A translational multimodal machine‐learning prototype predicting valproate response in epilepsy treatment
Ein translationaler multimodaler Machine-Learning-Prototyp zur Vorhersage des Valproat-Ansprechens bei Epilepsiebehandlung
Epilepsia · 2026-08 · 67 (8) · 4037-4050
DOI: 10.1002/epi.70300 ↗Auf einen Blick
Die Studie entwickelte und validierte ein multimodales Modell zur Vorhersage des Valproat-Ansprechens bei Epilepsie. In der unabhängigen Validierung erreichte es eine balanced accuracy von 63 %.
Fragestellung
Kann ein multimodales Modell das Ansprechen auf Valproinsäure bei Epilepsie vorhersagen?
Methode
Querschnittliche Studie zur Modellierung des Behandlungsansprechens mit Daten aus Epi25 und einer kanadischen Validierungskohorte; Merkmale aus genetischen, zellulären und klinischen Daten.
Ergebnisse
In der unabhängigen Validierungskohorte erreichte der multimodale Klassifikator eine balanced accuracy von 63 % (95 %-KI 52 %–73 %), NPV von 70 % (95 %-KI 51 %–85 %) und PPV von 60 % (95 %-KI 46 %–72 %).
Einschränkungen
Proof-of-concept; noch nicht für die klinische Anwendung bereit.
Schlussfolgerung der Autor:innen
Die Autor:innen berichten, dass die Integration genetischer, zellulärer und klinischer Daten eine Vorhersage des VPA-Ansprechens mit klinisch relevanter Genauigkeit ermöglicht.
Kernaussagen
- Multimodales Modell zur Vorhersage des VPA-Ansprechens
- Balanced accuracy 63 % in der unabhängigen Validierung
- Modelle mit nur einer oder zwei Datenmodalitäten waren schlechter
Deutsche Zusammenfassung: KI-gestützt aus dem Original-Abstract erstellt; alle Zahlen wurden automatisch mit dem Original abgeglichen. Maßgeblich ist der Original-Abstract.
Kurz gefasst
Fachartikel in der Zeitschrift Epilepsia, online erschienen am 20.05.2026. Verfasst von Simeon Platte und 15 weiteren Autor:innen. Beteiligte Einrichtung: Department of Child and Adolescent Psychiatry, Psychosomatics, and Psychotherapy, Goethe University Frankfurt, University Hospital, Frankfurt, German…. Hauptthemen laut MEDLINE: Valproic Acid, Anticonvulsants, Epilepsy, Machine Learning. Die Arbeit wurde bisher 2-mal zitiert (Crossref) und nennt 36 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: European Commission, Horizon Europe, European Commission; ERA PerMed, European Commission, Horizon 2020.
Abstract (Originalsprache)
Objective
Epilepsy affects ~1% of the global population and often requires lifelong antiseizure medication (ASM) therapy. Valproic acid (VPA) is a commonly prescribed first‐line ASM, yet only approximately half of patients achieve sustained seizure freedom. Treatment selection remains largely empirical. We aimed to develop and independently validate a multimodal predictive model to estimate response to VPA and support more individualized treatment strategies.
Methods
This cross‐sectional treatment response modeling study used data from a subset of the international Epi25 cohort (Belgium, Finland, Germany). Individuals with epilepsy were included if they had received VPA monotherapy and had available genetic or clinical data. Discovery data (1965–2021, 58% female) were split into a training set ( n = 196) and test set ( n = 133). Independent validation was performed in a Canadian cohort (2021–2022, n = 156, 40% female). The primary outcome was binary VPA response. Responders achieved ≥12 months of seizure freedom attributed to VPA; nonresponders had >50% seizure recurrence or discontinued VPA due to inefficacy, adverse effects, or unclear reasons. The predictive algorithm integrated features derived from common and rare variants in VPA pharmacokinetic and pharmacodynamic genes, in vitro neuronal VPA response measures, and clinical features. Model performance was assessed using accuracy, predictive values (negative predictive value [NPV]/positive predictive value [PPV]), and area under the curve (AUC).
Results
In the independent validation cohort, the multimodal classifier achieved a balanced accuracy of 63% (95% confidence interval [CI] = 52%–73%), NPV of 70% (95% CI = 51%–85%), PPV of 60% (95% CI = 46%–72%), and AUC of .73 (95% CI = .63–.83). Models restricted to single or dual data modalities showed consistently lower predictive performance. Significance This proof‐of‐concept study demonstrates that integrating genetic, cellular, and clinical data enables prediction of VPA treatment response with clinically meaningful accuracy. Although not yet ready for clinical application, this approach supports the feasibility of biomarker‐informed ASM selection and may ultimately reduce time to effective seizure control.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Stichwörter der Autor:innen
Substanzen
Autor:innen und Institutionen
- Department of Child and Adolescent Psychiatry, Psychosomatics, and Psychotherapy, Goethe University Frankfurt, University Hospital, Frankfurt, Germany.
- Laboratory of Experimental Neurology, Université Libre de Bruxelles, Brussels, Belgium.
- Department of Neurology and Neurosurgery, Montreal Neurological Institute, McGill University, Montreal, Quebec, Canada.
- Department of Clinical Neurosciences, Cumming School of Medicine, University of Calgary, Calgary, Alberta, Canada.
- Hotchkiss Brain Institute, University of Calgary, Calgary, Alberta, Canada.
- O'Brien Institute for Public Health, Cumming School of Medicine, University of Calgary, Calgary, Alberta, Canada.
ORCID-Profile: Simeon Platte, Afsheen Kumar, Massimo Pandolfo, Denise Haslinger, Colin B Josephson, Guillermo Delgado-García, Chantal Depondt, Reetta Kälviäinen, Felix Rosenow, Sophie von Brauchitsch, Karl Martin Klein, Andreas G Chiocchetti
Förderung
- European Commission, Horizon Europe
- European Commission; ERA PerMed
- European Commission, Horizon 2020
- European Commission
Zitieren
APA
Platte, S., Kumar, A., Guerini, G., Pandolfo, M., Haslinger, D., Josephson, C. B., Delgado-García, G., Kaur, N., Lux, M. P., Stempfle, H., Depondt, C., Kälviäinen, R., Rosenow, F., von Brauchitsch, S., Klein, K. M., & Chiocchetti, A. G. (2026). A translational multimodal machine‐learning prototype predicting valproate response in epilepsy treatment. Epilepsia, 67(8), 4037-4050. https://doi.org/10.1002/epi.70300
Vancouver
Platte S, Kumar A, Guerini G, Pandolfo M, Haslinger D, Josephson CB, et al. A translational multimodal machine‐learning prototype predicting valproate response in epilepsy treatment. Epilepsia. 2026;67(8):4037-4050. doi:10.1002/epi.70300
BibTeX
@article{Platte2026,
author = {Platte, Simeon and Kumar, Afsheen and Guerini, Giorgia and Pandolfo, Massimo and Haslinger, Denise and Josephson, Colin B and Delgado-García, Guillermo and Kaur, Navprabhjot and Lux, Michaela-Pauline and Stempfle, Heiko and Depondt, Chantal and Kälviäinen, Reetta and Rosenow, Felix and von Brauchitsch, Sophie and Klein, Karl Martin and Chiocchetti, Andreas G},
title = {A translational multimodal machine‐learning prototype predicting valproate response in epilepsy treatment},
journal = {Epilepsia},
year = {2026},
volume = {67},
number = {8},
pages = {4037--4050},
doi = {10.1002/epi.70300}
}
EUVIMED