Objective: This study aimed to create a phantom, which allows reproducible computed tomography perfusion experiments, and to identify the influence of contrast bolus configuration, scan parameters, and scanner hardware on the calculation of perfusion parameters.
Methods: A discoid perfusion phantom with centrifugally directed flow was constructed. Brain parenchyma was simulated by inert polyoxymethylene spheres. Repeated measurements were performed with variations of the above-mentioned factors, and their effects on perfusion results were analyzed.
Results: Calculated flow values measured during experiments were reproducible and showed good correlation with the true flow (R = 0.995, P < 0.01). Tube voltage, injection rate of the contrast agent, the mathematical perfusion algorithm, and the hardware of the scanner hardware had a reproducible influence on calculated perfusion results.
Conclusions: In the long term, perfusion phantoms might be helpful in identifying hardware-specific and protocol-related factors of different computed tomography scanners to improve comparability of different scanners and scanning protocols.