PURPOSE: To evaluate the significance of collateral veins, as seen on chest CT, in the diagnosis of superiorvena cava obstruction. MATERIAL AND METHOD: We retrospectively reviewed the records of 81 patients in whomcollateral veins were seen on chest CT. On spiral CT(n=49), Contrast material was infused via power injector, andon conventional CT(n=32), 50ml bolus infusion was followed by 50ml drip infusion. Obstruction of the SVC wasevaluated on chest CT ; if, however, evaluation of the SVC or its major tributaries was difficult, as in fivecases, the patient underwent SVC phlebography. Collateral vessels were assigned to one of ten categories. RESULT: On conventional CT, the common collaterals showing statistically significant difference between the two groupswere the jugular venous arch in the group with venous obstruction(n=6/15, 40.0%), and around the back and scapula(n=15/17, 88.2%) and paravertebral system(n=12/17, 70.6%) in the group without venous obstruction. On helical CT,the most common collaterals were around the back and scapular (n=9/14, 64.3%; n=26/35, 74.3%, respectively) andthe paravertebral system (n=9/14, 64.3%, n=22/35, 62.9%, respectively) in both groups, with or without venousobstruction. No collateral showed a statistically significant difference between the two groups. CONCLUSION: Onconventional CT, the jugular venous arch is the only collateral vessel to predict SVC obstruction ; on spiral CT,however, collateral vessels are not helpful in the diagnosis of SVC obstruction, but are a nonspecific finding.