Abstract
BACKGROUND AND PURPOSE: Parkinson disease is a prevalent disease, with olfactory dysfunction recognized as an early nonmotor manifestation. It is sometimes difficult to differentiate Parkinson disease from atypical parkinsonism using conventional MR imaging and motor symptoms. It is also known that olfactory loss occurs to a lesser extent or is absent in atypical parkinsonism. To the best of our knowledge, no study has examined olfactory bulb changes to differentiate Parkinson disease from atypical parkinsonism, even in an early diagnosis, and its association with conventional MR imaging findings. Hence, we aimed to assess the utility of olfactory bulb measurements in differentiating Parkinson disease from atypical parkinsonism even in the early stage. MATERIALS AND METHODS: In this retrospective study, we enrolled 108 patients with Parkinson disease, 13 with corticobasal syndrome, 15 with multiple system atrophy, and 17 with progressive supranuclear palsy who developed parkinsonism. Thirty-nine age-matched healthy subjects served as controls. All subjects underwent conventional MR imaging and 3D FIESTA for olfactory bulb measurements using manual ROI quantification of the cross-sectional olfactory bulb area using the coronal plane. Bilateral olfactory bulb measurements were averaged. For group comparisons, we used the Welch t test, and we assessed diagnostic accuracy using receiver operating characteristic analysis. RESULTS: Patients with Parkinson disease had a mean olfactory bulb area of 4.2 (SD, 1.0 mm(2)), significantly smaller than in age-matched healthy subjects (6.6 [SD, 1.7 mm(2)], P < .001), and those with corticobasal syndrome (5.4 [SD, 1.2 mm(2)], P < .001), multiple system atrophy (6.5 [SD, 1.2 mm(2)], P < .001), and progressive supranuclear palsy (5.4 [SD, 1.2 mm(2)], P < .001). The receiver operating characteristic analysis for the olfactory bulb area measurements showed good diagnostic performance in differentiating Parkinson disease from atypical parkinsonism, with an area under the curve of 0.87, an optimal cutoff value of 5.1 mm(2), and a false-positive rate of 18%. When we compared within 2 years of symptom onset, the olfactory bulb in Parkinson disease (4.2 [SD, 1.1 mm(2)]) remained significantly smaller than in atypical parkinsonism (versus corticobasal syndrome (6.1 [SD, 0.7 mm(2)]), P < .001; multiple system atrophy (6.3 [SD, 1.4 mm(2)]), P < .001; and progressive supranuclear palsy (5.2 [1.3 mm(2)], P = .003, respectively). CONCLUSIONS: 3D FIESTA-based olfactory bulb measurement holds promise for distinguishing Parkinson disease from atypical parkinsonism, especially in the early stage.