To examine in vitro whether an assessment of flow in normal and obstructed vessels is essentially possible using modern multislice CT-scanners. An experimental model allowed known stenoses to be perfused at defined flow rates. Aorta and coronary arteries were simulated by silicone tubes. A pulsatile pump was used to perfuse water through the system with intermittent injection of a bolus of radio-opaque contrast agent. CT-measurements were carried out with slice orientation perpendicular to the tubes. 50-90% concentric stenoses were examined 5 times at 4 different stenosis slice distances. A mathematical algorithm calculated the temporal density changes within a ROI in the tube cross-sections. Quantitative assessment of the data simultaneously acquired with the 16-slice system for the "coronary" and "aortal" time-density curves showed that the model allowed for exclusion of a ≥ 80% stenosis grade with a 99% probability when the slopes of the density increase quotient was > 0.79; a stenosis grade of ≥ 90% could be excluded when the slopes of the density increase quotient was > 0.52. A Quotient > 0.94 for "peak density" was associated with a 99% probability of a stenosis grade ≥ 70%. The 64-slice system allowed stenosis grades of ≥ 80% to be discriminated from lower grades. The general feasibility of the in vitro approach was verified in an in vivo model. The spatial, contrast and temporal resolution of CT scanners with at least 16 detector rows enables qualitative and semiquantitative assessment of stenotic changes in flow.