[Multislice computed tomographic angiography in the assessment of central veins for endovascular treatment planning: comparison with phlebography].

The dysfunction of a vascular access for hemodialysis and its loss may depend on drainage difficulties of the superficial or deep venation due to hemodynamically significant stenosis or obstruction of a central vein, which generally involve the innominate-subclavian veins or superior vena cava. These alterations are often neglected due to their central and deep location; when there is hemodynamic compensation, they may remain asymptomatic. For these reasons every suspect clinical sign for central vein stenosis (gross arm syndrome or venous hypertension in an arteriovenous fistula) must not be ignored, as timely intervention is essential for functional recovery of the vessel and for the protection of the arteriovenous fistula. The modern imaging techniques ensure thorough diagnostic assessment, while the possibilities of endovascular treatment with interventional radiology allow, in a large proportion of cases, optimal minimally invasive treatment, but above all the recovery of venation in a hemodialyzed patient. We report our experience with multislice computed tomographic angiography (MS-CTA) and reconstruction software for treatment planning of central vein stenosis or obstruction. Forty-nine patients were studied with MS-CTA (GE 16). Images were acquired in the venous phase (120-180 seconds after contrast medium injection) followed by digital vascular reconstruction (AutoBone for bone removal, vessel analysis for caliber and length measurements, thin and curved MIP, MPR). Within a week control phlebography was performed. The venous tree was divided into seven segments and analyzed in a double-blind fashion with a distinction between patent segments, 50-70% stenosis, >70% stenosis, occlusion, and collateral vascular beds. There was excellent correspondence in all the examined segments for patency, >70% stenosis, and occlusion, with high sensitivity (98%), specificity (99.3%), and diagnostic accuracy (99.1%). The binomial test demonstrated a highly significant concordance (alpha=0.99) for all patients and in all vascular segments with the exception of 70% stenoses, in which MS-CTA gave a slight overestimate. In the central venous district, color Doppler ultrasonography may not be as effective as for the peripheral study of arteriovenous fistulae, and second-level imaging techniques such as MS-CTA are more useful. We suggest that endovascular treatment must be preceded by MS-CTA. This examination shows the lesions that may benefit from endovascular treatment and recognizes ''uncrossable'' lesions, ie, the ones that will not benefit from treatment. Moreover, it allows accurate planning of endovascular treatment by showing the lesion type (stenosis or obstruction), the position and extension of the involved vessels, the vessel caliber above and below the lesion, and the possible presence of a collateral vascular bed. MS-CTA with dedicated reconstruction software, if correctly performed and accurately reconstructed, is a precious tool for diagnosis and treatment planning.