Objective: To test the hypothesis that accesible and reproducible measures of atrophy may have high clinical utility for the study of patients with the behavioral variant of frontotemporal dementia (bvFTD), we determined the diagnostic and prognostic value of six previously-validated visual atrophy scales (VAS) and the Magnetic Resonance Parkinsonism Index (MRPI).

Methods: We gathered data from 235 patients with bvFTD and 225 age- and MRI-matched healthy controls from three centers. One hundred twenty-one bvFTD participants had underlying frontotemporal lobar degeneration (bvFTD-FTLD) and nineteen did not have FTLD (bvFTD-noFTLD). Blinded clinicians applied VAS and the MRPI was calculated with a fully-automated approach. Cortical thickness and subcortical gray matter volumes were also calculated for comparison. We used linear mixed and Cox regression models to evaluate the ability of atrophy measures to predict clinical deterioration and survival.

Results: The bvFTD atrophy score (sum of atrophy scores in orbitofrontal, anterior cingulate, anterior temporal, medial temporal lobe and frontal insula regions) was usefull to discriminate bvFTD-FTLD from healthy controls (AUROC=.930; 95%CI, .903 to .957) and bvFTD-FTLD from bvFTD-noFTLD (AUROC=.880; 95%CI, .787 to .972) and had similar accuracy than cortical thickness and subcortical gray matter volumes. MRPI was increased in patients with bvFTD with underlying 4R tauopathies. The bvFTD atrophy score was an independent predictor of clinical deterioration and survival in bvFTD.

Conclusions: In bvFTD, the bvFTD atrophy score can increase the diagnostic certainty of underlying FTLD and predict disease progression. Moreover, the MRPI can identify bvFTD participants with an increased risk of underlying 4R tauopathies.