Evaluation of failing hemodialysis fistulas with multidetector CT angiography: Comparison of different 3D planes

Purpose: To evaluate failing hemodialysis fistula complications using 16-detector MDCTA, and to assess the accuracies of different 3D planes. Materials and methods: Thirty patients (16 men, 14 women, aged 27-79 years) were referred for hemodialysis access dysfunction. Thirty-one MDCTA exams were done prior to fistulography. For MDCTA, contrast was administered (2 mL/kg at 5 mL/s) via a peripheral vein in the contralateral arm. Axial MIP, coronal MIP, and VRT images were constructed. Venous complications were evaluated on axial source images, on each 3D plane, and on all-planes together. Results were analyzed using McNemar test. Results: Axial MIP, VRT and all-planes evaluations were most sensitive for fistula site detection (93%). Coronal MIP had the highest sensitivity, specificity and accuracy (35%, 96%, and 85%, respectively) for detecting venous stenosis. VRT and all-planes had the highest sensitivity and accuracy for detecting aneurysms (100%). All-planes and axial MIP were most sensitive for detecting venous occlusion (61% and 54%). Comparisons of detection frequencies for each venous pathology between the five categories of MDCTA revealed no significant differences (P > 0.05). MDCTA additionally showed 3 partially thrombosed aneurysms, 4 anastomosis site stenosis and 12 arterial complications. Conclusion: MDCTA overall gives low sensitivity for detection of central vein stenosis and moderate sensitivity for occlusion. For most pathology, all-planes evaluation of MDCTA gives highest sensitivity and accuracy rates when compared to other planes. For venous stenosis and occlusion, MDCTA should be considered when ultrasonography and fistulography are inconclusive. MDCTA is helpful in identifying aneurysms, collaterals, partial venous thromboses and additional arterial, anastomosis site pathologies