Clinical biomarkers are increasingly shaping how drug developers design trials, seek accelerated approval, and argue that a biological signal can predict meaningful patient benefit. Yet the journey from biomarker discovery to regulatory-grade surrogate endpoint remains slow, expensive, and scientifically unforgiving. The United States Food and Drug Administration and the European Medicines Agency have both created structured pathways for using biomarkers and novel methodologies in medicine development, but neither regulator treats biological plausibility as a substitute for clinical validation. That tension is now central to drug development in oncology, neurodegeneration, metabolic liver disease, rare genetic disorders, and immune-mediated diseases.
The issue is simple to state and difficult to solve. A biomarker can show that a drug is hitting a target, changing a pathway, shrinking a lesion, reducing a protein, or improving a laboratory value. A surrogate endpoint must do more. It must be sufficiently reliable to stand in for how a patient feels, functions, or survives. That distinction explains why many biomarkers become useful development tools, while only a smaller group advance far enough to support drug approval, accelerated approval, label expansion, payer confidence, and guideline acceptance.
Why does FDA distinguish between a biomarker signal and a surrogate endpoint that can support approval?
FDA’s public surrogate endpoint resources describe surrogate endpoints as measures that may substitute for direct clinical outcomes when those outcomes would take too long to observe or when the relationship between the surrogate and clinical benefit is already well understood. The agency’s accelerated approval pathway allows earlier approval of drugs for serious conditions with unmet medical need when a surrogate endpoint is reasonably likely to predict clinical benefit. That phrase is doing heavy regulatory lifting. It allows earlier access, but it also leaves sponsors with the obligation to verify benefit through confirmatory trials.
This is where the slow march begins. Biomarkers often look attractive in early development because they are measurable, repeatable, and mechanistically close to the drug target. Regulators, however, need to know whether changing that biomarker changes patient outcomes. A tumour response rate may matter if durable response correlates with survival or symptom relief in a defined cancer setting. Amyloid plaque reduction may matter in Alzheimer’s disease only if clinical decline is also slowed in a robust trial population. Fibrosis improvement in metabolic dysfunction-associated steatohepatitis may matter only if it eventually translates into fewer cases of cirrhosis, liver decompensation, transplantation, or death.
For sponsors, this distinction affects almost every major decision. It shapes trial size, follow-up duration, assay validation, statistical design, regulatory meeting strategy, and post-approval evidence planning. For patients, it determines whether earlier access arrives with enough confidence that the biomarker is more than a shiny laboratory signal. For payers, it defines whether a high-cost therapy is supported by clinical value or only by a biologically plausible bet.
How recent FDA examples show both the promise and the risk of surrogate endpoint approvals
The modern surrogate endpoint debate is not theoretical. It is already visible across therapeutic areas. Madrigal Pharmaceuticals’ Rezdiffra, or resmetirom, became the first FDA-approved therapy for adults with noncirrhotic nonalcoholic steatohepatitis with moderate to advanced fibrosis. The label states that approval was granted under accelerated approval based on improvement in nonalcoholic steatohepatitis and fibrosis, with continued approval contingent on confirmatory clinical benefit.
That decision mattered because metabolic liver disease had long struggled with trial designs that required histologic endpoints, invasive biopsies, and long-term clinical outcome studies. Rezdiffra’s approval showed that regulators can accept earlier biological and histological evidence where disease burden is high and development barriers are substantial. But it also showed the commercial catch: accelerated approval can open the market, while still leaving payers, hepatologists, and guideline committees to ask how strongly histological improvement predicts long-term outcome reduction.
Multiple myeloma offers another important signal. FDA’s January 2026 guidance on minimal residual disease and complete response in multiple myeloma provides recommendations for using minimal residual disease and complete response as endpoints in trials intended to support accelerated approval. This is significant because minimal residual disease is one of the most closely watched biomarker-based endpoints in haematology. It can detect deep response earlier than progression-free survival or overall survival, but the regulatory problem is not whether the assay is sensitive. The problem is whether a given threshold, timing, testing platform, and treatment context can reliably predict durable clinical benefit.
That is a critical commercial distinction. If minimal residual disease becomes more widely accepted as a surrogate endpoint in multiple myeloma, sponsors could shorten development timelines, test combinations faster, and move promising regimens into earlier lines of therapy more efficiently. But weaker assay harmonisation or inconsistent correlation with long-term outcomes could create the opposite result: more regulatory caution, more payer scepticism, and more demand for mature progression-free survival or overall survival data.
