Duplex Criteria To Identify Carotid Stent Patients

Back to Program

INTRODUCTION AND OBJECTIVES:
The CMS carotid stent (CAS) policy covers high-risk symptomatic patients ONLY if carotid arteriogram (AGRAM) confirms >70% stenosis. This CMS policy magnifies need for accurate carotid duplex ultrasound (DUS). We use a 3-parameter DUS algorithm (PSV, EDV, & ICA/CCA PSV ratio) to determine ICA stenosis. Each parameter has decile thresholds (e.g. 50-59%, 60-69% etc). Final stenosis estimate is range that straddles all 3 parameters (e.g. PSV & EDV in 50-59%, ICA/CCA ratio in 60-69%, diagnosis is 50-69%). Our PSV threshold for 70% stenosis is 340 cm/sec, while a 2003 national Consensus Panel recommended PSV 230 cm/sec. We tested accuracy of our DUS method and Consensus Panel recommendation.
METHODS:
Retrospective accuracy analysis of prospectively applied criteria for 70% ICA stenosis in consecutive 7-year series of patent evaluable DUS with AGRAM confirmation. Blinded observers measured AGRAMs by NASCET method. ROC curves generated by SPSS.
RESULTS:
In 203 bifurcations with DUS and AGRAM within 8 weeks, Consensus Panel PSV 230 was sensitive (96%) but not specific (85%) or accurate (89%) for diagnosing a 70% ICA stenosis. PSV 340 was less sensitive (89%) but much more specific (94%), with highest accuracy, 93%. After combining stenosis categories that spanned or exceeded 70% stenosis, our 3-parameter algorithm was highly sensitive (98%), specific (91%), and accurate (92%).
CONCLUSIONS:
The three parameter algorithm demonstrated best combination of sensitivity (98%) and specificity (91%) to identify Medicare CAS candidates. Our single parameter PSV 340 cm/sec had highest accuracy, but its 11% false negative rate, while acceptable for asymptomatic patients, may be undesirable for symptomatic patients. Consensus Panel PSV 230 cm/sec has a high false positive rate and will result in unnecessary AGRAMs in 15% of potential CAS. Best DUS analysis may involve PSV, EDV and ICA/CCA ratio. Labs should reexamine their DUS accuracy for Medicare CAS candidates.