Neurology · Quantitative brain MRI/CT volumetry and ARIA decision-support
icobrain
icometrix (a GE HealthCare company)
icobrain is the longest-standing quantitative brain-imaging platform on the US market — its volumetry engine has carried FDA clearance since 2016, across six successive 510(k)s that extended it from MRI atrophy measurement to acute head-CT triage support and CT perfusion. Its newest module, icobrain aria, is the first AI tool cleared in the United States and CE-marked in Europe for detecting and grading the imaging side-effects of the new Alzheimer's antibody therapies, a genuinely first-in-class indication. The evidence base is real but uneven: the strongest independent validations are multi-reader and retrospective rather than randomised, and the pivotal icobrain aria reader study is manufacturer-authored — and one independent traumatic-brain-injury validation found experienced attendings still outperformed the model. A broad FDA-plus-CE footprint, a clean public safety record and deployment at scale place it at two marks; the absence of randomised evidence and of independently verifiable approvals beyond the US and EU keep it from three.
Performance Metrics
Clinical Evidence
icobrain's evidence base spans four clinical domains, and its quality varies by domain — a distinction that matters more than any single headline metric. In dementia, the strongest study is the largely independent REMEMBER multi-centre cohort (Wittens et al., J Alzheimers Dis 2021; PMID 34334402), which evaluated icobrain dm volumetry against FreeSurfer on 820 clinical brain MRIs spanning cognitively healthy controls, subjective cognitive decline, mild cognitive impairment and Alzheimer's dementia. icobrain dm distinguished healthy controls from Alzheimer's dementia with an AUC of 0.914 (sensitivity 86.3%, specificity 83.0%) and ran with zero processing failures against FreeSurfer's 67, at minutes per scan rather than hours — a robustness and throughput advantage relevant to routine clinical deployment. In traumatic brain injury, the most informative study is independent and appropriately cautionary (Vande Vyvere / multi-reader study, Neuroradiology 2023; PMID 37269414), which validated icobrain tbi v5.0 on 100 non-contrast head-CT scans read by eight reviewers of varying seniority. The model showed high specificity (0.88) and positive predictive value (0.96) for classifying NIRIS severity, and using it as an assistive tool improved trainees' agreement with ground truth — but experienced neuroradiology attendings still outperformed the model unaided, a finding the authors frame as needing further refinement before clinical reliance. For icobrain aria, the pivotal evidence is a manufacturer-authored multireader multicase reader study (Bracchi / icometrix authors, JAMA Network Open 2024; PMID 38345816) in which 16 radiologists read 199 retrospective brain-MRI pairs from patients on amyloid-beta-directed therapy with and without the software. Software assistance was associated with a statistically significant improvement in detection — reported as a roughly 16-point gain in ARIA-E sensitivity and a 10-point gain in ARIA-H sensitivity — alongside shorter reading times and higher inter-reader agreement. The result is meaningful but should be read with its authorship and retrospective, enriched-pair design in mind. The structural caveat across all four domains: no randomised controlled trial of any icobrain module has been published as of the review date, and several supporting validations (multiple-sclerosis atrophy reproducibility in particular) are mixed or manufacturer-authored. The platform's value is best read as quantification and standardisation of routine reads, not autonomous diagnosis.
| Study | Design | n | Sensitivity | Specificity | AUC | Published |
|---|---|---|---|---|---|---|
| Wittens MMJ, Sima DM, Houbrechts R, et al. (REMEMBER, multi-centre, largely independent) | RetrospectiveRetrospective | 820 | 86.3% | 83.0% | 0.914 (healthy controls vs Alzheimer's dementia) | J Alzheimers Dis, 2021; PMID 34334402 |
| Multi-reader validation, 8 reviewers (independent; Stanford / MD Anderson / Karolinska) | RetrospectiveRetrospective | 100 | — | 0.88 | PPV 0.96 for NIRIS severity classification | Neuroradiology, 2023; PMID 37269414 — attendings outperformed the model unaided |
| icobrain aria reader study, 16 radiologists (manufacturer-authored MRMC) | RetrospectiveRetrospective | 199 | +16 pts ARIA-E, +10 pts ARIA-H (assisted vs unassisted) | — | — | JAMA Network Open, 2024; PMID 38345816 (199 retrospective MRI pairs) |
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
K240712
Class II
Class IIa
Safety Record
No icometrix or icobrain device recalls or FDA safety communications were identified in publicly available sources as of June 2026, and a search of the FDA MAUDE adverse-event database for "icobrain" returned no entries at the time of the source dossier (2026-06-11). Because MAUDE and the FDA enforcement database could not be queried directly from this environment, this should be read as "none found in public reporting" rather than an exhaustive audit. The modules are decision-support and reporting overlays — they quantify structures and flag findings but do not alter source images, remove studies from the reading queue or drive treatment automatically — which bounds the harm surface. The principal documented risk is over-reliance: independent traumatic-brain-injury validation showed experienced readers outperforming the model, and the dementia and ARIA evidence rests on retrospective and multi-reader designs, so a quantitative output or the absence of a flag does not substitute for the radiologist's independent read.
Intended Use & Indications
icobrain is a modular quantitative-neuroimaging platform that procurement reviewers should treat as several distinct regulatory products under one brand. The core volumetry engine performs automatic labelling, visualisation and volumetric quantification of segmentable brain structures from MR or non-contrast CT (NCCT) images and supports longitudinal comparison of scans acquired at different times — marketed clinically as icobrain ms (multiple sclerosis lesion and atrophy tracking), icobrain dm (dementia and Alzheimer's volumetry) and icobrain ep (epilepsy). The CT pathway adds icobrain tbi, which quantifies haemorrhagic volumes, cisternal compression and midline shift on NCCT in acute traumatic brain injury, and icobrain ctp, which computes perfusion parameters from CT-perfusion datasets of the head. icobrain aria is a separately cleared computer-aided detection and diagnosis (CADe/CADx) module that detects, measures and grades amyloid-related imaging abnormalities — ARIA with oedema/effusion (ARIA-E) and ARIA with haemorrhage/siderosis (ARIA-H) — on routine brain MRI of patients receiving amyloid-beta-directed antibody therapies. In every module the system is decision support: it generates quantitative reports and findings alongside the radiologist's read, and final diagnostic and treatment decisions remain with the interpreting clinician.