GLOSSARY

Imaging Biomarker

What is an imaging biomarker?

An imaging biomarker is a measurable characteristic extracted from a medical image that objectively indicates a biological process, disease state, or response to treatment — without requiring a tissue sample or invasive procedure.

Imaging biomarkers provide objective, reproducible measurements that reduce the subjectivity inherent in visual radiological assessment. In clinical trials they are used to confirm patient eligibility, track disease progression, and measure therapeutic response. The shift from qualitative radiological reads to quantitative imaging biomarkers is one of the defining trends in modern drug development.

Imaging biomarker — simple definition A quantitative measurement extracted from a medical image that indicates a biological process, disease state, or response to treatment.

How do imaging biomarkers differ from clinical endpoints and surrogate endpoints?

Understanding this distinction matters for trial design and regulatory planning.

An imaging biomarker is a measurement — the quantitative value derived from an image, such as a lesion volume in cubic millimetres or a brain atrophy rate in millilitres per year. The biomarker is the tool.

A clinical endpoint is the outcome the trial is designed to demonstrate — overall survival, functional decline, quality of life. Clinical endpoints measure what ultimately matters to the patient.

A surrogate endpoint is an imaging biomarker (or other measurement) accepted by regulators as a substitute for a clinical endpoint — because there is sufficient evidence that the biomarker predicts or is reasonably likely to predict clinical benefit. FDA recognises three tiers of surrogate endpoint: candidate (biological rationale but limited validation, used in exploratory contexts); reasonably likely (used to support accelerated approval, with post-market confirmation required); and validated (used for regular approval, supported by a substantial evidentiary base). Not every surrogate endpoint must be formally qualified through the Biomarker Qualification Program to be used in drug development — accelerated approval pathways allow biomarkers accepted as "reasonably likely" surrogates to support initial approval while confirmatory trials are completed.

The practical implication for trial design: most imaging biomarkers used in current trials are not formally validated surrogate endpoints and have not completed the full Biomarker Qualification Program. Many are used legitimately as exploratory, secondary, or supportive endpoints without surrogate qualification — and some have been accepted as reasonably likely surrogates for accelerated approval. For primary efficacy endpoints in submissions seeking regular approval, sponsors should discuss endpoint acceptability with FDA or EMA no later than end-of-phase-2 meetings.

What are the types of imaging biomarkers?

Imaging biomarkers are classified by their clinical function. The classification determines where in the trial the biomarker can serve as an endpoint.

Diagnostic biomarkers support the identification of a disease or its subtype. Amyloid PET positivity in Alzheimer's disease is a diagnostic imaging biomarker that is embedded in current clinical and research diagnostic frameworks — including the NIA-AA biological definition of Alzheimer's disease and FDA-approved treatment eligibility criteria. The Central Vein Sign in multiple sclerosis is a diagnostic imaging biomarker proposed for incorporation into updated MS diagnostic criteria; research indicates it can improve diagnostic specificity, though further prospective validation is ongoing.

Prognostic biomarkers predict the likely course of disease independent of treatment. Brain volume loss rate in neurodegeneration is a prognostic biomarker — it indicates how aggressively a disease is progressing regardless of which therapy a patient receives.

Predictive biomarkers indicate whether a specific patient is likely to respond to a specific treatment. Tumour volumetrics at baseline can predict response to immunotherapy in certain cancers.

Pharmacodynamic biomarkers measure whether a drug is having its intended biological effect. Changes in lesion burden following treatment with a disease-modifying therapy confirm target engagement.

Surrogate endpoints replace a clinical outcome — such as survival or functional decline — with an imaging-derived measurement accepted by regulators as a valid proxy. This is the most demanding classification: the biomarker-to-outcome relationship must be demonstrated in regulatory-grade evidence, and formal qualification is required before the measurement can serve as a primary regulatory endpoint.

Planning a trial that uses imaging biomarkers as endpoints?
QMENTA's imaging scientists can advise on biomarker selection, analytical validation requirements, and platform configuration for your protocol.
Speak to an imaging expert →

Imaging biomarker vs. digital biomarker — what is the difference?

These terms are often used interchangeably but refer to different concepts, and the distinction matters for regulatory classification, reimbursement, and validation requirements.

An imaging biomarker is derived from a medical image — an MRI, CT, PET, or OCT scan — and requires specialised imaging equipment and analysis software. The measurement is typically performed at a clinical site or core laboratory.

