Radiology · Quantitative multiparametric MRI post-processing for liver-tissue characterisation
LiverMultiScan
Perspectum Ltd (formerly Perspectum Diagnostics Ltd), Oxford, UK
LiverMultiScan is best understood as a quantitative measuring instrument rather than a diagnostic reader — and its authority rests on how thoroughly that instrument has been validated. Out of a short non-contrast MRI it derives iron-corrected T1 (cT1), a composite marker of liver fibro-inflammation, along with liver fat and iron. The evidence base is unusually deep for an imaging tool and much of it is independent: the original multiparametric-MRI validation against liver histology reported areas under the curve above 0.9 for fibrosis, fat and iron; a prospective cohort followed patients with chronic liver disease and found the imaging predicted liver-related clinical events; a multicentre pooled analysis put cT1's accuracy for high-risk NASH at 0.78, ahead of MRI liver fat at 0.69; and a UK Biobank sub-study established population reference ranges, with a low-risk median around 666 milliseconds. The device has been FDA 510(k)-cleared and marketed in the United States since 2015 and is CE-marked in Europe, with a clean public safety record. What holds it to two marks rather than three is the shape of the evidence and the footprint: there is no randomised trial of the device's own diagnostic accuracy — cT1 appears in randomised drug trials as an endpoint, which is a different thing — and a third-jurisdiction approval beyond the US and EU could not be confirmed from public primary sources. It is a rigorously characterised quantification aid, and the entry treats it as exactly that.
Performance Metrics
Clinical Evidence
LiverMultiScan's evidence base is one of the deepest in quantitative body MRI, and it should be read for what the device actually claims — measurement of tissue properties — rather than as evidence of autonomous diagnosis. The foundational study is Banerjee and colleagues (Journal of Hepatology 2014; PMID 24036007), a single-centre validation that compared multiparametric MRI against liver biopsy and reported areas under the receiver-operating curve above 0.9 for the separate quantification of fibrosis, steatosis and iron. It established cT1 as a histology-correlated metric and set the early cut-offs; as a single-centre diagnostic study it is a strong but not definitive anchor. The most clinically weighty result is prospective. Pavlides and colleagues (Journal of Hepatology 2016; PMID 26471505) followed 112 patients with chronic liver disease for a median of 27 months and found that the multiparametric-MRI read predicted liver-related clinical events — the kind of outcome signal that a purely cross-sectional accuracy figure cannot provide, and the basis for grading this entry's evidence as prospective. Breadth comes from a multicentre pooled analysis (Clinical Gastroenterology and Hepatology 2022; PMID 34626833), which combined individual data across five studies of suspected NAFLD patients with paired biopsy and reported cT1's accuracy for identifying high-risk NASH at an AUROC of 0.78 (95% CI 0.74–0.82), superior to MRI liver fat at 0.69. Population context comes from a UK Biobank sub-study (Mojtahed and colleagues, Abdominal Radiology 2019; PMID 30032383), which characterised cT1 in 1,037 participants at low risk of fatty-liver disease and reported a median of 666 milliseconds (IQR 643–694), with age and sex having minimal effect — the reference range clinicians read individual values against. A crucial honesty point: there is no randomised controlled trial of LiverMultiScan's own diagnostic accuracy. cT1 and PDFF do appear as quantitative endpoints inside randomised MASLD/MASH drug trials, and post-approval work has begun to define response thresholds (a cT1 reduction on the order of 80 milliseconds has been proposed as a monitoring threshold for steatohepatitis resolution), but using a biomarker as a trial endpoint is not the same as randomising patients to test the device's clinical accuracy. The entry does not conflate the two.
