On this page
A biomarker in drug development is a measurable biological characteristic that predicts treatment response or disease state. Biomarkers identify which patients are most likely to benefit from a drug, serve as early indicators of efficacy and safety, and can become required diagnostic tests after approval. They are not clinical outcomes themselves but predictive signals that guide development decisions.
- Biomarkers predict treatment response or disease course; clinical outcomes are what patients actually experience, and biomarkers are surrogate measures that may or may not correlate with final outcomes.
- Prognostic biomarkers predict disease trajectory independent of treatment, predictive biomarkers indicate response to a specific drug, and pharmacodynamic biomarkers show a drug is hitting its intended target.
- Between 2010 and 2012, the FDA approved 45 percent of new drugs based on a surrogate endpoint, and the agency maintains a table of over 200 surrogate markers accepted to support drug approval [F4, F6].
- Biomarker failure is pervasive: of 883 EMA assessment reports, only 37 predictive biomarkers for 41 drugs were identified, with most candidates failing early in discovery or analytical validation.
- A biomarker becomes qualified for regulatory use through a formal three-stage submission process under the 21st Century Cures Act, requiring analytical validation, clinical validation, and demonstration of benefit over current methods [F1, F2, F3].
Biomarkers appear throughout drug development filings and clinical trial reports, but their role is often misunderstood. A biomarker is a measurable biological indicator, a protein level in blood, a genetic mutation, an imaging finding, that predicts how a patient will respond to treatment or how their disease will progress. Biomarkers are not clinical outcomes. A clinical outcome is what patients experience: did the drug reduce their symptoms, extend their life, or prevent disease? A biomarker is a signal that suggests an outcome might follow.
Biomarkers function as decision tools in the development pipeline. They identify which patients are most likely to benefit from a drug before enrollment, they serve as early indicators of whether a drug is working before waiting for long-term clinical data, and they can become required diagnostic tests that determine which patients receive a drug after approval. Understanding what biomarkers are, how they differ from clinical outcomes, and why they matter is essential for reading trial designs, regulatory filings, and the reasoning behind development decisions.
What a biomarker is in drug development
A biomarker is a measurable biological characteristic that can be detected and quantified in a patient's body. It may be a protein concentration in blood serum, a genetic sequence variant, a metabolite level, an imaging finding, or any other observable biological marker. The defining feature is that it predicts something: how a patient will respond to a treatment, whether disease will progress, or whether a drug is causing harm.
Biomarkers are not diagnoses and not treatments. They are predictive signals. A high cholesterol level is a biomarker for cardiovascular risk, but measuring cholesterol does not treat heart disease. A mutation in a cancer cell's DNA repair gene is a biomarker that predicts sensitivity to certain chemotherapies, but identifying the mutation does not cure the cancer. The biomarker informs a decision about which patients should receive which drugs, and whether those drugs are working.
When a biomarker has undergone formal regulatory validation, the FDA calls it qualified. Qualification means the biomarker has been studied rigorously enough that regulators are confident in its interpretation and can rely on it in drug development and approval decisions. Not all biomarkers are qualified; many are used in research or early trials without formal regulatory endorsement.
How biomarkers work in the drug development process
Biomarkers serve three concrete functions across the drug development pipeline: they identify which patients to enroll in trials, they provide early evidence that a drug is working, and they monitor for safety signals. These roles reduce development risk and accelerate the path to approval by allowing developers to focus resources on patients most likely to benefit and to detect efficacy or harm before waiting years for clinical outcomes to accumulate.
Consider a hypothetical oncology drug designed to target a specific genetic mutation in tumor cells. The mutation is a biomarker, it predicts which patients' cancers will respond to the drug. Rather than enrolling all cancer patients, developers can screen for the mutation first, enroll only patients who carry it, and measure whether their tumors shrink. This enrichment strategy increases the probability that the drug will show benefit in a smaller, faster trial. Without the biomarker, the same drug might fail because it would be diluted across patients whose tumors lack the target, masking the signal in the responsive subgroup.
Patient selection and stratification
Biomarkers identify which patients are most likely to respond to a drug before enrollment. This is called patient stratification or enrichment. A developer measures a biomarker in a population, identifies patients with a specific biomarker profile, and enrolls only those patients in the trial.
This approach reduces trial failure risk. If a drug works only in patients with a particular genetic variant, inflammatory marker level, or disease subtype, enrolling the entire population will dilute the drug's apparent effect. The drug may work brilliantly in the responsive subgroup but appear ineffective overall because the non-responsive majority masks the signal. Biomarker-driven stratification solves this problem by ensuring the trial population enriches for responders. The result is a smaller, faster, and more statistically efficient trial that is more likely to show benefit if the drug truly works in its intended population.
Measuring drug efficacy and safety
Biomarkers provide early indicators of whether a drug is working or causing harm, often months or years before clinical outcomes become visible. This matters because clinical outcomes, survival, disease progression, symptom relief, take time to accumulate. A biomarker can shift much faster.
