Stroke Agent and outcome prediction model
Individualized outcome prediction in acute ischemic stroke is challenging but crucial for treatment decision-making and communication with patients and their families. In our projects, we are developing models for prognoses and treatment decision support in acute ischemic stroke, ultimately improving patient outcomes. By uniting statistical with deep learning approaches, we aim at generating interpretable, multimodal models that are not only accurate but transparent, which is crictial for clinical translation. In different applications, we have shown state-of-the-art prediction performance while outperforming medical experts in experimental settings.
To advance our models, benchmark performance against physicians and increase our understanding of treatment-decision making in acute stroke care, we currently host the USZ Thrombectomy Outcome Prediction in large vessel occlusion STROKE (TOP-STROKE) study, a public rating challenge which is open to all medical experts.
See publications below:
L. Herzog, N. Blindenbacher, C. Globas, M. I. Haeberlin, P. Baumgartner, F. Capecchi, C. Inauen, B. Sick, C. Majoie, W. van Zwam and S. Wegener, "Human Intuition vs. Computational Precision: Neurologists, Feature-based Models, and Deep Learning for Stroke Prognosis," medRxiv, 2026.
L. Herzog, J. Brändli, M. Schneeberger, L. Avci, N. Dari, M. Hänsel, H. Baazaoui, P. Bühler, S. Wegener and B. Sick, "Explainability in mulimodal deep transformation models for stroke outcome prediction," Accepted at MICCAI, 2026.
L. Herzog, P. Bühler, E. de la Rosa, B. Sick and S. Wegener, "Outcome Prediction and Individualized Treatment Effect Estimation in Patients with Large Vessel Occlusion Stroke," in Image Analysis in Stroke Diagnosis and Interventions: 5th International Workshop, SWITCH 2025, Held in Conjunction with MICCAI 2025, Daejeon, South Korea, September 23, 2025, Proceedings, 2025.
L. Herzog, L. Kook, J. Hamann, C. Globas, M. R. Heldner, D. Seiffge, K. Antonenko, T. Dobrocky, L. Panos, J. Kaesmacher, U. Fischer, J. Gralla, M. Arnold, R. Wiest, A. R. Luft, B. Sick and S. Wegener, "Deep Learning Versus Neurologists: Functional Outcome Prediction in LVO Stroke Patients Undergoing Mechanical Thrombectomy," Stroke, vol. 54, 2023.
L. L. Herzog & Kook, A. Götschi, K. Petermann, M. Hänsel, J. Hamann, O. Dürr, S. Wegener and B. Sick, "Deep transformation models for functional outcome prediction after acute ischemic stroke," Biometrical Journal, vol. 65, p. 2100379, 2023.
J. L. Hamann & Herzog, C. Wehrli, T. Dobrocky, A. Bink, M. Piccirelli, L. Panos, J. Kaesmacher, U. Fischer, C. Stippich, J. Luft, M. Arnold, R. Wiest, B. Sick and S. Wegener, "Machine-learning based outcome prediction in stroke patients with middle cerebral artery-M1 occlusions and early thrombectomy," European Journal of Neurology, vol. 28, no. 4, p. 1234–1243, 2021.
L. Herzog, E. Murina, O. Dürr, S. Wegener and B. Sick, "Integrating uncertainty in deep neural networks for MRI based stroke analysis," Medical Image Analysis, vol. 65, no. 11, 2020.