Biomarker Utilization for Regulatory Decision Making: A Landscape Analysis of Neurological Drug Products Approved by FDA (2008–2024)

Abstract

Over the past 15 years, biomarkers have evolved into a significant component of neurological drug development and regulatory evaluation, with expanded roles and increasing impact. This landscape analysis covers New Molecular Entity (NME) products approved by FDA from 2008 to 2024 for neurological diseases, which demonstrates the growing utilization of biomarkers in regulatory decision-making. This includes their use as surrogate endpoints, confirmatory evidence, and basis for dose selection. Our analysis suggested prominent roles of biomarkers in various therapeutic modalities including small molecules, oligonucleotides, and monoclonal antibodies, as well as in the context of rare or slowly progressive diseases. However, several challenges hinder the effective utilization of biomarker data, including lack of established clinical relevance, variability in data quality, and bioanalytical issues. Addressing these challenges will require cross-sector collaboration, rigorous analytical validation, and a clear demonstration of the linkage between biomarker changes and meaningful clinical benefits. By strengthening the evidentiary foundation of biomarker data, stakeholders can accelerate innovative drug development for neurological diseases and other therapeutic areas.

Neurological diseases are imposing an escalating burden on healthcare systems and devastating impact on patients and their caregivers worldwide. These diseases are often characterized by complex disease pathophysiology and heterogeneous clinical presentations and trajectories, creating significant challenges for drug development¹. Additionally, the high failure rate in clinical trials and prolonged drug development process reflect the urgent need for innovative approaches in this field².

Biomarkers, as measurable indicators of pathogenic processes or responses to therapeutic interventions, are a promising tool to accelerate drug development and to inform regulatory decision making³. Over the past 15 years, substantial scientific advancements have been achieved in the field of biomarker discovery. This progress has been driven by technological innovations, such as genomics, proteomics, imaging, and computational biology using artificial intelligence⁴. Concurrently, there has been a notable increase in the inclusion of biomarker data in FDA new drug and biologic applications, underscoring the growing importance of biomarkers in the drug development process⁵.

Incorporating biomarkers into drug development offers several advantages. Biomarkers can facilitate early disease detection, patient stratification, and real-time monitoring of therapeutic efficacy and safety⁵, and are generally less affected by placebo effect⁶. This is particularly beneficial in certain neurological indications, where defining clinical endpoints can be challenging and disease progressions are slow or variable⁷.

This landscape analysis summarizes the use of biomarkers in the regulatory approval of new molecular entity (NME) products, including NME New Drug Applications (NDAs) and original Biologics License Applications (BLAs) for neurological diseases approved from 2008 to 2024, based on FDA review documents and United States Prescribing Information (USPI). This analysis seeks to provide valuable insights that can inform future drug development strategies, assist in regulatory decision-making, and identify opportunities and challenges in leveraging biomarker data to facilitate drug development and approval.

Data Source, Analysis Approach, and Biomarker Selection Criteria

The analysis was conducted with the FDA Center for Drug Evaluation and Research (CDER)-approved neurology drugs and biologics classified as NMEs, which had biomarker data submitted and reviewed during the regulatory approval process. The NMEs approved from January 1, 2008 to December 31, 2024 for neurological indications were identified using data processed from the FDA’s internal Data Analysis Search Host (DASH) database. Neurological indications in this analysis were defined as those currently managed by Division of Neurology 1 (DN1) or Division of Neurology 2 (DN2) within the Office of Neuroscience (ON) at CDER’s Office of New Drugs.

Next, NMEs with biomarker-related information were identified through a keyword search in FDA documents available on Drugs@FDA⁸. The reviewed documents included USPI, summary reviews, clinical pharmacology reviews, and multidisciplinary/integrated reviews (Table S2), which summarize the basis for regulatory approval, including how biomarker data influenced the assessment of a drug’s benefit-risk profile and labeling recommendations. Keywords used in the search were ‘biomarker’, ‘marker’, and ‘pharmacodynamic’ (PD). The identified NMEs with biomarker data were then selected for further analysis, and the information relevant to biomarker was manually extracted. The roles of biomarkers were identified through a comprehensive review of the documents. Subsequently, all identified NMEs and the utilities of biomarkers were analyzed and visualized by trends, therapeutic indications, and drug modalities. The representative cases illustrating challenges associated with biomarker data were also summarized.

