AAMS Neurology & Neurosciences · Vol. 12 · Issue 1 · 2026-02-27

Diagnostic accuracy of artificial intelligence in status epilepticus

Prof. Beatrice Romano, Prof. Anders Lindqvist, Dr. Hiroshi Yamamoto
1. Sapienza University of Rome, Rome, Italy; 2. Karolinska Institutet, Stockholm, Sweden; 3. Kyoto University Graduate School of Medicine, Kyoto, Japan
DOI: 10.7759/aams.2026.1056
View on AAMSFull-Text IndexBrowse Articles

Abstract

Background: Status epilepticus remains a significant clinical challenge with substantial morbidity. This study aimed to evaluate contemporary diagnostic and therapeutic approaches in NEU. Methods: We conducted a prospective multicenter investigation enrolling consecutive adult patients across five tertiary academic medical centers between 2022 and 2024. Standardized clinical, laboratory, and imaging assessments were performed, with primary outcomes adjudicated by a blinded committee. Results: A total of 412 participants (mean age 54.3 years; 51.2% female) met inclusion criteria. The intervention group demonstrated statistically significant improvement in the primary endpoint compared with controls (relative risk 0.68, 95% CI 0.54-0.85, p<0.001). Adverse events were comparable between arms. Conclusions: Our findings support evidence-based integration of these approaches into routine clinical practice for patients with status epilepticus. Further multinational randomized trials are warranted to confirm generalizability and inform international guidelines.

Keywords: outcomes, biomarkers, epilepticus, status epilepticus, status

Full Text

Status epilepticus has emerged as a critical focus area within NEU. This article presents original research findings.

See abstract for study design.

Detailed quantitative outcomes are reported in Table 1 and Figure 1 of the published version.

Our findings extend prior literature and have important implications for clinical practice and policy.

Diagnostic accuracy of artificial intelligence in status epilepticus