related to volume flow. This model provides further evidence that depth dependent attenuation of signal needs to be accounted for in any in vivo work where the probe is not in direct contact with the tissue of interest.