Biological Feature-based Cluster Analysis Reveals Inflammatory Response Heterogeneity in Patients with Cryptogenic New-onset Refractory Status Epilepticus
Abstract number :
1.221
Submission category :
2. Translational Research / 2C. Biomarkers
Year :
2024
Submission ID :
726
Source :
www.aesnet.org
Presentation date :
12/7/2024 12:00:00 AM
Published date :
Authors :
Martin Guillemaud, BS – Paris Brain Institute
Mario Chavez, PhD – Paris Brain Institute
Annamaria Vezzani, PhD – Istituto Mario Negri
Anthony Jimenez, BS – Yale University School of Medicine
Maysaa Basha, MD, FAES, FANA – Wayne State University School of Medicine
Ayush Batra, MD – Feinberg School of Medicine, Northwestern University
Sophie Demeret, MD – Pitie Salpetriere Hospital, Paris
Onome Eka, MBBS – Icahn School of Medicine at Mount Sinai, New York City
Krista Eschbach, MD – University of Colorado Anschutz Medical Campus – Children’s Hospital Colorado
Brandon Foreman, MD – University of Cincinnati Medical Center
Nicolas Gaspard, MD, PhD – Hôpital Universitaire de Bruxelles
Elizabeth Gerard, MD – Feinberg School of Medicine, Northwestern University
Teneille Gofton, MD – London Health Sciences Centre (LHSC)
Hiba Haider, MD – University of Chicago Medicine
Charles Howe, PhD – Department of Neurology, Mayo Clinic, Rochester
Amy Jongeling, MD, PhD – NYU Grossman School of Medicine
Mariel Kalkach Aparicio, MD, MBE – University of Wisconsin-Madison
Padmaja Kandula, MD – Weill Cornell Medicine, Department of Neurology, New York City,
Karnig Kazazian, BS – London Health Sciences Centre (LHSC)
Minjee Kim, MD – Feinberg School of Medicine, Northwestern University
yichen lai, MD – Texas Children's Hospital, Baylor College of Medicine
clemence marois, MD – Pitie Salpetriere Hospital, Paris
andrew mellor, BS – Children’s Hospital Colorado, Aurora
wazim mohamed, MD – Wayne State University School of Medicine
mikaela morales, BS – Divison of Pediatric Neurology, University of Washington, Seattle WA
vincent Navarro, MD, PhD – Pitie Salpetriere Hospital, Paris
cederic pimentel, MD – Neurocritical Care, Emory University, Atlanta, GA
alexandra ramirez, BS – University of Cincinnati, Department of Neurology and Rehabilitation Medicine
Claude Steriade, MD – NYU Grossman School of Medicine
Aaron Struck, MD – University of Wisconsin-Madison
Olga Taraschenko, MD, PhD – University of Nebraska Medical Center
Nathan Torcida Sedano, MD – Hôpital Universitaire de Bruxelles
Mark Wainwright, MD, PhD – Seattle Children's Hospital
Jiyeoun (Jenna) Yoo, MD – Icahn School of Medicine at Mount Sinai
Lawrence Hirsch, MD – Yale University School of Medicine
Presenting Author: Aurélie Hanin, PharmD, PhD – Yale University School of Medicine
Rationale: Cryptogenic New-Onset Refractory Status Epilepticus (cNORSE) is thought to be a disorder of immunity leading to exacerbated inflammation. We previously identified cytokine dysregulations in cNORSE patients and found correlations between cytokine levels and outcomes, suggesting the involvement of innate immunity-related inflammation in cNORSE pathogenesis. However, not all patients with cNORSE presented with the same cytokine profile, possibly revealing different etiologies. Therefore, we investigated if clusters of patients with cNORSE can be identified based on a highly multiplexed serum cytokine profile.
Methods:
We measured 96 cytokines (96-plex Discovery Assay Array, Eve Technology, Canada) in the serum of 62 patients with cNORSE, hospitalized in ICU in 25 centers and enrolled into NORSE biorepositories (Yale/Paris). Correlations among the cytokine levels were used to define groups of cytokines. The cytokine levels for each patient were then replaced by the principal component of their respective groups. Using these new features, a similarity matrix between all patients was obtained and a non-parametric graph clustering procedure (Louvain’s algorithm) was applied to identify clusters of patients. Finally, clinical data, including functional outcomes (GOS-E), were compared between clusters of patients.
Results:
Patients had a median age of 31 years (IQR 21-60). 66% of patients qualified as FIRES and 74% received at least one immunotherapy before serum collection. Three groups of cytokines were identified, corresponding mainly to innate inflammation and cytokine storm (group 1), mucosal immune response (group 2), and innate/autoimmune markers (group 3). Based on these 3 biological groups, the similarity matrix highlighted 8 clusters of patients (Fig 1). Besides different cytokine profiles, clusters differed in patient age and number of anti-seizure medications received. The graph revealed similarities between clusters 3 and 4 (group A), clusters 1 and 5 (group B), and clusters 0, 2, and 6 (group C). In contrast, cluster 7 was similar to most clusters.
By comparing the 3 groups of patient clusters (A, B, and C), we noticed significant differences in age, treatments received, and short- and long-term outcomes. Patients from group A were older, had worse outcomes, and expressed elevated levels of group 1 markers, including IL6, CXCL8, MCP-1, and IP-10. Patients from group C had better outcomes and expressed elevated levels of group 2 markers, including IL-29 and IL-35 with anti-inflammatory properties. Despite a similarly poor outcome as group A, patients from group B did not present a high inflammatory profile, suggesting different pathophysiological mechanisms.
Conclusions: This study sheds light on the heterogeneity of patients with cNORSE. By studying a large number of cytokines, we were able to identify clusters of patients who may have a similar pathophysiological mechanism or even etiology.
Funding: NORSE/FIRES Research Fund at Yale, Wong Neurology Research Fund
Translational Research