A prospective database included all patients who underwent lower extremity endovascular interventions between 2004 and 2009, and patients with de novo stenotic lesions in the femoropopliteal segment were selected. A study 1 analyzed duplex ultrasound (DUS) and digital subtraction angiography (DSA) data pairs ≤30 days apart in 275 lesions (200 patients). Receiver operator characteristic (ROC) analysis showed that to detect ≥70% stenosis, a peak systolic velocity (PSV) of 200 cm/s had 89.2% sensitivity and 89.7% specificity, and a velocity ratio (Vr) of 2.0 had 88.7% sensitivity and 90.2% specificity. To differentiate between <50% and ≥50% stenosis, PSV of 150 cm/s and Vr of 1.5 were highly specific and predictive. Combining PSV 200 cm/s and Vr 2.0 for ≥70% stenosis gave 79.0% sensitivity, 99.0% specificity, 99.0% positive predictive value, and 85.0% negative predictive value.
A systematic review 2 included 14 studies. Of these 7 assessed the aortoiliac segment, 12 the femoropopliteal segment, and 3 the infragenicular segment.
For the aortoiliac segment, the sensitivity for the detection of greater or equal to 50% stenosis was 80% (95% CI 61% to 93%), and specificity 95% (95% CI 91% to 98%).
For the femoropopliteal segment, the sensitivity for the detection of greater or equal to 50% stenosis was 82% (95% CI 67% to 92%), and specificity 96% (95% CI 93% to 98%).
The sensitivities and specificities for detecting total occlusion in the aortoiliac and femoropopliteal segment were even better than those for the detection of at least 50% stenosis.
For the infragenicular segment, the sensitivity for the detection of occlusion or greater or equal to 50% stenosis was 83% (95% CI 59% to 96%), and specificity 84% (95% CI 69% to 93%).
Comment: The quality of evidence is upgraded by high quality of some original studies and the review.
Primary/Secondary Keywords