BACKGROUND Qualitative change in carotid plaques was prospectively evaluated by Gray Scale Median (GSM) analysis at repeated examinations "in vivo", in relation to quantitative change in carotid arterial geometry, as assessed by high-resolution magnetic resonance imaging (HR-MRI). PATIENTS AND METHODS Duplex ultrasound with GSM analysis and HR-MRI at the carotid level were performed at baseline and 1- and 2-year follow up in 30 patients with < 70% carotid stenosis. Changes in GSM values (ΔGSM) were evaluated as the intra-individual difference between 2-year and baseline values. HR-MRI studies were evaluated for lumen area (LA), total vessel area (TVA), vessel wall area (VWA = TVA-LA) and normalized wall index (NWI = VWA/TVA). RESULTS ΔGSM value distribution was divided into quartiles. Predominantly echolucent plaques with ΔGSM value in the lowest quartile (ΔGSM ≤- 8) showed a significantly greater mean 2-year LA (28.62 ± 10.9 mm2 vs. 17.88 ± 4.8 mm2, p = 0.04) and a greater mean 2-year TVA (83.64 ± 19.4 mm2 vs. 63.26 ± 9.2 mm2, p = 0.02) than predominantly echogenic plaques with ΔGSM value in the highest quartile (ΔGSM ≥8). CONCLUSIONS Increasing echolucency during the 2-year follow up was associated with a 2-year lower degree of stenosis and higher tendency toward lumen preservation. By corroborating that plaque vulnerability is highly independent of stenosis severity, our study provided a possible new combined "in vivo" noninvasive approach for the assessment of carotid plaque vulnerability.