Why Alzheimer’s disease turned amyloid into the most politically sensitive surrogate endpoint debate
Alzheimer’s disease has become the cautionary tale for biomarker-driven regulation. The approval of anti-amyloid therapies placed amyloid plaque reduction under intense scrutiny because it raised the question of whether removing a pathological hallmark of disease reliably predicts clinically meaningful slowing of decline. The controversy around Biogen’s Aduhelm showed the reputational risk of approving a therapy where the surrogate endpoint did not carry broad confidence among clinicians, payers, and advisory committee members.
The later trajectory of Eisai and Biogen’s Leqembi and Eli Lilly’s Kisunla shifted the debate by tying amyloid reduction to clinical outcome data showing slowing of cognitive and functional decline in early Alzheimer’s disease populations. Even so, these drugs did not erase the surrogate endpoint controversy. They reframed it. The question moved from whether amyloid matters at all to when amyloid reduction is sufficient, which patient populations benefit most, how safety risks such as amyloid-related imaging abnormalities should be managed, and whether health systems can absorb the diagnostic, infusion, and monitoring burden.
This is where biomarker validation becomes a market access issue. A biomarker may satisfy a regulator in one setting, but payers may still restrict coverage if the clinical effect size is modest, monitoring costs are high, or real-world patient selection is difficult. Alzheimer’s disease has shown that the commercial value of a surrogate endpoint depends not only on regulatory acceptance, but also on clinician trust, payer policy, diagnostic infrastructure, and patient risk tolerance.
Why EMA’s qualification pathway shows how regulators want context, not generic biomarker claims
The European Medicines Agency’s qualification of novel methodologies pathway allows developers to seek regulatory acceptability for biomarkers, imaging methods, statistical approaches, digital measures, and other development tools within a specific context of use. That context matters. A biomarker may be useful for patient enrichment, dose selection, early pharmacodynamic response, safety monitoring, or disease progression modelling without being suitable as a primary efficacy endpoint.
This creates a practical hierarchy. The first step is often analytical validation: does the test measure what it claims to measure accurately and reproducibly? The second is clinical validation: does the biomarker correlate with disease state, progression, or treatment response in the intended population? The third is regulatory qualification: can it support a defined decision in drug development? Only after that does the more ambitious question arise: can it replace a clinical endpoint in a pivotal trial or support accelerated approval?
Sponsors sometimes want to move quickly from biological rationale to pivotal trial utility. Regulators usually do not. They want evidence that the biomarker behaves consistently across populations, interventions, laboratories, assays, and disease stages. That is why the march is slow. The risk of a false-positive surrogate is not just a failed trial. It can mean exposing patients to toxicity, allocating payer budgets to uncertain benefit, crowding out better therapies, and weakening confidence in regulatory science.
How oncology explains why some surrogate endpoints move faster than others
Oncology has historically been the most fertile field for surrogate and intermediate endpoints because waiting for overall survival can be impractical, ethically complex, or confounded by subsequent lines of therapy. FDA oncology guidance has recognised endpoints such as objective response rate, duration of response, progression-free survival, pathologic complete response, and disease-free survival in different contexts.
But oncology also shows why no surrogate endpoint is universally portable. Objective response rate may be compelling in a refractory cancer with few options if responses are deep and durable. The same response rate may be less persuasive in a crowded market if the effect is short-lived, toxic, or disconnected from survival. Pathologic complete response may support accelerated approval in some early breast cancer settings, but it does not automatically validate the same logic across every tumour type. Minimal residual disease may be powerful in haematologic malignancies, but it requires assay discipline and disease-specific validation.
The commercial implication is blunt. A sponsor cannot assume that one accepted endpoint creates a shortcut across an entire platform. Regulators increasingly ask whether the biomarker-endpoint relationship holds in the exact disease, line of therapy, mechanism, patient subgroup, and treatment setting being studied. That makes endpoint strategy a board-level issue, not a biostatistics footnote.
Why payers may become tougher than regulators on biomarker-based approvals
Regulatory approval is only one gate. Payer acceptance is becoming the second, and in some therapeutic areas it may become the harder one. Accelerated approval based on a surrogate endpoint can bring a drug to market before long-term clinical benefit is confirmed. That helps patients with serious diseases, but it also creates uncertainty around value, budget impact, and treatment sequencing.
Payers are likely to ask whether the surrogate endpoint has a clear relationship to outcomes that matter economically: fewer hospitalisations, delayed disease progression, reduced need for procedures, avoided transplantation, improved function, or longer survival. In rare diseases, they may accept greater uncertainty because patient populations are small and unmet need is high. In large chronic diseases such as Alzheimer’s disease, metabolic liver disease, obesity-related conditions, and cardiovascular disease, the threshold may be higher because even a modestly priced therapy can create enormous aggregate cost.
