Flow Phantom for the Validation and Quantitative Analysis of Computed Tomography Perfusion

Objective: We create a flow phantom using a simple brain model with well-defined flow behavior. This allows for the validation of CT perfusion (CTP), as CTP data can be compared with theoretical calculations and gold standard flow meter values. Our approach offers a calibration system for existing CTP scanners, facilitating the selection of optimal data processing algorithms and parameter settings. Methods: The phantom was built using 37 parallel glass capillary tubes of internal diameter 1.0 mm fixed with epoxy. Saline was pumped into the inlet portion of a Y-shaped tubing system with one inlet and two outlets. One Y outlet led to the phantom, which was in series with a flow meter. The other outlet flowed to a waste container. A contrast bolus was introduced at the Y inlet and perfusion imaging was performed with a Toshiba 320 slice Aquilion One scanner. Phantom flow rates of 3, 5, 7.5 and 20 mL/min were used. Time activity curves (TACs) were calculated by averaging along various phantom segments. Results: CBV could be accurately recovered for all flow levels but 3 mL/min. MTT values were slightly higher than expected. Compared to TACs at the base of the Y, TACS just prior to the phantom showed significant dispersion. MTT measured across the phantom for length of 4.1 cm (i.e. internal volume 1 mL) with a flow meter rate of 20 mL/min was 0.066 min, whereas 0.05 min was expected. However, the same measurement for a flow of 3 mL/min had an MTT of 0.32 min, close to the expected value of 0.33 min. Conclusions: We have developed a robust system for the analysis of CTP. CBV was accurate provided sufficient scan duration. MTT values were mildly exaggerated due to dispersion of the contrast bolus. This is analogous to measuring an early arterial TAC, such as the carotid TAC, when tissue segments are fed by smaller branches. This dispersion causes a lower bound for MTT, an effect that is greater at higher flows. Our system has the potential for the calibration and optimization of existing commercial CTP devices.