Suppose a new heart failure drug is designed to reduce a specific inflammatory protein in blood. Developers can measure that protein weeks after patients start treatment. If the protein level drops sharply in treated patients but not in controls, that biomarker signal suggests the drug is hitting its intended mechanism. Biomarkers can provide an early read on mechanism and efficacy during drug development. This allows trials to be stopped early if the biomarker signal is strong, or halted if the signal is absent or reversed (suggesting no benefit or potential harm). Biomarkers thus compress the timeline for detecting both efficacy and safety problems.
Companion diagnostics and approval
A biomarker can become a required diagnostic test after drug approval. This test, called a companion diagnostic, determines which patients are eligible to receive the drug in clinical practice. Once a drug is approved with a companion diagnostic, prescribers must order the test before dispensing the drug to confirm the patient has the biomarker that predicts response.
This approach protects patients by ensuring they receive drugs only when evidence suggests they will benefit. It also protects the drug's market by concentrating its use in the population most likely to show clinical benefit, which supports real-world efficacy and safety monitoring. Companion diagnostics are especially common in oncology, where genetic and protein biomarkers in tumors often determine which patients respond to targeted therapies.
Biomarkers versus clinical outcomes
The most important distinction in drug development is the difference between a biomarker and a clinical outcome. This distinction is often blurred in trial reports and regulatory filings, and misunderstanding it leads to confusion about what a trial actually proved.
A clinical outcome is what a patient experiences: their symptoms improve, their disease does not progress, they live longer, or they experience fewer adverse events. These are the outcomes that matter to patients. A biomarker is a measurable biological signal that is believed to predict whether a clinical outcome will follow. Blood pressure is a biomarker; the clinical outcome is stroke prevention. A tumor shrinkage on imaging is a biomarker; the clinical outcome is prolonged survival. Cholesterol level is a biomarker; the clinical outcome is reduced heart attack risk.
The critical tension is that a biomarker may predict a clinical outcome in theory but fail to do so in practice. A drug might lower a biomarker dramatically without improving the clinical outcome. This happens because biology is complex and our understanding of disease mechanisms is incomplete. A biomarker that correlates with disease in one population may not do so in another. For example, many clinical trials using blood pressure lowering medications have demonstrated that reducing systolic blood pressure reduced the risk of stroke, establishing blood pressure as a validated surrogate for stroke risk. But not all biomarkers have this level of evidence. When regulators approve a drug based on a biomarker that has not been fully validated against clinical outcomes, they are making a bet that the biomarker will predict benefit. That bet sometimes pays off, and sometimes it does not.
Types of biomarkers in drug development
Biomarkers are categorized by their function in development and their relationship to treatment. Understanding these categories helps readers recognize biomarkers in trial designs and regulatory documents and understand what role each biomarker is playing.
Prognostic biomarkers
A prognostic biomarker predicts how a patient's disease will progress independent of treatment. It forecasts the natural course of disease in an untreated or standard-treated patient. Prognostic biomarkers identify high-risk and low-risk patients before any intervention.
In cancer, tumor stage, grade, and certain genetic mutations are prognostic biomarkers. They predict which patients are likely to relapse and which are likely to remain disease-free, regardless of which drug they receive. In heart disease, ejection fraction and troponin levels are prognostic biomarkers that predict risk of future cardiac events. Prognostic biomarkers are useful for patient stratification and for designing trials that enrich for patients at highest risk, where a new drug might have the greatest impact.
Predictive biomarkers
A predictive biomarker predicts response to a specific drug or class of drugs. Unlike prognostic biomarkers, which forecast disease course regardless of treatment, predictive biomarkers are treatment-specific. They identify which patients will benefit from a particular intervention.
The presence of a specific genetic mutation in a tumor that makes it sensitive to a targeted drug is a predictive biomarker. A patient with the mutation is predicted to respond; a patient without it is predicted not to respond. Predictive biomarkers are valuable for drug development because they allow enrichment of trial populations and companion diagnostic testing after approval.
Pharmacodynamic biomarkers
A pharmacodynamic biomarker shows that a drug is hitting its intended biological target, regardless of whether clinical benefit is yet visible. It is a direct measure of drug activity in the body. Pharmacodynamic biomarkers can confirm mechanism of action during drug development.
Suppose a new drug is designed to inhibit a specific enzyme involved in inflammation. Developers can measure the enzyme's activity in patient blood or tissue after dosing. If the enzyme activity drops sharply, the biomarker confirms the drug is engaging its target. This pharmacodynamic signal provides confidence that the drug is working as intended, even before clinical outcomes appear. Pharmacodynamic biomarkers are especially useful for go/no-go decisions in early development: if the biomarker signal is absent, the drug likely will not work, and development can be halted before investing in expensive Phase 3 trials.
Why biomarkers matter in drug development decisions
Biomarkers reduce uncertainty and risk at every stage of drug development. They allow developers to identify the patient population most likely to benefit before running large, expensive trials. They provide early evidence of efficacy and safety, compressing timelines and allowing go/no-go decisions based on biological signal rather than waiting for clinical outcomes to accumulate. They enable precision medicine by matching drugs to patients most likely to respond.
