Abstract #3313

Automatic assessment of corpus callosum malformation from structural MRI images to improve diagnosis reproducibility.

Denis Peruzzo¹, Umberto Castellani², Fabio Triulzi^1,3, Andrea Righini⁴, Cecilia Parazzini⁴, and Filippo Arrigoni¹

¹Neuroimaging Unit, Scientific Institute IRCCS “Eugenio Medea”, Bosisio Parini, Italy, ²Department of Computer Science, University of Verona, Verona, Italy, ³Department of Neuroradiolody, Fondazione IRCCS “Ca’ Granda” Ospedale Maggiore Policlinico, Milano, Italy, ⁴Department of Pediatric Radiology and Neuroradiology, Children Hospital “Vittorio Buzzi”, Milano, Italy

The diagnosis of brain malformations is usually based on the visual inspection of MRI images by trained neuroradiologists. The resulting procedure is therefore subjective and mainly provides a qualitative description of the detected malformations. In this study, we propose an assisted diagnosis tool (ADT) for the analysis of the corpus callosum from structural T1-weighted images. The method detects and characterizes different kind of malformations (local/diffuse, homogeneous/heterogeneous). Inter-subject reproducibility experiments showed that the agreement rate significantly improved from 67.5% to 79.3% using the proposed method.

This abstract and the presentation materials are available to members only; a login is required.

Join Here