Multivariate Analysis of 18F-DMFP PET Data to Assist the Diagnosis of Parkinsonism

• DOI: 10.3389/fninf.2017.00023
• Publication Date: 2017-03-29T23:39:38Z
• Authors: Fermín Segovia; Juan M. Górriz; Javier Ramírez; Francisco J. Mart...; Johannes Levin; Madeleine Schuberth; Matthias Brendel; Axel Rominger; Kai Bötzel; Gaëtan Garraux; Christophe Phillips
• Keywords: Binary classification

An early and differential diagnosis of parkinsonian syndromes still remains a challenge mainly due to the similarity their symptoms during onset disease. Recently, 18F-Desmethoxyfallypride (DMFP) has been suggested increase diagnostic precision as it is an effective radioligand that allows us analyze postsynaptic dopamine D2/3 receptors. Nevertheless, analysis these data poorly covered its use limited. In order address this challenge, paper shows novel model automatically distinguish idiopathic parkinsonism from non-idiopathic variants using DMFP data. The proposed method based on multiple kernel support vector machine uses linear version classifier identify some regions interest: olfactory bulb, thalamus supplementary motor area. We evaluated for both, binary separation multigroup variants. These systems achieved accuracy rates higher than 70%, outperforming DaTSCAN neuroimages purpose. addition, system combined was assessed.