Journal Article

MRI data-driven algorithm for the diagnosis of behavioural variant frontotemporal dementia

Masellis, Mario, Knopman, David, Arriba, María de, Neason, Mollie, Fede, Giuseppe Di, Thonberg, Hakan, Cosseddu, Maura, Seeley, William W., Rossor, Martin, Moore, Katrina M, Flanagan, Toby, Prix, Catharina, Peakman, Georgia, Laforce Jr, Robert, Antonell, Anna, Archetti, Silvana, Binetti, Giuliano, Galimberti, Daniela, Mitchell, Sara, Ferrari, Camilla, Finger, Elizabeth, Duro, Diana, Warren, Jason, Levin, Johannes, Fox, Nick, Tagliavini, Fabrizio, Gabilondo, Alazne, Balasa, Mircea, Rosa-Neto, Pedro, Otto, Markus, Black, Sandra, Benussi, Alberto, Meeter, Lieke, Bras, Jose, Vandenbulcke, Mathieu, Sanchez-Valle, Raquel, Ferreira, Carlos, Mead, Simon, Shafei, Rachelle, Pievani, Michela, Van Swieten, John Cornelis, Taipa, Ricardo, Maruta, Carolina, Tábuas-Pereira, Miguel, Vandenberghe, Rik, Lladó, Albert, Rogaeva, Ekaterina, Anderl-Straub, Sarah, Santana, Isabel, Ourselin, Sebastien, Woollacott, Ione, Arighi, Andrea, Greaves, Caroline, Boeve, Bradley F., Synofzik, Matthis, Jelic, Vesna, Santiago, Beatriz, Schönecker, Sonja, Boxer, Adam L., Piaceri, Irene, Freedman, Morris, Nicholas, Jennifer, Fenoglio, Chiara, Scar

2021

Overview

IntroductionStructural brain imaging is paramount for the diagnosis of behavioural variant of frontotemporal dementia (bvFTD), but it has low sensitivity leading to erroneous or late diagnosis.MethodsA total of 515 subjects from two different bvFTD cohorts (training and independent validation cohorts) were used to perform voxel-wise morphometric analysis to identify regions with significant differences between bvFTD and controls. A random forest classifier was used to individually predict bvFTD from deformation-based morphometry differences in isolation and together with semantic fluency. Tenfold cross validation was used to assess the performance of the classifier within the training cohort. A second held-out cohort of genetically confirmed bvFTD cases was used for additional validation.ResultsAverage 10-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 the separate validation cohort of definite bvFTD, accuracy was 88% (81% sensitivity, 92% specificity) with MRI and 91% (79% sensitivity, 96% specificity) with added semantic fluency scores.ConclusionOur results show that structural MRI and semantic fluency can accurately predict bvFTD at the individual subject level within a completely independent validation cohort coming from a different and independent database.

• Publisher: BMJ Publishing Group Ltd,BMJ Publishing Group LTD
• Subject: Atrophy / Biomarkers / Cognitive neurology / Datasets / Dementia / Machine learning / Magnetic resonance imaging / Medical diagnosis / Medical imaging / Mutation / Neuropsychology / Parameter identification / Principal components analysis / Proteins / Quality control / Random variables