Multimodal Lesion Phenotyping Improves Seizure Outcome Classification after Traumatic Brain Injury: An EpiBioS4Rx Study
Rachael Garner1, Alexis Bennett1, Akul Sharma1, Michael Douglas Morris2, Marianna La Rocca1,3, Giuseppe Barisano1, Ruskin Cua4, Paul Vespa2, Arthur W Toga1, and Dominique Duncan1
1USC Stevens Neuroimaging and Informatics Institute, Keck School of Medicine, University of Southern California, Los Angeles, CA, United States, 2David Geffen School of Medicine, University of California, Los Angeles, Los Angeles, CA, United States, 3Dipartimento Interateneo di Fisica, Università di Bari, Bari, Italy, 4USC Department of Radiology, Keck School of Medicine of USC, University of Southern California, Los Angeles, CA, United States
Posttraumatic epilepsy (PTE), or recurrent seizures after traumatic brain injury (TBI), is a debilitating complication of TBI. We present a multimodal approach to classify seizure outcomes using computed tomography (CT) and magnetic resonance imaging (MRI) features that characterize lesion phenotypes. Five logistic regression models to predict seizure outcome are presented, using patient demographics and clinical information in conjunction with CT and MRI variables that describe lesion characteristics including contusion type and location as well as whole-brain lesion volumetrics. The optimal model utilized all four categories of features, yielding 91.4% sensitivity, 75% specificity, and 0.886 area under the curve performance.
This abstract and the presentation materials are available to members only;
a login is required.