Data-driven algorithm for the diagnosis of behavioral variant frontotemporal dementia

INTRODUCTION Brain structural imaging is paramount for the diagnosis of behavioral variant of frontotemporal dementia (bvFTD), but it has low sensitivity leading to erroneous or late diagnosis. METHODS A total of 515 subjects from two different bvFTD databases (training and validation cohorts) were included to perform voxel-wise deformation-based morphometry analysis to identify regions with significant differences between bvFTD and controls. A random forest classifier was used to individually predict bvFTD from morphometric differences in isolation and together with bedside cognitive scores. RESULTS Average ten-fold cross-validation accuracy was 89% (82% sensitivity, 93% specificity) using only MRI and 94% (89% sensitivity, 98% specificity) with the addition of semantic fluency. In a separate validation cohort of genetically confirmed bvFTD, accuracy was 88% (81% sensitivity, 92% specificity) with MRI and 91% (79% sensitivity, 96% specificity) with added cognitive scores. DISCUSSION The random forest classifier developed can accurately predict bvFTD at the individual subject level.