1 Definition and scope
An imaging biomarker is a measurable feature observed on a medical image that can indicate a normal biological process, a disease process, or a response to treatment. The feature may be qualitative or quantitative and may reflect anatomy, tissue composition, physiology, or molecular activity. Imaging biomarkers are used in both routine care and research, where they can support diagnosis, staging, prognosis, therapy selection, and follow-up.
1.1 Biomarker terminology
In general biomedical usage, a biomarker is any objectively measured characteristic that serves as an indicator of normal biology, pathology, or response to an intervention. Imaging biomarkers are a subgroup of biomarkers derived from visualized data rather than blood, tissue, or other laboratory specimens. They may be anatomical, functional, or molecular, depending on the information extracted from the image.
1.2 Imaging-based measurement
Imaging biomarkers are obtained from modalities such as X-ray, computed tomography, magnetic resonance imaging, ultrasound, positron emission tomography, and single-photon emission computed tomography. Measurements can include size, shape, signal intensity, tracer uptake, blood flow, diffusion characteristics, or derived computational features. The reliability of the measurement depends on acquisition technique, analysis method, and the biological stability of the feature being assessed.
1.3 Clinical and research roles
In clinical practice, imaging biomarkers help clinicians detect disease, estimate severity, guide intervention, and monitor change over time. In research, they are used as surrogate or exploratory endpoints, as well as tools for patient stratification and drug development. A useful biomarker should be reproducible, interpretable, and linked to a clinically relevant outcome.
2 Types of imaging biomarkers
Imaging biomarkers are commonly grouped according to the kind of biological information they represent. Some provide structural information, while others capture physiology, cellular behavior, or molecular processes. Many modern applications combine several biomarker types within a single examination.
2.1 Structural biomarkers
Structural biomarkers describe visible anatomical features on imaging. They are often the most familiar type and include measurements of organ size, tissue architecture, and lesion morphology. These features are widely used because they are comparatively easy to visualize and quantify.
2.1.1 Organ size and morphology
Changes in organ size, contour, or internal architecture can indicate disease presence or progression. Examples include enlargement of the heart, atrophy of the brain, or distortion of an organ by a mass. Morphologic assessment is especially useful in chronic diseases where structural change accumulates over time.
2.1.2 Lesion detection and characterization
Imaging can identify focal abnormalities and help distinguish benign from malignant or active from inactive lesions. Characteristics such as margins, density, signal pattern, calcification, enhancement, and internal heterogeneity may provide diagnostic clues. These features often support further testing, biopsy, or surveillance.
2.2 Functional biomarkers
Functional biomarkers reflect how tissues or organs work rather than how they look. They are often derived from dynamic or specialized imaging protocols and can reveal abnormalities before obvious structural change appears.
2.2.1 Perfusion and blood flow
Perfusion imaging measures blood delivery to tissue and can indicate ischemia, inflammation, tumor vascularity, or treatment response. Techniques may estimate flow, blood volume, or transit time. These measurements are useful in organs where circulation strongly influences function.
2.2.2 Diffusion and cellularity
Diffusion-based imaging, especially in MRI, assesses the movement of water molecules within tissue. Restricted diffusion may suggest high cellular density, as seen in some tumors or acute injury. Such measures can provide insight into microstructural organization and tissue integrity.
2.3 Molecular biomarkers
Molecular imaging biomarkers show biological activity at the level of receptors, enzymes, metabolism, or other specific targets. They are often obtained with labeled tracers in nuclear medicine or with contrast agents designed to probe particular pathways.
2.3.1 Receptor expression
Some imaging studies measure the distribution or density of receptors on cell surfaces or within tissues. This can help determine whether a lesion expresses a target relevant to therapy. Receptor-based imaging is especially valuable in personalized treatment planning.
2.3.2 Metabolic activity
Metabolic imaging captures processes such as glucose utilization, oxygen consumption, or amino acid transport. Increased or altered metabolic activity may signal tumor growth, infection, or tissue stress. Because metabolic changes can occur early, these biomarkers may detect disease before anatomic change becomes pronounced.
2.4 Quantitative imaging biomarkers
Quantitative imaging biomarkers are numerical features extracted from images by measurement or computation. They are intended to reduce subjectivity and improve comparability across time, patients, and institutions. Their usefulness depends on consistency in acquisition and analysis.
2.4.1 Volumetric analysis
Volumetric analysis measures the volume of an organ, lesion, or subregion. It is more informative than single diameter measurements in many settings because it better captures irregular shapes and change over time. Serial volume assessment is widely used in oncology, neurology, and organ transplantation.