Between 2008 and 2024, a total of 67 NMEs were approved for the treatment of neurological diseases. Our analysis was focused on the 37 submissions which had biomarker data submitted and reviewed by FDA (Table S1). Twenty-nine of the 37 submissions incorporated biomarkers into their labeling, with 28 of them describing biomarkers in Section 12.2 Pharmacodynamics. By integrating biomarkers into the labeling, sponsors and regulatory authorities communicate critical, data-driven information that can influence clinical decision-making. Out of the 37 NMEs, 8 of them included more than one biomarker. These NMEs were often associated with conditions where the pathophysiology has not been fully elucidated while a significant unmet medical need persists. Representative biomarkers were selected for each NME for further analysis (Table S1), with a priority on those serving as surrogate endpoints or providing confirmatory evidence. This approach allowed us to focus on the primary drivers of regulatory decisions.

Increasing Trends of NME Approvals and Incorporation of Biomarker Data

There has been an increasing trend in the number of regulatory submissions that leverage biomarker data for multiple purposes, underscoring the evolving and expanding role of biomarkers in regulatory decision making. Among the 67 NMEs in our analysis, 25 of the 50 NDAs included biomarker data, whereas 12 of the 17 BLAs included biomarker data. Figure 1 reflects an evolution of biomarker data inclusion into regulatory submissions over time. Prior to 2015, the inclusion of biomarker data in regulatory submissions was relatively limited. A notable trend was the marked increase in NME approvals for neurological diseases since 2016, which was concurrent with a rising number of BLA/NDA submissions that included biomarker data (Figure 1). This period witnessed an increase in the approval of oligonucleotide therapies, including siRNA and antisense oligonucleotides, all of which utilized biomarker data to demonstrate efficacy. A similar trend was observed with monoclonal antibodies (mAb). The growing acceptance of biomarkers as surrogate endpoints and confirmatory evidence reflects increasing knowledge and confidence in their ability to predict clinical outcomes and inform therapeutic decisions. After 2021, the trend continued with an even greater emphasis on biomarker data.

Figure 1.

Approved Neurology NMEs With or Without Biomarker Data in FDA Reviews (2008–2024)

Meanwhile, the regulatory frameworks and descriptions pertaining to the roles of biomarker data in regulatory review documents have been evolving over the past two decades. Earlier submissions often lacked explicit language indicating the use of biomarkers as confirmatory evidence. In contrast, recent regulatory reviews frequently use “confirmatory evidence” to describe the role of biomarkers in supporting substantial evidence of effectiveness, sometimes further specified as “confirmatory evidence from pharmacodynamic/mechanistic data”. This transformation was driven by progress in biomarker discovery and validation, which led to increased availability of robust biomarker data and acceptance of using biomarkers during regulatory assessment. The FDA’s guidance on demonstrating evidence of effectiveness^{9 10 11} and biomarker qualification¹², along with the rise of model-informed drug development (MIDD) approaches, has further exemplified the importance of biomarkers in regulatory decision-making.

Roles of Biomarkers in Regulatory Decision-Making

Our analysis of FDA review documents elucidated multiple roles of biomarkers in regulatory decision-making, particularly as surrogate endpoints, confirmatory evidence, and basis for dose selection (Figure 2 and Table S1). Figure 2 illustrates the proportion of NMEs that employ biomarkers in these three major roles. The majority of NMEs integrate biomarkers in multiple capacities to support regulatory approval, with dose selection being the most frequently observed role. Table S1 presents a detailed summary of each NME utilizing biomarkers, highlighting the most representative biomarkers for each NME and the specific roles of these biomarkers. The following sub-sections describe each of these roles in more detail.

