PICTORIAL ESSAY Collateral Pathways in Thoracic Central Venous Obstruction Three-Dimensional Display Using Direct Spiral Computed Tomography Venography Hyo-Cheol Kim, MD, Jin Wook Chung, MD, Chang Jin Yoon, MD, Whal Lee, MD, Hwan Jun Jae, MD, Young Il Kim, MD, and Jae Hyung Park, MD Abstract: We illustrate the spectrum of venous collateral pathways caused by the thoracic central venous obstruction with direct spiral computed tomography (CT) venography and 3-dimensional recon- struction images. Venous structures that can be demonstrated with CT venography include the jugular veins; the subclavian and brachioce- phalic (innominate) veins; the internal and lateral thoracic veins; the superior and inferior venae cavae; the pericardiophrenic veins; the azygos, hemiazygos, and accessory hemiazygos veins; and the inter- costal veins. Key Words: angiography, computed tomography (CT), obstruction, stenosis, 3-dimensional, veins, venae cavae (J Comput Assist Tomogr 2004;28:24–33) O bstruction or stenosis of the superior vena cava (SVC) and its tributaries is a complication of malignant and be- nign diseases and results in the formation of collateral veins. 1,2 Although the extent and location of collateral veins are ex- tremely variable, they are often categorized into four main col- lateral pathways, such as lateral thoracic, internal mammary, azygos, vertebral pathways. 1–5 With the advent of helical com- puted tomography (CT), CT venography can be a valuable mo- dality in identifying the exact level of venous blockage and the collateral pathways in patients with SVC obstruction. 6,7 It is less invasive than conventional venography and can also pro- vide 3-dimensional reconstruction images that enhance the de- lineation of the complex vascular anatomy and facilitate the clinician’s understanding of the thoracic central venous ob- struction. We illustrate the spectrum of venous collateral path- ways caused by the thoracic central venous obstruction with CT venography and 3-dimensional reconstruction images. CT SCANNING TECHNIQUE CT examinations were performed with Somatom Plus 4 (Siemens, Erlangen, Germany), and an MX8000 multidetec- tor-row CT scanner (Marconi Medical Systems, Cleveland, OH). Unenhanced images were first obtained from lower neck level to diaphragm level with parameters of 10-mm beam col- limation, 10-mm/s table speed, 5-mm slice thickness, and 5-mm reconstruction interval, which was used to determine the scan range of CT venography. In the Somatom Plus 4 scanner, the CT parameters for CT venography were 3-mm beam col- limation, 5–6-mm/s table speed, 3-mm slice thickness, and 2-mm reconstruction interval. Scan range was 16–18 cm from the first thoracic vertebrae level to the upper one-third level of the right atrium, and scan time was 30–32 seconds. In the MX8000 scanner, the parameters were 2.5-mm detector colli- mation, 20-mm/s table speed, 3.2-mm slice thickness, and 1.6- mm reconstruction interval. Scan range was 28–34 cm from the mandible angle level to the upper abdomen level, and scan time was 14–17 seconds. CT venography was accomplished during a single breath-hold in all patients. We prepared 200 mL of diluted contrast media by mix- ing nonionic iopromide (Ultravist 370; Schering, Berlin, Ger- many) and sterile isotonic 0.9% saline solution in a ratio of 1:2. The diluted contrast material was simultaneously injected into both upper extremities or only into a diseased upper extremity with a mechanical injector. We used ipsilateral (unilateral) in- jection when the suspected level of venous obstruction was unilateral subclavian or brachiocephalic (innominate) vein. The injection rate was 2 mL/s in each extremity in case of bi- lateral injection and was 3 mL/s in cases of unilateral injection. The total amount of diluted contrast media was 200 mL in bi- lateral injection and 180 mL in unilateral injection. Therefore, the injection duration was 50 seconds in bilateral injection and 60 seconds in unilateral injection. Scan delay time ranged from 20 to 30 seconds for the single detector scanner and from 30 to 45 seconds for the four-detector scanner because it was set as long as possible for sufficient enhancement of collateral ves- sels and central thoracic veins and was decided by subtracting scan time from the injection duration. From the Department of Radiology, Seoul National University College of Medicine, Institute of Radiation Medicine, Seoul National University Medical Research Center, Seoul; and Clinical Research Institute, Seoul National University Hospital, Seoul, Korea. Reprints: Jin Wook Chung, MD, Department of Radiology, Seoul National University Hospital, # 28 Yongon-dong, Chongno-gu, Seoul, 110-744, Korea (e-mail: chungjw@radcom.snu.ac.kr). Copyright © 2004 by Lippincott Williams & Wilkins 24 J Comput Assist Tomogr • Volume 28, Number 1, January/February 2004