| Study | Design | n | Sensitivity | Specificity | AUC | Published |
|---|---|---|---|---|---|---|
| Banerjee R, Pavlides M, Tunnicliffe EM, et al. (University of Oxford) | RetrospectiveRetrospective | 79 | — | — | >0.9 for fibrosis, steatosis and iron vs liver biopsy (single-centre validation) | J Hepatol 2014;60(1):69–77; PMID 24036007 |
| Pavlides M, Banerjee R, Sellwood J, et al. (University of Oxford) | ProspectiveProspective | 112 | — | — | Multiparametric MRI predicted liver-related clinical events; median follow-up 27 months | J Hepatol 2016;64(2):308–315; PMID 26471505 |
| Multicentre pooled analysis, 5 studies of suspected NAFLD with paired biopsy | RetrospectiveRetrospective | 0 | — | — | cT1 AUROC 0.78 (95% CI 0.74–0.82) for high-risk NASH, superior to MRI liver fat 0.69 | Clin Gastroenterol Hepatol 2022; PMID 34626833 (pooled individual-participant data) |
| Mojtahed A, Kelly CJ, Herlihy AH, et al. (UK Biobank imaging sub-study) | RetrospectiveRetrospective | 2,816 | — | — | Low-risk reference range: median cT1 666 ms (IQR 643–694) in 1,037 low-risk participants | Abdom Radiol 2019;44(1):72–84; PMID 30032383 |
Clinical Pulse
Verified clinician sentiment
Verified clinician reviews launching soon · Apply to contribute →
Reviews are signed by clinicians verified against their regulator's public register.
Inside the algorithm
Editorial featureHow LiverMultiScan turns a short liver MRI into standardised numbers.
Five stages — from raw input to verdict — drawn from manufacturer documentation and the public regulatory record.
- INGEST
- NORMALISE
- DETECT
- LOCALISE
- VERDICT
Stage 01 · INGEST
A non-contrast MRI is acquired on a compatible scanner.
LiverMultiScan works from a short, non-contrast MRI of the liver acquired with a defined protocol on compatible scanners. The relevant sequences are T1 mapping, a proton-density fat-fraction acquisition and a T2*/R2* acquisition — the raw physics from which the biomarkers are computed.
Nothing is diagnosed at this stage; the platform is receiving standardised image data. Because the outputs are only as reliable as the input, protocol and scanner compatibility are part of the device's cleared use, not an afterthought.
Input
Non-contrast liver MRI
Inference
Quantitative post-processing (not autonomous diagnosis)
Inside the Auris+ Listing
Five more sections complete this device’s Auris+ Listing.
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
Safety Record
No Perspectum or LiverMultiScan recall, FDA safety communication or MAUDE adverse-event report was identified in publicly available sources as of July 2026; because the FDA MAUDE and enforcement databases could not be queried directly from this environment, this should be read as "none found in public reporting" rather than an exhaustive audit. The residual-risk profile is bounded by what the software does: it is a post-processing measurement aid, not an autonomous or diagnostic reader, so its principal failure modes are the ordinary ones for quantitative decision-support — a measurement is only as good as the input image, acquisition on a non-validated scanner or protocol can bias results, and a cT1, fat or iron value supports but does not replace the clinician's interpretation. Iron loading is a known confounder of T1 that the cT1 correction is specifically designed to address, which is why the uncorrected value is not reported as the fibro-inflammation marker.
Intended Use & Indications
LiverMultiScan is a standalone post-processing software application that analyses magnetic-resonance images of the liver acquired on compatible scanners and returns standardised, quantitative measures of liver-tissue characteristics. Its principal output is iron-corrected T1 (cT1), a T1-mapping metric that reflects extracellular fluid changes associated with inflammation and fibrosis and is corrected for the confounding effect of iron; alongside it the software reports liver fat as proton-density fat-fraction and iron content via T2*/R2*. The analysis segments and quality-controls regions of interest within the liver, computes the biomarkers, and presents them for a clinician to interpret against published reference ranges. The regulatory scope is deliberately narrow and matters throughout: the device is cleared to measure and report tissue properties as an aid to the clinician, not to diagnose, stage, or grade liver disease, and not to operate autonomously. It is used across hepatology clinics, imaging research and — most visibly — as a quantitative endpoint in metabolic-liver-disease (MASLD/MASH) drug trials, where cT1 and PDFF track disease and treatment response without a biopsy. In every setting the numbers support, but do not replace, the reading clinician's judgement.