Figure 2. Number of Approved Neurology NMEs Categorized by the Roles of Biomarker Data.

*Footnote: The roles of PD biomarkers were not identified for the 14 products in “Other” category based on FDA review documents, except for a few cases described in the “Other Roles” section of this manuscript.

Surrogate Endpoints

A surrogate endpoint is generally a biomarker that is thought to predict clinical benefit but is not itself a measure of clinical benefit¹³. A surrogate endpoint that is reasonably likely to predict a drug’s intended clinical benefit (i.e., a reasonably likely surrogate endpoint) could be the basis for accelerated approval^13,14. For example, dystrophin protein production has been accepted as a surrogate endpoint for the accelerated approval of novel therapies to treat Duchenne muscular dystrophy (DMD), such as eteplirsen, golodirsen, casimersen, and viltolarsen.

The approval of tofersen for superoxide dismutase 1 amyotrophic lateral sclerosis (SOD1-ALS) relies on the reduction in plasma neurofilament light chain (NfL) concentration as a surrogate endpoint. Mechanistic evidence demonstrated that tofersen reduces SOD1 protein levels, while additional data supported the prognostic value of plasma NfL in predicting disease progression and survival in ALS. Furthermore, the observed correlation between reductions in NfL and slower declines in clinical outcomes provided strong support for the use of NfL as a surrogate endpoint.

For Alzheimer’s Disease (AD), the reduction of brain amyloid beta (Aβ) plaque observed through positron emission tomography (PET) imaging was used as a surrogate endpoint for the accelerated approval of lecanemab. A dose- and time-dependent decrease in Aβ plaque following treatment with lecanemab demonstrated its efficacy. Additionally, corresponding reductions in the decline of clinical outcome measures further confirmed that the reduction in Aβ plaque is likely to predict meaningful clinical benefits.

Confirmatory Evidence

Biomarkers can provide confirmatory evidence to support approval. In the cases where pharmacodynamic biomarkers offer valuable information about clinical outcomes, the results of a single adequate and well-controlled clinical investigation can be substantiated by confirmatory evidence from pharmacodynamic data¹⁰. For example, transthyretin (TTR) is a biomarker utilized as the confirmatory evidence for approval of NMEs indicated for polyneuropathy, such as patisiran, vutrisiran, and eplontersen. The sponsors demonstrated a reduction in serum TTR levels, and these data provide strong mechanistic support for their therapeutic efficacy.

Dose Selection

Biomarker data has been playing a significant role in dose selection throughout the drug development process, even when not pivotal for regulatory approval. In early-phase clinical trials, pharmacodynamic measurements using biomarkers have been utilized to guide optimal dosing strategies for subsequent studies. This approach ensures that the doses selected for phase 3 pivotal trials maximize therapeutic benefits while minimizing adverse effects. A notable example is the use of B-cell counts to inform dose selection for ublituximab-xiiy, a CD20-targeting mAb indicated for multiple sclerosis. Clinical assessments showed that an initial 150 mg IV dose on Day 1, followed by 450 mg on Day 15, achieved sustained B-cell depletion, with no added benefit from increasing the maintenance dose to 600 mg. These data supported the 450 mg maintenance dose in pivotal Phase 3 trials.

Biomarkers have also been used to optimize the dosing interval. For instance, the dosing intervals of ocrelizumab were selected based on B-cell counts, which demonstrated rapid depletion following IV infusion and evidence of repletion 6 months post-dose. This observation supported the selection of a 6-month dosing interval to maintain sustained B-cell depletion and ensure continued therapeutic efficacy. These examples highlight the integral role of biomarkers in optimizing dosing regimens for precise and effective therapeutic strategies.

