Color Doppler sonographic imaging in the assessment of upper-extremity deep venous thrombosis.

Color Doppler flow imaging with linear-array transducers was performed in 130 extremities in 91 patients with suspected upper-extremity deep venous thrombosis. Thrombi were detected in 39 studies (30%). Eighty-four patients had either imaging examinations (22 studies comprising venography in 20 patients, CT in one, and MR imaging in one) or clinical follow-up (99 studies) for correlation. In studies with imaging correlation, the sensitivity of the Doppler studies was 78% and the specificity was 92%. There were two false-negative examinations in which the proximal subclavian vein either could not be imaged or the anatomy was distorted. There was one false-positive color Doppler diagnosis of deep venous thrombosis in a case of partial obstruction of the subclavian vein caused by extrinsic compression. There were no known additional errors in color Doppler imaging in patients who had clinical follow-up. Four cases of isolated superior vena cava or proximal innominate vein obstruction were missed by color Doppler imaging. Color Doppler flow imaging is an accurate, noninvasive method for the evaluation of upper-extremity deep venous thrombosis. Other imaging tests may be required when the color Doppler study is negative and central venous thrombosis is suspected.