Applying New Framework Improves Big Data Analysis for Pediatric and Adult Epilepsies

Published on

An international study led by researchers from Children’s Hospital of Philadelphia (CHOP) affiliated with the CHOP Epilepsy Neurogenetics Initiative (ENGIN) made major improvements to a framework that seeks to standardize the clinical descriptions of epilepsy used for computational analysis. This framework helps translate disease classification, which enables clinicians and researchers to better diagnose epilepsy. The study was published online by the journal Epilepsia.

Epilepsy is one of the most common neurological disorders in children. Even though seizures are the common feature across all epilepsy, the way that seizures present can be different between individuals. In particular, the way that various seizures present in a child can be very complex. In prior studies, researchers from CHOP have utilized Human Phenotype Ontology (HPO), a method that standardizes a patient’s clinical features and allows that data to be translated and analyzed similar to genetic data. Much like genetic data, the more information that is available, the more likely this descriptive information can lead to more accurate diagnoses and potentially improve precision medicine treatment options. In their current study, the researchers completely redesigned how seizures are classified by HPO to improve its consistency and support how this data can be used to guide research into epilepsy and its genetic causes.

“We formed an international team and spent two years arranging the current and historical expert classifications of seizures into a map that allows people’s seizures to be compared precisely, even if described from different perspectives,” said lead author David Lewis-Smith, PhD, Wellcome Trust Clinical PhD Fellow at the Translational and Clinical Research Institute at Newcastle University in England and a researcher at CHOP. “This will help providers make genetic diagnoses for people with epilepsy and researchers make the most of data from large numbers of people with epilepsy, even reusing data from prior studies, to discover new causes and better treatments.”

The CHOP-led international study team used the 2017 International League Against Epilepsy (ILAE) Operational Classification of Seizure Types as a basis for creating a new dictionary for describing the clinical features of seizures, or a so-called subontology. This new HPO seizure subontology integrated concepts at different levels of detail, such as seizures lasting for more than 30 minutes, seizures that occur when a child has a high fever, seizures caused by specific stimuli, and seizures occurring in newborns. This new HPO seizure subontology was compared to existing clinical descriptions of these patients prior to and following the revision based on seizure data from 791 patients from three independent cohorts, including 150 newly recruited patients.

Ingo Helbig, MD Ingo Helbig, MD The specific advantage of the HPO is the fact that it allows for the depth and quality of information to be measured similar to empirical data. Using the new seizure dictionary built during their study, the researchers increased the number of descriptive concepts for seizures five-fold. The number of descriptive terms for seizures that could be annotated to the cohort increased by 40% and the total amount of information about each individual’s seizures increased on average by 38%.

“This first-of-its-kind study essentially rewrites the code on how epilepsies will be handled in the future by any diagnostic lab or research study that uses this common language,” said senior author Ingo Helbig, MD, attending physician and director of the genomic and data science core of ENGIN. “We have demonstrated that we can capture significantly more information on the clinical presentation of children and adults with epilepsy, allowing us to make these large data resources tractable.”

This work was supported by the Wellcome Trust, the Hartwell Foundation, the National Institute for Neurological Disorders and Stroke (K02 NS112600), the Center Without Walls on ion channel function in epilepsy (“Channelopathy-associated Research Center”, U54 NS108874), the Eunice Kennedy Shriver National Institute of Child Health and Human Development, the Intellectual and Developmental Disabilities Research Center (IDDRC) at CHOP and the University of Pennsylvania (U54 HD086984), and intramural funds from CHOP through ENGIN. Research reported in this publication was also supported by the National Center for Advancing Translational Sciences of the National Institutes of Health (UL1TR001878), the Institute for Translational Medicine and Therapeutics’ (ITMAT) Transdisciplinary Program in Translational Medicine and Therapeutics at the Perelman School of Medicine of the University of Pennsylvania, the EuroEPINOMICS-Rare Epilepsy Syndrome (RES) Consortium by the German Research Foundation (HE5415/3-1) within the EuroEPINOMICS framework of the European Science Foundation, by the German Epilepsia Research Foundation (DFG; HE5415/5-1, HE5415/6-1) and by the DFG/FNR INTER Research Unit FOR2715 (He5415/7-1, INTER/DFG/17/11583046), in part by a research grant from Science Foundation Ireland (SFI) under Grant Number 16/RC/3948 and co-funded under the European Regional Development Fund and by FutureNeuro industry partner Congenica, and by the European Commission (Solve-RD; 779257).

Lewis-Smith et al, “Modelling seizures in the Human Phenotype Ontology according to contemporary ILAE concepts makes big phenotypic data tractable.” Epilepsia, online May 5, 2021. DOI: 10.1111/epi.16908.