2.4.2 Radiomics features
Radiomics refers to the extraction of a large number of quantitative image features, such as texture, shape, and intensity distribution. These features are then analyzed statistically or with machine learning methods to identify patterns associated with disease behavior or outcome. Radiomics remains an active research area, with ongoing work on standardization and validation.
3 Imaging modalities used
Different imaging modalities provide different kinds of biomarker information. The best choice depends on the target organ, clinical question, required sensitivity, and acceptable burden to the patient. Some modalities are best for anatomy, while others excel at function or molecular characterization.
3.1 X-ray and mammography
X-ray imaging provides rapid, widely available structural information, especially for bones, chest, and certain soft tissue findings. Mammography is a specialized X-ray technique used to assess breast tissue and detect lesions or microcalcifications. These methods are often used for screening and first-line evaluation.
3.2 Computed tomography
Computed tomography produces cross-sectional images with excellent spatial resolution and is commonly used for detecting masses, trauma, vascular disease, and lung abnormalities. Quantitative CT measures can include lesion attenuation, calcification, and volume. It is often useful when precise anatomy is required.
3.3 Magnetic resonance imaging
Magnetic resonance imaging offers high soft-tissue contrast and multiple sequence types that can probe anatomy, diffusion, perfusion, and chemical composition. It is especially valuable in the brain, musculoskeletal system, pelvis, and liver. MRI supports both structural and functional biomarker development.
3.4 Ultrasound
Ultrasound is portable, relatively low cost, and capable of real-time imaging. It can assess anatomy, motion, blood flow, and tissue stiffness using specialized techniques such as Doppler and elastography. Its performance may depend strongly on operator skill and patient factors.
3.5 Positron emission tomography
Positron emission tomography detects radiotracer distribution and is widely used for molecular and metabolic imaging. It can reveal abnormal uptake linked to tumor activity, inflammation, or specific biological pathways. PET is often combined with CT or MRI to improve localization.
3.6 Single-photon emission computed tomography
Single-photon emission computed tomography uses gamma-emitting radiotracers to image physiologic or molecular processes. It is commonly applied in cardiology, bone imaging, and certain neurologic or oncologic studies. Although generally less sensitive than PET, it remains clinically important and broadly available.
4 Clinical applications
Imaging biomarkers have a broad range of clinical uses. They can support initial diagnosis, define disease extent, estimate likely course, and help determine whether treatment is working. Their value increases when the measured feature is closely linked to the underlying disease process.
4.1 Disease detection
A biomarker may identify abnormal tissue before symptoms become severe or before the disease is obvious on examination. Screening and early detection depend on markers that are sensitive enough to identify subtle changes while maintaining acceptable specificity. Detection is often the first step toward further characterization.
4.2 Staging and classification
Staging describes the extent of disease, while classification groups lesions or patients into categories with shared characteristics. Imaging biomarkers can help determine local spread, nodal involvement, organ burden, or structural severity. Accurate staging is essential for consistent treatment decisions and prognostic estimates.
4.3 Prognostic assessment
Some imaging features are associated with future disease course, including survival, recurrence, or functional decline. Prognostic biomarkers may reflect aggressiveness, extent of invasion, or residual functional reserve. Their value lies in helping clinicians and patients anticipate risk.
4.4 Treatment selection and planning
Imaging biomarkers can influence the choice of therapy by identifying target expression, tumor extent, or anatomic constraints. They are also used to plan surgery, radiation, and interventional procedures. In some contexts, biomarker findings help avoid ineffective treatment by showing a lack of expected target.
4.5 Therapy monitoring
Serial imaging allows comparison of biomarker values over time to determine whether a disease is stable, improving, or worsening. Monitoring can reveal response earlier than symptoms or laboratory tests in some conditions. Consistent follow-up technique is important for reliable interpretation.
4.6 Outcome prediction
Outcome prediction uses imaging-derived information to estimate the probability of future events or treatment benefit. This may include the chance of recurrence, progression, complications, or durable response. Predictive use is especially important in precision medicine, where therapy is matched to biological characteristics.
5 Validation and qualification
Before an imaging biomarker is adopted widely, it must be shown to measure what it claims to measure and to do so in a dependable way. Validation is a multi-step process that links the image feature to biological truth and clinical usefulness. Qualification may be formal or informal depending on the setting.
5.1 Analytical validity
Analytical validity refers to whether the measurement is accurate, precise, and technically sound. It addresses issues such as scanner performance, acquisition consistency, and algorithm reliability. A biomarker with poor analytical validity is unlikely to be useful, even if it appears biologically plausible.
5.2 Clinical validity
Clinical validity is the degree to which the imaging biomarker correlates with a clinically meaningful condition or outcome. For example, a lesion feature may correlate with pathology, disease stage, or recurrence risk. This relationship must be demonstrated in appropriate patient populations.
