Purpose: To determine the accuracy and optimal threshold values of duplex u
ltrasonography (US) in assessing restenosis of renal artery stents.
Methods: Twenty-four consecutive patients with 33 renal arteries that had p
reviously been treated with placement of a Palmaz stent underwent duplex US
prior to intraarterial digital subtraction angiography (DSA), which was th
e reference standard. Diagnostic accuracy of in-stent peak systolic velocit
y (PSV) and reno-aortic ratio (RAR = PSV renal stent/PSV aorta) in detectin
g > 50% in-stent restenosis were evaluated by the receiver operating charac
teristic curve. Sensitivity and specificity were determined using the optim
al threshold values, and using published threshold values: RAR > 3.5 and in
-stent PSV > 180 cm/sec.
Results: Six examinations were technically inadequate. Nine stents had resi
dual or restenosis > 50% at DSA. The two duplex parameters were equally acc
urate since areas under the curves were similar (0.943). With optimal thres
hold values of 226 cm/sec for PSV and 2.7 for RAR, sensitivities and specif
icities were 100% and 90%, and 100% and 84%, respectively. Using the publis
hed duplex criteria resulted in sensitivities and specificities of 100% and
74% for PSV, and 50% and 89% for RAR.
Conclusion: Duplex US is a sensitive modality for detecting in-stent resten
osis if laboratory-specific threshold values are used.