Correlation Between Contrast Time-Density Time on Digital Subtraction Angiography and Flow: An in Vitro Study

Digital subtraction angiography (DSA) provides an excellent anatomic characterization of cerebral vasculature, but hemodynamic assessment is often qualitative and subjective. Various clinical algorithms have been produced to semiquantify flow from the data obtained from DSA, but few have tested them against reliable flow values.

An arched flow model was created and injected with contrast material. Seventeen injections were acquired in anterior-posterior and lateral DSA projections, and 4 injections were acquired in oblique projection. Image intensity change over the angiogram cycle of each DSA run was analyzed through a custom MATLAB code. Time-density plots obtained were divided into 3 components (time-density times, TDTs): TDT_10%-100% (time needed for contrast material to change image intensity from 10% to 100%), TDT_100%-10% (time needed for contrast material to change image intensity from 100% to 10%), and TDT_25%-25% (time needed for contrast material to change from 25% image intensity to 25%). Time-density index (TDI) was defined as model cross-sectional area to TDT ratio, and it was measured against different flow rates.

TDI_10%-100%, TDI_100%-10%, and TDI_25%-25% all correlated significantly with flow (P < 0.001). TDI_10%-100%, TDI_100%-10%, and TDI_25%-25% showed, respectively, a correlation coefficient of 0.91, 0.91, and 0.97 in the anterior-posterior DSA projections (P < 0.001). In the lateral DSA projection, TDI_100%-10% showed a weaker correlation (r = 0.57; P = 0.03). Also in the oblique DSA projection, TDIs correlated significantly with flow.

TDI on DSA correlates significantly with flow. Although in vitro studies might overlook conditions that occur in patients, this method appears to correlate with the flow and could offer a semiquantitative method to evaluate the cerebral blood flow.