Other Roles

Biomarkers serving other purposes were less frequently observed and thus not listed in Table S1. For example, biomarkers were used to define the indicated population. Efgartigimod and rozanolixizumab are indicated for the treatment of generalized myasthenia gravis (gMG) in patients who are defined by the positivity of anti-acetylcholine receptor (AChR) or anti-muscle-specific tyrosine kinase (MuSK) antibodies, besides that the serum levels of these autoantibodies were used as confirmatory evidence for approval. Similarly, the indicated population for inebilizumab was limited to anti-aquaporin-4 (AQP4) antibody positive patients with neuromyelitis optica spectrum disorder (NMOSD).

Besides PD biomarkers, pharmacogenomic biomarkers also contributed to clinical study design and labeling recommendations. For example, ApoE ε4 carrier status was used for patient stratification in the pivotal efficacy studies of lecanemab and donanemab for Alzheimer’s disease. While these therapeutics were approved for general population regardless of the genetic status, the labeling recommends testing for ApoE ε4 status prior to initiation of treatment to inform the risk of developing amyloid related imaging abnormalities (ARIA). Similarly, nusinersen and risdiplam for the treatment of spinal muscular atrophy used the pharmacogenomic biomarker SMN2 copy number for patient enrichment in the pivotal efficacy study, and the indicated population for approval was regardless of SMN2 copy number. In the case of amifampridine phosphate, the N-acetyltransferase gene 2 (NAT2) phenotype was used to guide individualized dosing recommendations for the treatment of Lambert-Eaton myasthenic syndrome (LEMS) in adults and pediatric patients.

Furthermore, biomarkers have been used for safety monitoring, such as the evaluation of anti-JCV antibody that supported re-approval of natalizumab. Blood lymphocyte count was measured to evaluate the risk of infections for the NMEs indicated for multiple sclerosis. In addition, the decrease in morning cortisol levels was evaluated for vamorolone to assess the risk of Cushing’s syndrome (hypercortisolism).

Target Disease and Drug Modality Associated with Biomarker Use

The prevalence of biomarker data in NME submissions varies across different indications in neurology (Figure 3A). Biomarker data are frequently submitted and reviewed for drug products targeting rare diseases, as well as chronic and progressive conditions such as multiple sclerosis, Alzheimer’s disease, and Parkinson’s disease. The new drug development for these diseases often faces challenges due to an incomplete understanding of disease pathophysiology and limited patient populations of rare diseases¹⁵. Biomarker data can address these challenges by supporting innovative clinical trial designs or by providing valuable insights into disease progression and treatment response.

Figure 3. Approved Neurology NMEs Categorized by Drug Modality and Indications.

*Footnote: The categories in Figure 3A are mutually exclusive. The “Rare diseases” category includes all the rare diseases in the analysis, except for ganaxolone (seizure), vigabatrin (seizure), and dalfampridine (multiple sclerosis).

The NMEs with or without biomarker data were further categorized by their drug modalities (Figure 3B). Among the 37 NMEs that included biomarker data, small molecules, oligonucleotides, and mAbs each accounted for approximately one third of the total (Figure 3B, left). In contrast, NMEs without biomarker data were predominantly small molecules, comprising over 80% of this subgroup (Figure 3B, right). Since the approval of eteplirsen in 2016, there has been a remarkable increase in the number of yearly approved oligonucleotide drugs, including siRNA and antisense oligonucleotides. All 10 approved oligonucleotides included biomarker data to demonstrate or support their effectiveness (Table S1). A similar trend was observed for mAbs, with the majority of approved mAb applications (10 of 13) incorporating biomarker data. Notably, the three mAbs approved without biomarker data were all indicated for migraine, an acute condition where the necessity for biomarkers in clinical outcomes is less pronounced. Collectively, the higher prevalence of biomarker use in mAb and oligonucleotide therapies likely reflects the growing development and increasing number of regulatory approvals of these modalities in recent years.

Lessons Learned from Biomarker Data Utilization

Despite the increased integration of biomarker data into regulatory submissions, there are various challenges with utilizing biomarker data in the regulatory decision-making process. Table 1 highlights key obstacles in integrating biomarker data into regulatory considerations, including establishing clinical outcome relevance, ensuring biomarker data quality, and validation of bioanalytical method.