Biomarkers also reduce the number of patients who must be enrolled in trials. A trial enriched for biomarker-positive patients can be smaller and faster than an unselected trial, because the effect size in the responsive population is larger. This means fewer patients are exposed to an experimental drug, reducing safety risk. It also means fewer patients receive placebo when a drug might help them, an ethical advantage.
From a regulatory perspective, biomarkers allow the FDA to approve drugs based on surrogate endpoints rather than waiting for long-term clinical outcomes. Between 2010 and 2012, the FDA approved 45 percent of new drugs based on a surrogate endpoint. The FDA maintains a publicly available table of surrogate endpoints that have been accepted to support drug approval, which includes over 200 surrogate markers. These approvals accelerate patient access to new treatments, though they come with the understanding that the biomarker's correlation with clinical benefit must be monitored as the drug is used in broader populations.
Limitations and risks of biomarker-driven development
Biomarkers are predictive, not definitive. A strong biomarker signal does not guarantee clinical benefit. The core risk of biomarker-driven development is that a biomarker may correlate with disease or treatment response in one population or context but fail to do so in another. Developers and regulators must account for this uncertainty when making approval decisions and when monitoring drugs after they reach the market.
Validation and reproducibility
A biomarker must be validated across different populations and settings before it can be relied upon. Biomarkers may not generalize across different populations or contexts. Biomarker failure is pervasive in development: of 883 EMA assessment reports, only 37 predictive biomarkers for 41 drugs were identified, with most candidates failing early in discovery or analytical validation. Only a small number of clinically useful biomarkers have been successfully validated for routine clinical practice despite thousands of papers written in biomarker discovery projects. In cancer specifically, only about 0.1 percent of potentially clinically relevant biomarkers described in literature progress to routine clinical use.
These statistics reflect the difficulty of moving from biomarker discovery to clinical utility. A biomarker may perform well in a carefully controlled research setting but fail when applied to diverse patient populations in clinical practice. Analytical validation (confirming the assay is accurate and reproducible) is necessary but not sufficient. Clinical validation (confirming the biomarker predicts the intended outcome in real patients) is far more challenging and time-consuming.
Regulatory acceptance
Not all biomarkers are accepted by regulators as sufficient for drug approval. Some biomarkers are accepted as surrogate endpoints that can support approval on their own. Others are accepted only as supportive evidence that must be paired with clinical outcome data. Still others are not accepted at all because the evidence linking them to clinical benefit is insufficient.
Biomarker qualification under the 21st Century Cures Act involves a three-stage submission process to develop a biomarker for regulatory use. This formal process ensures that biomarkers accepted by the FDA have undergone rigorous evaluation. However, not every biomarker that shows promise in early trials will complete this process or receive qualification. Developers must present evidence that the biomarker has clinical validity, analytical validity, and utility, that it measures what it claims to measure, that it predicts the intended outcome reliably, and that it provides benefit over current methods. Meeting these standards requires substantial investment and often multiple clinical trials.
Biomarkers and precision medicine
Biomarkers are the foundation of precision medicine, the shift toward targeted therapies matched to individual patient characteristics. Rather than treating all patients with a disease the same way, precision medicine uses biomarkers to identify which patients are most likely to benefit from which drugs, then tailors treatment accordingly.
This approach is transforming drug development and clinical practice. In oncology, genetic biomarkers in tumors now drive treatment selection. In immunology and rheumatology, inflammatory biomarkers predict response to specific biologics. In cardiology, genetic and protein biomarkers inform risk stratification and drug choice. As biomarker science advances and assays become cheaper and faster, the ability to match drugs to patients will become routine.
Precision medicine also changes how drugs are developed and approved. Rather than developing a single drug for a broad disease category, developers increasingly identify a specific biomarker-defined subpopulation and design trials that enrich for that population. This makes trials smaller and faster, but it also means approved drugs may have narrower indications tied to specific biomarkers. Patients must be tested for the biomarker to determine eligibility. This requires coordination between drug developers, diagnostic companies, regulators, and clinicians to ensure that companion diagnostics are available, accurate, and accessible when a drug reaches the market.
Do this automatically
Measuring biomarkers across patient populations, analyzing genetic sequences, and running repeated assays throughout clinical trials requires substantial laboratory time and resources that scale with each new drug candidate and trial phase. Software systems can automate the collection, analysis, and interpretation of biomarker data to accelerate these measurements and reduce the manual effort required at each stage of drug development. Healthcare research reads ClinicalTrials.gov records, regulatory filings, and published literature directly, and cites each claim to the record it came from, across as many subjects as you need, with Mimir attaching a citation to every figure. It will not predict a trial's outcome or advise on anyone's treatment. Those decisions stay with clinicians and regulators, and deliberately so. Every plan includes it, including the free one.
Start your first report and run it on something you already know well.
This article is general information about healthcare and life-sciences topics. It is not medical advice, diagnosis, or treatment guidance, and it is not a substitute for a qualified professional. Any treatment decision should be discussed with a clinician.
