Objectives: Perfusion computed tomography (PCT) is increasingly getting popular with the advent of computed tomography (CT) systems with adequate temporal resolution and spatial coverage. We sought to develop a biological phantom for perfusion measurements in CT to design, improve, and validate scan protocols and postprocessing algorithms in vitro.
Materials and methods: A special technique was applied to prepare and preserve a fresh porcine kidney. The kidney was connected to an open circuit driven by a peristaltic pump with the option to inject contrast material. We evaluated repeated dynamic contrast-enhanced CT acquisitions with different input flow rates and the relation to calculated parenchymal flow results of the phantom. Flow was calculated with 2 different algorithms. Identical scans were performed with a time interval of 1 year to check long-term stability of the phantom. Different bolus geometries were designed and bolus dispersion was measured for the setup using a tubing array.
Results: We found a linear relationship between the input flow rate of the circuit and the calculated phantom tissue flow with a correlation coefficient rr2 = 0.99 for both algorithms. Both algorithms resulted in very similar absolute values, the mean difference was 3.1 mL/100 mL/min. Perfusion measurements with contrast material injection and storage did not alter the phantom. The enhancement properties did not change over the time of 1 year. With our setup, it was possible to design typical bolus geometries as they occur in clinical practice. Bolus dispersion was small: peak enhancement and bolus width changed by about only 5% over 2-m tube length.
Conclusions: A phantom for parenchymal flow measurements suitable for repeated measurements over a long period of time was developed. The setup allows the design of diverse bolus geometries with negligible dispersion.