Pulsed Doppler frequency and carotid stenosis

Two hundred and forty-two internal carotid arteries (ICA) were evaluated by independently interpreted arteriography and pulsed Doppler spectrum analysis using ultrasonic arteriography to evaluate the ability of peak systolic frequency (PSF) to predict the degree of internal carotid stenosis. Mean PSF in the 129 (53.3%) high grade ICA stenoses of greater than 50% diameter reduction was 6.55 +/- 0.14 (SEM) khz, while mean PSF in the 113 (46.7%) low grade (less than 50% diameter reduction) stenoses was 3.38 +/- 0.12 (SEM) kHz (P less than 0.0001). Receiver-operator characteristic (ROC) analysis revealed that PSFs of 4.5 kHz (sensitivity 87%, specificity 88%) and 5.0 kHz (sensitivity 83%, specificity 93%) were best for identifying a 50% diameter stenosis. Positive predictive value of 5.0 kHz was 93% (107/115) and negative predictive value was 82.7% (105/127). Linear regression analysis of PSF in kHz versus percentage diameter reduction yielded the equation: % stenosis = 10.7 (PSF) - 4.1 (r = 0.76). A nonlinear equation was also derived: % stenosis = 61.9 - 33.5 (PSF) + 8.7 (PSF)2 - 0.5 (PSF)3 (r = 0.77). Based on this analysis peak systolic frequency criteria measured by pulsed Doppler spectrum analysis appear to be useful for distinguishing high grade from low grade stenoses. Both the linear and nonlinear equations further suggest that PSF can more precisely quantitate the degree of ICA luminal narrowing.