Combining semi-quantitative rating and automated brain volumetry in MRI evaluation of patients with probable behavioural variant of fronto-temporal dementia: an added value for clinical practise?

• Published in: Neuroradiology 2023-06, Vol.65 (6), p.1025-1035
• Main Authors: Calloni, Sonia Francesca, Vezzulli, Paolo Quintiliano, Castellano, Antonella, Leone, Riccardo, Basaia, Silvia, von Loon, Almar, Spinelli, Edoardo Gioele, Magnani, Giuseppe, Caso, Francesca, Agosta, Federica, Filippi, Massimo, Falini, Andrea
• Format: Article
• Language: English
• Subjects: Atrophy; Atrophy - pathology; Automation; Brain; Brain - diagnostic imaging; Brain - pathology; Dementia; Dementia disorders; Diagnostic Neuroradiology; Evaluation; Frontotemporal dementia; Frontotemporal Dementia - diagnostic imaging; Frontotemporal Dementia - pathology; Humans; Imaging; Magnetic resonance imaging; Magnetic Resonance Imaging - methods; Medicine; Medicine & Public Health; Neuroimaging; Neurology; Neuropsychological Tests; Neuroradiology; Neurosciences; Neurosurgery; Observers; Radiology; Software
• ISSN: 0028-3940; 1432-1920
• DOI: 10.1007/s00234-023-03133-w

Purpose To evaluate the diagnostic value of combined semiquantitative and quantitative assessment of brain atrophy in the diagnostic workup of the behavioural-variant of frontotemporal dementia (bvFTD). Methods Three neuroradiologists defined brain atrophy grading and identified atrophy pattern suggestive of bvFTD on 3D-T1 brain MRI of 112 subjects using a semiquantitative rating scale (Kipps’). A quantitative atrophy assessment was performed using two different automated software (Quantib® ND and Icometrix®). A combined semiquantitative and quantitative assessment of brain atrophy was made to evaluate the improvement in brain atrophy grading to identify probable bvFTD patients. Results Observers’ performances in the diagnosis of bvFTD were very good for Observer 1 ( k value = 0.881) and 2 ( k value = 0.867), substantial for Observer 3 ( k value = 0.741). Semiquantitative atrophy grading of all the observers showed a moderate and a poor correlation with the volume values calculated by Icometrix® and by Quantib® ND, respectively. For the definition of neuroradiological signs presumptive of bvFTD, the use of Icometrix® software improved the diagnostic accuracy for Observer 1 resulting in an AUC of 0.974, and for Observer 3 resulting in a AUC of 0.971 ( p -value