OBJECTIVE This study compared sensitivity, specificity and diagnostic accuracy of Echoduplex and Magnetic Resonance Angiography (MRA) in the evaluation of carotid bifurcation stenosis. METHODS Twenty-five patients with clinical signs suggestive for cerebrovascular insufficiency (CVI) were studied with Duplex scan, MRA and Digital Subtraction angiography (DSA). Gold standard was the angiographic examination. RESULTS on equal value of sensitivity (80.5%), MRA showed 96% specificity versus 81% of Duplex scan, 89.4% diagnostic accuracy versus 80.9% for Duplex scan. As for stenosis over 31%, the value of sensitivity did not change i.e., 80.5%; MRA showed 100% specificity versus 97.1% for Duplex, 91.2% diagnostic accuracy versus 89.1% respectively. These values compared by the test for categorial analysis and correspondence analysis (p < 0.05) did not indicate any statistically significant difference. CONCLUSIONS on the basis of our experience and as shown by current literature [2, 12] we can state that both MR-angiography and Duplex scan fail in quantifying correctly carotid stenosis with consequent over- and underestimation [3, 4]. However, they can be considered effective diagnostic procedures in a screening program [12, 16]; they are accurate, safe and accepted by the population. CONCLUSIONS on the basis of the cost in planning the screening of a population at risk for CVI, Duplex scanning is still to be considered the elective procedure.