Abstract
BACKGROUND AND OBJECTIVES: Progressive supranuclear palsy (PSP) is characterized by progressive atrophy of brainstem, subcortical, and frontal regions. It is unknown whether longitudinal atrophy patterns differ between PSP-Richardson syndrome (PSP-RS) and cortical and subcortical clinical variants of PSP and whether atrophy rates could be useful outcome measures for treatment trials. We aimed to determine whether regional rates of atrophy differ across PSP clinical variants and determine sample size estimates for clinical treatment trials. METHODS: We identified participants with PSP-RS, PSP-cortical, or PSP-subcortical variants recruited by the Neurodegenerative Research Group who underwent 2 annual 3T MRI scans. Volumes were calculated for 10 regions of interest, and mixed-effects regression analysis was performed to compare annualized rates of atrophy across groups, accounting for confounders. Sample sizes required to power placebo-controlled treatment trials to detect a 20% treatment effect with 80% power were calculated for rates of atrophy and clinical metrics, including the PSP Rating Scale. RESULTS: Fifty patients with PSP-RS (baseline age = 69 [interquartile range 64 to 75], 46% female), 18 with PSP-cortical (74 [70 to 76], 50% female), 20 with PSP-subcortical (71 [67 to 77], 60% female), and 32 controls (69 [66 to 72], 72% female) were included. Rates of midbrain, cerebellar dentate, and subthalamic atrophy were greater in all PSP variants than in controls. The PSP-RS group showed greater midbrain atrophy rates than PSP-subcortical (-1.6% difference [95% CI -2.4 to -0.7], p < 0.001) and PSP-cortical (-1.1% [-2 to -0.2], p = 0.02) groups. The PSP-cortical group showed greater rates of pallidum (12% [2.7 to 21], p = 0.02), putamen (2.1% [0.4 to 3.8], p = 0.02), and caudate (1.6% [0.1 to 3], p = 0.04) atrophy than the PSP-RS group. PSP-RS and PSP-cortical groups showed greater rates of superior frontal (2.2% [0 to 4.3], p = 0.049, and 4.7% [1.9 to 7.4], p = 0.002, respectively) and precentral (2.6% [0.1 to 5.1], p = 0.046, and 4.3% [1 to 7.5], p = 0.01, respectively) atrophy than controls, with no differences between variants. The smallest sample size estimates were obtained using a metric combining the PSP Rating Scale and the best-performing regional volume, providing sample size estimates of 107 participants/arm for PSP-RS, 88 for PSP-cortical, and 368 for PSP-subcortical. DISCUSSION: Regional patterns of brain atrophy differ across PSP clinical variants, although midbrain atrophy was common in all variants. Sample size estimates suggest that combining the PSP Rating Scale and targeted MRI volumes provides the optimum outcome measure for clinical treatment trials recruiting patients with PSP clinical variants.