Midbrain atrophy in progressive supranuclear palsy: comparison of two dimensional planimetric- with three dimensional volumetric measurements

Abstract

Two-dimensional (2D) planimetric measurements of the brainstem could be useful to differentiate patients with the clinical diagnosis of progressive supranuclear palsy (PSP) from those with other parkinsonian disorders intra vitam (1). However, planimetric 2D measurements of small structures may be inaccurate owing to partial volume effects, which depend on the rater-dependent selection of the midsagittal image and the manual traces to identify the structure. We aim here to compare the use of 2D with three-dimensional (3D) volumetric measurements. On a semiautomated basis, we measured 33 patients diagnosed with PSP (Eight confirmed at autopsy)(2). All measured brain regions were normalized by the total intracranial volume (TICV). Although the normalized midbrain volume was the most powerful predictor to discriminate between PSP patients and control subjects, there was some overlap between these two groups (Figure B). We found that the midbrain atrophy in patients with PSP correlated with the Hoehn and Yahr Scale, bradykinesia, gait disturbance, and falls (Table). We applied a 2D technique (1) which consists in a manual tracing of brainstem structures on a midsaggital MRI-slice followed by an automatic planimetric measurement on the images of our patients with PSP (n=33) and corticobasal degeneration (CBD; n=18) and age-matched control subjects (n=22). To this purpose, we reconstructed the midsagittal plane using the anterior and posterior commissures as landmarks. We found this feature useful to discriminate between these groups (Figure A). Therefore, 2D measurements discriminate not only PSP patients from control subjects but also from those with CBD. In contrast to our volumetric approach, the 2D determination of the midbrain showed no significant correlation with clinical variables (Table). Thus, 2D measurement of the brainstem appears to be a useful clinical tool to rapidly discriminate PSP from control and other atypical parkinsonian syndromes, but should be questioned as a surrogate marker to evaluate disease severity or disease progression. For clinicoanatomic correlation, the 3D-volumetric approach appears to be superior.