Neurology · Acute stroke imaging
e-Stroke Suite
Brainomix
Brainomix 360 is among the most independently scrutinised AI platforms in acute stroke imaging — externally validated against masked expert panels in a 4,100-patient study across nine trials, evaluated prospectively in unselected real-world cohorts, and studied at national scale across 107 NHS England hospitals. The evidence is deep and largely independent, but it is prospective-observational and external-validation rather than randomised, and the verified regulatory footprint spans FDA and CE without a confirmable third-jurisdiction approval — placing the suite firmly at two marks rather than three.
Performance Metrics
Clinical Evidence
The e-Stroke suite carries one of the more independent evidence bases in stroke AI. Its anchor study is the Real-world Independent Testing of e-ASPECTS Software (RITeS) external validation (Mair et al., Annals of Neurology 2022; 92(6):943–957), in which Brainomix staff were excluded from the research plan, image processing, statistical analysis, and writing. RITeS drew on 4,100 patients pooled from nine completed stroke trials and observational studies, each CT scored by masked central expert panels. It concluded that e-ASPECTS was non-inferior to expert readers, with greater sensitivity but lower specificity than the experts — a profile consistent with a triage-leaning decision-support tool. Performance figures vary substantially by setting, which procurement reviewers should weigh. A retrospective single-centre analysis (Frontiers in Neuroinformatics 2025; 1,029 non-contrast CT studies in 954 patients, March 2020–December 2024) reported 95.8% sensitivity (95% CI 93.6–97.3), 96.9% specificity (95% CI 94.7–98.2), and 96.3% accuracy for detecting acute ischaemia, with strong correlation to radiologist ASPECTS (ρ = 0.953). By contrast, a prospective evaluation in an unselected hyperacute cohort (Mallon et al., Stroke & Vascular Neurology 2023; 1,163 studies in 551 UCLH patients) found e-Stroke more specific (83.5%) than sensitive (58.6%) for acute middle-cerebral-artery-territory ischaemia, with 77.0% overall accuracy — a reminder that headline accuracy degrades against unselected, noisy real-world inputs. The suite's strongest claim is on care-pathway impact rather than standalone accuracy. A prospective observational study across NHS England (Harston et al., The Lancet Digital Health 2025) analysed more than 450,000 patients admitted to 107 hospitals over five years, with Brainomix 360 deployed to 26 of them and more than 71,000 ischaemic-stroke patients in the deployment cohort; a pre/post-implementation analysis associated deployment with roughly a doubling of endovascular thrombectomy rates and reduced inter-hospital transfer delays. A companion external validation of the e-CTA module (Mair et al., Annals of Clinical and Translational Neurology 2023) assessed CT-angiography accuracy for occlusion detection. The suite originated as an Oxford University spin-out and was an early NHS AI in Health and Care Award recipient (2020).
| Study | Design | n | Sensitivity | Specificity | AUC | Published |
|---|---|---|---|---|---|---|
| Mair G, White P, Bath PM, et al. (RITeS external validation) | RetrospectiveRetrospective | 4,100 | — | — | — | Annals of Neurology, 2022 (92:943–957); e-ASPECTS non-inferior to expert panels |
| Frontiers in Neuroinformatics — single-centre retrospective | RetrospectiveRetrospective | 954 | 95.8% | 96.9% | — | Frontiers in Neuroinformatics, 2025 (n=1,029 NCCT); accuracy 96.3%, ρ=0.953 |
| Mallon D, Cowley P, et al. (UCLH prospective real-world) | ProspectiveProspective | 551 | 58.6% | 83.5% | — | Stroke & Vascular Neurology, 2023 (n=1,163 studies); accuracy 77.0% |
| Harston G, et al. (NHS England prospective observational) | ProspectiveProspective | 71,000 | — | — | — | The Lancet Digital Health, 2025; ~100% increase in thrombectomy rates |
Clinical Pulse
Verified clinician sentiment
Verified clinician reviews launching · Apply to contribute →
Reviews are signed by clinicians verified against their regulator's public register.
Inside the Auris+ Listing
Six more sections complete this device’s Auris+ Listing.
Inside the algorithm
ProPro unlocks the five-stage editorial walkthrough of how this device's AI reaches a verdict — model architecture, training data, inference latency, and where the clinician stays in the loop.
Decision Ledger
ProPro unlocks a private, cross-vendor log of every case you read on this device — what the AI called, what you concluded, and a one-line reason if you overrode it. Ready for the EU AI Act's deployer logging obligations when they land in 2028.
Clinical Evidence Deep Dive
ProPro unlocks the structured clinical-evidence summary — study count, target patient population, and a tabular accuracy-metrics view drawn from peer-reviewed sources.
Peer-Reviewed Publications
ProPro unlocks the curated peer-reviewed publication list with PubMed cross-links — the citation backbone of every editorial verdict.
Post-Market & Regulatory Conditions
ProPro unlocks the post-market surveillance summary, recall record, and the conditions of approval that bound real-world use.
AI Algorithm Version History
ProPro unlocks the chronological record of algorithm version changes — what changed when, drawn from manufacturer changelogs and regulatory filings.
Regulatory Approvals
Class IIa
K243294
Class II
Safety Record
No Brainomix device recalls or FDA safety communications were identified in publicly available sources as of the review date; the MAUDE adverse-event database could not be queried directly from this environment, so this should be read as "none found in public reporting" rather than an exhaustive MAUDE audit. The platform is decision-support software — it does not alter image acquisition, deliver therapy, or interact with implanted hardware, which limits the harm surface. The principal documented risk is performance variability: the prospective UCLH cohort recorded 58.6% sensitivity for acute MCA ischaemia, so a negative e-ASPECTS read does not exclude early infarction and does not substitute for clinician review.
Intended Use & Indications
Brainomix 360 (commercially "e-Stroke") is a modular decision-support platform for the hyperacute stroke pathway. The e-ASPECTS module analyses non-contrast head CT and synthesises imaging features into a single Alberta Stroke Program Early CT Score (ASPECTS), presented with an overlaid hypodensity heatmap that the reading clinician can review and modify. The e-CTA module detects suspected large-vessel occlusion on CT angiography and automates the CTA collateral score. The e-CTP module estimates ischaemic core and penumbra volumes, mismatch ratio, and hypoperfusion intensity ratio from CT perfusion data; a later-cleared feature also estimates ischaemic core volume directly from non-contrast CT. The platform is a clinical decision-support tool — final diagnostic, treatment, and transfer decisions remain with the treating clinician.