3D QUANTIFICATION OF MICRO‐CT ANGIOGRAMS IN MURINE HIND LIMB ISCHEMIA: STATISTICAL ROBUSTNESS AND NOVEL BIOLOGICAL INSIGHTS

Abstract

OBJECTIVE. Vascularity on micro‐CT angiograms is generally reported as the sum of vessel diameters assessed on serial CT slices. This over‐estimates the area of vessels not perpendicular to the Z‐axis (Fig. 1) thus we developed a 3D volume based measurement.METHODS. Hind limb ischemia was induced in 22 mice and post‐mortem angiography with bismuth/gelatine was performed after 1 (n = 6), 4 (n = 4), 7 (n = 4), 14 (n = 3) and 21 (n = 5) days. Scans were at 28 μ m resolution. After sub‐volume selection (Fig. 2) and thresholding at 800, 1000, and 1200 grey intensity levels (GIL), voxel based vessel volume was computed. Group mean volume in the non‐ischemic leg minus the volume in the ischemic leg was compared by t‐distribution and F test for paired comparisons.RESULTS. Statistical significance was not affected by threshold (800 GIL P = 0.012; 1000 GIL P = 0.019; 1200 GIL P = 0.023). After day 4 vessel volume in the ischemic leg was significantly greater than in the non‐ischemic leg irrespective of threshold. Volume in the non‐ischemic leg initially decreased (Fig. 3).CONCLUSIONS. Volumetric quantification of vessel re‐growth is statistically robust and not influenced by threshold. This method does not suffer from overestimation of vessel area on CT slices and is likely to be more accurate. The temporary decrease in vessel volume in the non‐ischemic leg is best explained by diameter changes in response to altered flow and pressure.