Neurocysticercosis-related Seizures: Epilepsy Biomarkers from Advanced Quantitative Magnetic Resonance Imaging

Epilepsy is a disorder of global importance, predisposing 70 million people to unprovoked seizure activity. There are myriad treatment options available, however roughly 30% of people diagnosed with epilepsy proceed to develop drug resistance, for which surgical intervention is often the only recourse. Neurocysticercosis (NCC) is a parasitic CNS infection endemic to low- and middle-income countries that primarily presents with acute seizure activity. A disproportionally large number of people with a history of NCC-related seizure activity continue to experience recurrent seizure activity. NCC has a well-defined (if pleomorphic) presentation, yet the pathomechanisms underlying the development of an epileptogenic state in NCC are unclear. Epilepsy is recognised to be a network disorder, where abnormal neuronal organisation facilitates the transition of the brain into an ictogenic state, the correlates of which have been extensively explored using quantitative magnetic resonance imaging (qMRI) approaches. The aim of this project was to apply qMRI methodologies, comparable to those used in cutting-edge epilepsy research, to a novel NCC-related seizure cohort in order to explore cortical network properties underlying seizure presentation and recurrence. We recruited 29 people with a first recent seizure due to probable NCC—as well as 11 representative controls—from the National Institute of Mental Health and Neuro Sciences, and augmented this sample with 53 similarly matched healthy controls from the Human Connectome Project. We recorded demographic and clinical data alongside research-quality structural, diffusion, and functional MRI. We first developed a TRIPOD-compliant clinical prediction model for seizure recurrence following acute NCC-related seizure activity from routinely acquired clinical data. Following this, we compared morphometric similarity and volumetric properties of the cortex and subcortical structures between the patients and controls in our sample. For analysis of the diffusion data, we computed lesion symptom network maps and extracted white matter microstructural properties from TractSeg-defined regions with a fixel-based methodology, which we used to synthesise tract metric profiles for each patient standardised to our controls. We applied a similar technique to the resting-state functional MRI data, first characterising group differences in regional grey matter properties before mapping differences in lesion symptom network maps between clinical groups. For the final analysis, we leveraged graph-theoretical connectomics to quantify the whole-brain network dynamics within our cohort. We first present an exemplar guideline-compliant model for predicting post-acute seizure recurrence using baseline clinico-demographic features, which (whilst underpowered) indicates the importance of considering both neuroinflmmatory potential and epileptogenicity at diagnosis. Following this, we present morphometric evidence that pallido-thalamic network abnormalities might predispose an individual to acute seizure activity and post-acute recurrence following NCC infection. Next, we show that whole brain diffusion metrics and functional connectivity were mostly preserved across our patient subgroups, providing rationale for more granular characterisation of the perilesional environment in future studies. However, lesion-network mapping results in both modalities are suggestive of a potential dichotomy in the connectivity of cysts related to post-acute seizure freedom, and post-acute seizure recurrence. Finally, graph-theoretical comparisons contextualise the findings of the previous chapters, and are indicative of a clustering/segregation imbalance related to both seizure history, and seizure recurrence. Whilst statistical power was a significant concern throughout this project, the identified network abnormalities highlight that there is little consistency between the compensatory plasticity engendered by prior seizure activity, perilesional and distal disruptions that occur in the acute stage of an NCC infection, and network characteristics that predispose to post-acute seizure recurrence. As advanced neuroimaging technology becomes more accessible globally, it is crucial that we consider how it can be implemented effectively. This project demonstrates the ability of qMRI approaches over conventional neuroradiological evaluation to identify correlates of seizure presentation and recurrence in NCC, but also reveals areas of potential oversight in these methods. Furthermore, we have identified a potential common pallido-thalamic biomarker of epileptogenesis following a focal insult, with implications for the prognostication of NCC and early treatment of acquired epilepsy in general.