Combining Genomic, EHR Data Could Improve Epilepsy Treatment

August 13, 2020 - By combining EHR data and genomic information, researchers were able to identify gene-specific signatures in childhood epilepsies, according to a study published in Genetics in Medicine.

The findings could lead to better treatments for the disorder, as well as improved clinical decision support tools.

“Our study is the first example in childhood neurological orders to systematically connect genomic information with the medical records,” said Ingo Helbig, attending physician at CHOP’s Epilepsy Neurogenetics Initiative (ENGIN), director of the genomic and data science core of ENGIN and lead investigator on this study.

“This is really important as we need to understand the clinical features that children with genetic brain disorders, especially children with genetic epilepsies, develop over time. Using the technologies that we have developed, we can use the available data in the electronic medical records to bridge the gap between genetics and outcomes."

Genetic factors have been increasingly implicated in childhood epilepsies, the team noted. In the last decade, researchers have identified more than 200 genetic causes of epilepsy. Genetic changes can be found in up to 30 percent of Developmental and Epileptic Encephalopathies (DEE), severe brain disorders that can cause aggressive seizures, cognitive and neurological impairment, and in some cases early death.

Identifying a causative gene is often the first step of improving treatment, because many children with these conditions don’t respond to current treatment methods.

While collectively common, each causal gene is only found in one percent or less of the patient population, making it difficult to generate enough clinical information to provide families and physicians with reliable information on how these conditions develop over time.

Moreover, although genomic data is collected in a standardized way, the patient’s phenotype – a set of clinical findings that may include seizures or developmental disabilities – historically has not been collected in the same way.

The researchers stated that initiatives to link genomic data with EHR information are already underway. However, because these initiatives are relatively new, the role of EHRs in studying how disease-causing genetic changes can impact patients over longer periods of time has not yet been explored.

Researchers examined 62,104 patient encounters in 658 individuals with known or presumed genetic epilepsies. To standardize clinical observations, the team leveraged the Human Phenotype Ontology (HPO), a catalogue that provides a standardized format to characterize a patient’s phenotypic features. HPO also allows researchers to process clinical information through data science techniques.

This resulted in a total of 286,085 HPO terms, which were then grouped to 100 three-month time intervals, with researchers assessing gene-phenotype associations at each interval.

Researchers were able to identify significant associations of various known genetic causes of epilepsy, including status epilepticus with the gene SCN1A at one year of age. Status epilepticus is a dangerous condition in which epileptic seizures last longer than five minutes or follow in a short sequence without full recovery in between them.

The team also discovered an association between severe intellectual disability with the gene PURA at age ten and infantile spasms with the gene STXBP1 at six months.

These associations mirror known clinical features of each of these conditions that were identified through an automated analysis framework evaluating more than 3,200 observational patient years. This amount of clinical data goes far beyond what researchers could have accomplished through a manual chart review.

The combination of EHR data and genomic information could help develop effective clinical decision support and treatments for neurological disorders in children, the researchers stated.

“With new precision therapies emerging, there is a pressing demand to understand the natural history of rare genetic epilepsies,” said Sudha Kilaru Kessler, MD, a pediatric neurologist who is part of the leadership of CHOP’s ENGIN and director of epilepsy surgery at CHOP’s Pediatric Regional Epilepsy Program.

“Electronic medical records are an untapped resource to learn about how very rare disorders present over time, which will allow us to include this information in our clinical practice. Finally, these tools will allow us to develop clinical decision support and learning health systems with the ultimate aim to improve the life of our patients.”