5.3 Clinical utility
Clinical utility asks whether using the biomarker improves patient care, decision-making, or health outcomes. A biomarker may be valid yet still have limited utility if it does not change management or if alternative methods are better. Utility is often the most important criterion for routine adoption.
5.4 Standardization and reproducibility
Standardization reduces variation caused by scanner settings, contrast timing, reconstruction methods, and analysis workflows. Reproducibility means that repeated measurements produce similar results under comparable conditions. High reproducibility is essential for longitudinal monitoring and multicenter research.
6 Data analysis and interpretation
The value of an imaging biomarker depends not only on what is seen but also on how the data are acquired, processed, and interpreted. Small differences in technique can influence the final measurement. Careful workflow design helps limit error and improve confidence in the results.
6.1 Image acquisition protocols
Acquisition protocols define how images are collected, including patient preparation, timing, contrast use, and scanner parameters. Consistent protocols are important because changes in technique can alter biomarker values. Protocol harmonization is particularly important in studies that compare multiple sites or repeated scans.
6.2 Segmentation and region of interest selection
Segmentation is the process of outlining a structure or lesion for measurement. Region of interest selection determines which part of the image is analyzed and can strongly affect the result. Manual, semi-automatic, and automatic methods each have advantages and limitations.
6.3 Quantification methods
Quantification methods convert image information into measurable values. These may include diameters, volumes, signal ratios, standardized uptake values, perfusion indices, or texture metrics. The chosen method should match the clinical question and should be defined clearly to allow comparison across studies.
6.4 Interobserver and intraobserver variability
Interobserver variability refers to differences between readers, while intraobserver variability refers to differences when the same reader repeats the measurement. Both can affect confidence in an imaging biomarker. Training, clear definitions, and automated tools can reduce these differences.
6.5 Artificial intelligence and machine learning
Artificial intelligence and machine learning are increasingly used to identify patterns, automate segmentation, and predict outcomes from imaging data. These methods can uncover complex relationships that are difficult to detect by visual inspection alone. However, their performance depends on training data quality, external validation, and careful handling of bias.
7 Advantages and limitations
Imaging biomarkers offer major practical advantages, but they also face technical and biological constraints. Their strengths make them useful in many settings, yet limitations must be recognized to avoid overinterpretation. Balanced evaluation is essential for both clinical and research use.
7.1 Noninvasive assessment
A major advantage of imaging biomarkers is that they can assess disease without the need for invasive sampling. This is especially important when tissue biopsy is risky, impractical, or unable to capture disease heterogeneity. Repeated measurements can also be obtained more easily than repeated tissue sampling.
7.2 Accessibility and cost
Some imaging methods are widely available and relatively affordable, while others require specialized equipment, tracers, or expertise. Cost, travel burden, appointment time, and radiation exposure can all influence feasibility. Accessibility varies across institutions and health systems.
7.3 Sources of error and bias
Errors may arise from motion, poor image quality, partial volume effects, reconstruction differences, or inconsistent interpretation. Selection bias and verification bias can also affect research findings when imaging outcomes are compared unevenly with reference standards. Careful study design helps limit these problems.
7.4 Biological and technical variability
Biological variability includes normal fluctuations in physiology, disease heterogeneity, and changes unrelated to the condition of interest. Technical variability arises from scanner differences, calibration drift, and analysis inconsistency. Both kinds of variability can reduce the apparent reliability of a biomarker.
8 Regulatory and research considerations
Because imaging biomarkers may influence diagnosis or treatment decisions, they are subject to scientific scrutiny and, in some settings, formal oversight. Research use often precedes routine implementation. Clear standards help ensure that results are trustworthy and interpretable.
8.1 Biomarker endpoints in trials
Imaging biomarkers are often used as endpoints in clinical trials to measure disease burden, biological response, or progression. They may serve as primary, secondary, or exploratory outcomes depending on the study design. When used as surrogate endpoints, they require strong evidence linking them to true clinical benefit.
8.2 Reporting standards
Structured reporting improves clarity, consistency, and reproducibility. Standards may specify how the image was acquired, how measurements were derived, and what thresholds or criteria were applied. Transparent reporting makes studies easier to compare and reproduce.
8.3 Multicenter studies
Multicenter studies increase sample size and generalizability, but they also introduce variability across scanners, protocols, and readers. Harmonization efforts are therefore critical, including quality control, shared definitions, and centralized analysis when appropriate. These studies are especially important for validation.
8.4 Future directions
Future work in imaging biomarkers is likely to emphasize integrated data analysis, improved automation, and stronger links between image features and biology. Advances in hybrid imaging, standardized computational pipelines, and robust external validation may expand clinical use. Continued collaboration between clinicians, physicists, and data scientists will remain important.