Authors :
Presenting Author: Gabriella Foca, BA – NYU Grossman School of Medicine
Aaron Smith, BSN, RN – NYU
Gregory Cascino, MD – Mayo Clinic, Rochester MN, USA.
Dennis Dlugos, MD, MSCE – Children’s Hospital of Philadelphia (CHOP), Philadelphia, PA, USA
Michael Gelfand, MD, PhD – Perelman School of Medicine at the University of Pennsylvania
Orrin Devinsky, MD – NYU Langone
Jacqueline French, MD – Department of Neurology and Comprehensive Epilepsy Center, New York University Grossman School of Medicine, NYU Langone Health
Manisha Holmes, MD, FAES – Westchester Medical Center Health Network
Rationale:
This study explores whether specific clinical or electrographic features are associated with lower inter-rater agreement among epilepsy specialists when classifying idiopathic generalized epilepsy (IGE). To understand the variability in expert interpretation and identify factors that contribute to diagnostic uncertainty in IGE, we examined initial adjudication data from the Human Epilepsy Project 3 (HEP3).
Methods:
HEP3 subjects were enrolled into three cohorts: 1) newly diagnosed (within 12 months of diagnosis), 2) treatment sensitive (>2 years ongoing treatment, < 2 ASMs failed for efficacy), and 3) treatment resistant. At enrollment, a standardized adjudication form summarizing each participant’s medical history, seizure semiology, and EEG findings was completed and submitted to a panel of five epileptologists. Adjudicators independently reviewed cases and assigned each participant to one of three confidence levels for IGE: < 50% (Low confidence, L), 50-80% (Medium confidence, M), or >80% (High confidence, H) and provided reasoning for reduced confidence using 1) age of onset, 2) seizure types, 3) EEG, and 4) other clinical features. Participants with >3 adjudicator responses indicating medium-high confidence in a diagnosis of IGE were accepted via group consensus. Highly ambiguous cases were those with at least one vote of low confidence in IGE diagnosis and less than 4 votes of high confidence.
Results:
22 cases met inclusion criteria for highly ambiguous, including 13/84 in the newly diagnosed cohort (C1), 4/68 in the treatment sensitive cohort (C2), and 5/49 in the treatment resistant cohort (C3). Adjudicator confidence vote (Table 1) varied widely across adjudicators and study cohorts. Reasons for reduced confidence were consistent across groups: EEG most often cited as the reason for doubt (n=44), followed by other clinical features (e.g., seizure semiology) (n=19). These categorical factors were most often cited together. Clinical features were less often cited alone (n=16) as a cause for reduced confidence in IGE diagnosis. Older age at seizure onset (n=7) was associated with greater uncertainty among reviewers, particularly in newly diagnosed patients (n=6).
Conclusions:
Newly diagnosed patients comprised the majority of the highly ambiguous diagnosis group (59%). The frequency of ambiguous diagnoses varied between newly diagnosed (15.5%), treatment sensitive (5.9%) and treatment resistant (10.2%) cohorts, suggesting diagnostic confidence may be associated with treatment response. EEG was the most frequently cited factor for diagnostic uncertainty in IGE. Future analysis will involve all enrolled participants.
Funding:
The HEP3 study is supported by the Epilepsy Study Consortium (ESCI), a non-profit organization dedicated to accelerating the development of new therapies in epilepsy to improve patient care. The funding provided to ESCI to support HEP3 comes from industry, philanthropy and foundations.