Rationale:
Epilepsy panel testing and exome sequencing are used for genetic diagnosis in epilepsy. We aimed to extend the understanding of yield of exome sequencing in epilepsy to assess yield of exome after a non-diagnostic epilepsy gene panel and the clinical predictors of a diagnostic result.
Method:
We evaluated patients with epilepsy from Boston Children’s Hospital who had epilepsy gene panel testing followed by clinical whole exome sequencing between 2011 and 2019. Results were considered diagnostic if identified genetic variants on initial testing or reanalysis (clinical or research based) were determined clinically to explain the patient’s epilepsy (pathogenic, likely pathogenic, or with literature after the laboratory report to suggest pathogenicity). We characterized genotypic and phenotypic features for each patient including genetic variant details and classification of pathogenicity, age of seizure onset, EEG patterns, family history, neuroimaging, developmental disorders, regression, systemic malformations, cerebral visual impairment (CVI), head size, muscle tone, and facial dysmorphic features. We assess association of clinical features with diagnostic results using logistic regression modeling.
Results:
Our cohort consisted of 108 patients (44% F, 56% M) with epilepsy onset at median 9 months (Interquartile range 3 months to 1.6 years). The additional diagnostic yield of exome sequencing in pediatric patients with epilepsy who had a negative epilepsy gene panel test was 37% (40/108) (Figure 1). Of the 40 patients with a diagnostic exome, medical management was clearly impacted in 9 (23%) including impact on anti-seizure medication choices, precision therapy, and monitoring for associated medical conditions. Univariate analysis suggested that the odds of a diagnostic result was higher in the setting of epilepsy age of onset ≤ 2 years and abnormal muscle tone (Table 1). After building a multivariable logistic regression model using a purposeful selection technique, only age of onset ≤ 2 years is significant. Median time between panel and exome was not statistically different between those with diagnostic and non-diagnostic exome results, 449 days (interquartile range 306, 1147) for those with a diagnostic exome result compared to median 412 days (interquartile range 259, 814) for those with a non-diagnostic result.
Conclusion:
Overall, this study indicates meaningful yield of exome after a non-diagnostic epilepsy gene panel including impact on medical management. Within the power of the current study, the only phenotypic feature predicting a diagnostic result is epilepsy onset before 2 years. Thus, it is reasonable to consider exome in any individual with epilepsy of suspected genetic etiology and negative gene panel test, especially those with onset in the first 2 years of life.
Funding:
:NINDS 1K23 NS107646-01, Harvard Medical School Eleanor and Miles Shore fellowship program
FIGURES
Figure 1