Non-diagnostic lung (V/Q) scans: further investigation with venous ultrasound (US) or helical computed tomography (CT)?

Full text: The approach to the diagnosis of thromboembolism in patients (pts) with non-diagnostic V/Q scans and suspected acute pulmonary embolism (PE) is controversial. It has been suggested that leg venous US screening is of low yield and that alternative tests such as chest helical CT may be preferable. Seven hundred and thirty-seven pts (445 women, 292 men) with mean age = 57 ± 18 years underwent V/Q scans over 12 months for suspected PE. V/Q scans were reported by revised PIOPED criteria with pt risk factors used to determine post-test likelihood of PE. There were 165 high-probability (22%), 277 non-diagnostic (38%) and 295 low-probability or normal (40%) V/Q scans. These pts also underwent 219 leg venous US and 51 chest helical CT scans (US in 33 normal or low-probability, 109 non-diagnostic and 77 high-probability; CT in 6 low, 40 non-diagnostic and 5 high). Venous US and helical CT were both performed in 23 pts, with 15 concordantly negative, 3 concordantly positive and 5 discordant results (+ CT, - US in 3; - CT, + US in 1; non-diagnostic CT, + US in 1). In summary: (1) the yield of venous US in non-diagnostic V/Q scans (30%) is higher than previously reported; (2) the likelihood of + US or CT is related to post-test likelihood of PE on V/Q scan; and (3) there is no significant difference in the yield of venous US vs CT in pts with non-diagnostic V/Q scans, thus suggesting that US would be a more cost-effective tool in this regard