A digital biomarker is derived from data collected by a digital device — a smartphone, wearable, or sensor — typically during routine daily activity. Gait speed measured by a phone accelerometer and heart rate variability from a wearable are digital biomarkers.

The two categories can overlap: a measurement derived from a smartphone camera analysed for signs of neurodegeneration is both a digital biomarker (generated by a consumer device) and an imaging biomarker (derived from image analysis). Regulatory guidance from both FDA and EMA is still evolving on how to classify hybrid cases.

How are imaging biomarkers validated?

Regulatory acceptance of an imaging biomarker as a trial endpoint requires a formal validation process. The FDA's Biomarker Qualification Program and EMA qualification procedures provide the framework, but validation takes years and requires data from multiple independent cohorts.

The core validation steps are:

1. Analytical validation — confirming the biomarker can be measured accurately, reproducibly, and consistently across different scanners, sites, and time points
2. Clinical validation — demonstrating that the biomarker measurement correlates with a clinically meaningful outcome in the target patient population
3. Qualification — formal regulatory acceptance of the biomarker for use in a specific context of use, defining the disease, patient population, and endpoint application it can support

Most imaging biomarkers used in current trials have completed analytical and clinical validation but are not formally validated surrogate endpoints under the full Biomarker Qualification Program. They are used legitimately as exploratory, secondary, or supportive endpoints — and some have been accepted as reasonably likely surrogates under accelerated approval pathways. Use as a primary efficacy endpoint for regular approval requires a more substantial evidentiary base and typically requires prior discussion with the regulatory agency.

What imaging biomarkers are used in neurological clinical trials?

Neurological diseases have driven the most rapid adoption of imaging biomarkers in clinical research, because the brain is not accessible for repeated biopsy and imaging is the only practical way to monitor disease progression longitudinally.

Key imaging biomarkers in current neurological trials include:

• White matter lesion volume — used in multiple sclerosis trials to measure disease activity and treatment response; assessable via central review using standardised AI analysis tools
• Brain atrophy rate — measured by tools including SIENAX and FreeSurfer; used across Alzheimer's, Parkinson's, and MS trials as a sensitive longitudinal marker
• Amyloid PET burden — used in Alzheimer's trials to confirm amyloid pathology at baseline and measure clearance with anti-amyloid therapies
• Central Vein Sign — a novel MRI biomarker incorporated into the 2025 McDonald Criteria for MS diagnosis
• DAT-SPECT binding ratios — used in Parkinson's disease trials to measure dopaminergic neuronal integrity

QMENTA's Imaging Hub provides over 50 AI imaging biomarker analysis tools across neurology, oncology, and cardiology — including open-source algorithms such as FreeSurfer, FSL, and SIENAX, and validated proprietary biomarkers from partner research institutions.¹ Algorithm versions are locked at trial initiation to prevent drift across sites.

Key takeaways

• An imaging biomarker is a quantitative measurement extracted from a medical image — not a qualitative radiological observation
• Imaging biomarkers are classified by function: diagnostic, prognostic, predictive, pharmacodynamic, and surrogate endpoints
• Surrogate endpoint is a regulatory concept with three tiers under FDA's framework — most imaging biomarkers are used as exploratory or secondary endpoints without formal surrogate qualification; some support accelerated approval as "reasonably likely" surrogates
• Imaging biomarkers differ from digital biomarkers in data source: imaging requires specialist equipment; digital biomarkers use consumer devices
• Validation has three steps: analytical validation, clinical validation, and regulatory qualification — only the last enables primary endpoint use
• QMENTA's Imaging Hub provides over 50 AI imaging biomarker tools deployable centrally across multi-site trials with version-locked algorithms
  

By Paulo Rodrigues, PhD, Chief Technology Officer and Co-Founder at QMENTA
Paulo Rodrigues leads technology strategy at QMENTA and writes about imaging clinical trials, protocol standardization, real-time QC, and compliance-ready neuroimaging workflows for multi-site studies. View executive leadership.

¹ QMENTA. AI Imaging Biomarker Catalog. qmenta.com/imaging-hub/ai-imaging-biomarkers See also: QMENTA. 2025 Year in Review. qmenta.com/blog/qmenta-2025-year-in-review — for the McDonald Criteria 2025 contribution reference.