That is why confirmatory trial design now has commercial importance from day one. A sponsor that treats post-approval evidence as a regulatory afterthought may win early approval but lose payer confidence. A sponsor that designs confirmatory trials to answer clinician and payer questions can turn a surrogate endpoint approval into durable market access. In today’s environment, the endpoint does not just have to satisfy the regulator. It has to survive the formulary committee.
Why the next phase of biomarker validation will favour platforms with longitudinal evidence
The future of surrogate endpoints is likely to be shaped by longitudinal datasets, real-world evidence, digital monitoring, multi-omic profiling, and more precise patient stratification. Biomarkers that are measured once at baseline may still be useful, but dynamic biomarkers that show how disease changes over time and how treatment modifies that trajectory could become more valuable.
This is especially relevant in neurodegeneration, autoimmune disease, cardiometabolic disease, and chronic kidney disease, where clinical outcomes may take years to mature. Developers that can connect biomarker movement to disease progression, functional outcomes, healthcare utilisation, and survival will have a stronger case. Developers that merely show target engagement may still struggle.
The scientific bar will also rise for assay validation. Regulators will want confidence that a biomarker result is not dependent on a single laboratory, proprietary platform, or fragile threshold. If an endpoint is going to support approval, physicians and payers must understand how it can be measured consistently in real-world practice. That is particularly important for companion diagnostics, digital biomarkers, minimal residual disease testing, imaging-based endpoints, and blood-based markers in diseases where diagnostic pathways are still evolving.
What this means for drug developers trying to turn biomarkers into regulatory-grade endpoints
The main lesson for sponsors is that biomarkers should be developed as evidence assets, not decorative science. A biomarker strategy needs to begin early, be embedded across trial phases, and be aligned with regulators before pivotal decisions are locked. It should define the context of use, assay performance, clinical relevance, statistical analysis plan, patient population, and confirmatory evidence strategy.
This also means that biomarkers can fail commercially even when they succeed scientifically. A beautifully measured biomarker that does not influence prescribing, reimbursement, or patient outcomes may remain a development tool. A messier biomarker with strong linkage to clinical consequences may become a market-shaping endpoint. The difference lies in validation, not enthusiasm.
The slow march from clinical biomarker to surrogate endpoint is therefore not regulatory obstructionism. It is the price of replacing direct patient outcomes with earlier biological signals. Faster drug development is a worthy goal, especially in serious diseases with limited treatment options. But speed only helps if the signal is trustworthy. In the next generation of drug approvals, the winners will not be the companies with the most biomarkers. They will be the companies that can prove which biomarkers truly matter.
Key takeaways: why surrogate endpoint validation is becoming a strategic drug development battleground
- The FDA’s accelerated approval pathway allows earlier drug approval when a surrogate endpoint is reasonably likely to predict clinical benefit, but confirmatory trials remain central to maintaining confidence in the approval.
- Clinical biomarkers are not automatically surrogate endpoints. They must show a reliable relationship with patient-relevant outcomes such as survival, function, symptoms, progression, or irreversible morbidity.
- Madrigal Pharmaceuticals’ Rezdiffra approval in nonalcoholic steatohepatitis showed how histological improvement can support accelerated approval, while still leaving long-term clinical benefit to be verified.
- FDA’s 2026 multiple myeloma guidance on minimal residual disease and complete response signals growing regulatory interest in earlier endpoints, but also raises expectations around assay validation and trial design.
- Alzheimer’s disease remains the most visible example of how biomarker-driven approvals can trigger payer, clinician, and public scrutiny when the link between biological effect and clinical benefit is contested.
- The European Medicines Agency’s qualification pathway reinforces that biomarkers must be accepted for a specific context of use, not as generic regulatory shortcuts.
- Oncology will remain the most active field for surrogate endpoint development, but regulators are unlikely to allow endpoint logic to transfer automatically across tumour types, mechanisms, or treatment settings.
- Payers may become tougher than regulators where surrogate endpoint approvals carry high budget impact, uncertain durability, or heavy monitoring requirements.
- Longitudinal real-world evidence and harmonised assays will become increasingly important as sponsors try to convert biomarkers into regulatory-grade endpoints.
- The companies best positioned in this shift will be those that treat biomarker validation as a core evidence-generation strategy, not as a late-stage regulatory rescue tool.