Table 1.

Challenges in Using Biomarker Data to Support Regulatory Decision Making

Challenges	Details	Examples*
Lack of Clinical Relevance	• The relationship between pharmacodynamic (PD) biomarkers and clinical outcomes is not well established, making it difficult to confirm that biomarker changes directly translate into clinical benefit	• Satralizumab • Risdiplam
Biomarker Data Quality	• High variability in biomarker expression limiting the reliability of the data • Limited sample size reducing the statistical power of analyses • Limited dose range in the study hindering exposure-response analysis	• Nusinersen • Viltolarsen • Satralizumab
Bioanalytical Issues	• Bioanalytical method validation not established • Variability in bioanalytical assays	• Sodium phenylbutyrate and taurursodiol • Dimethyl fumarate • Nusinersen

^*Footnote: The weblink of FDA reviews for each NME examples are summarized in Table S2

Clinical Relevance

In successful cases where the biomarker information has been used to support approval and included in product labeling, clinical relevance of the biomarker is critical especially if it is intended as a surrogate endpoint for approval. Evaluating the clinical relevance of a biomarker requires comprehensive and deliberate considerations of the evidence from all available data sources including literature and observed data from nonclinical/clinical studies. Exemplified by tofersen, the acceptability of plasma NfL concentration as a reasonably likely surrogate endpoint for SOD1-ALS was evaluated based on multiple factors, such as evidence from the literature demonstrating the prognostic value of plasma NfL for ALS progression and survival, and the observed correlation between NfL reduction and clinical outcomes during treatment. Additionally, reductions in another biomarker SOD1 protein support the mechanism of action of tofersen and provide confirmatory evidence, which further contributed to the substantial evidence of effectiveness to support accelerated approval.

Biomarker Data Analysis

Various approaches of quantitative analysis have been used in analyzing biomarker data to evaluate the relationships of dose, exposures, and responses. Following initial descriptive analysis that visualize the longitudinal change of biomarkers, mixed-effect models with repeated measures (MMRM) is frequently employed for data analysis. Of note, the change of a biomarker following treatment is not only evaluated by its statistical significance, but also considered regarding whether the change is large, robust, and convincing¹⁶. Population-PK/PD and exposure-response analysis could further elucidate the covariate effects and compare biomarker responses in different subgroups of patients. For example, to explore the reduction in beta-amyloid (Aβ) plaque as a tool to facilitate new drug development for Alzheimer’s disease, a meta-analysis was conducted at the randomized group level with multiple amyloid-targeting monoclonal antibodies, to evaluate the association between placebo-subtracted change in the biomarker and change in clinical endpoint CDR-SB¹⁷. Correlation analysis for establishing biomarker and clinical outcome relationships has been commonly used in multiple NME reviews^16,18,19. Additionally, ensuring the quality of biomarker data is critical for achieving interpretable results and meaningful conclusions, particularly given challenges from data variability and potential confounders. Efforts need to be made to ensure adequate sample size, appropriate time-point selection, and a well-justified dose range.

Bioanalytical Methods

Challenges from the bioanalysis of biomarkers are noteworthy. A recent study examining BLA submissions for neurological diseases revealed significant variability in the bioanalysis of biomarkers²⁰, with a lack of standardization in both the selection of analytical parameters and adherence to recommended validation practices²¹. Variability in bioanalytical results can also occur due to study design and execution, sample handling and preparation, and instrument performance. Consequently, there is a pressing need to establish consistent best practices for the bioanalysis of PD biomarkers. Of note, biomarkers can be used for a wide variety of purposes during drug development, and a fit-for-purpose approach should be used when determining the appropriate extent of method validation²¹. For example, in the scenario that some aspects of the bioanalytical validation are sub-optimal, the biomarker data may be more suitable for making qualitative comparisons rather than quantitative conclusions¹⁸.

Besides the regulatory considerations listed above, enhanced accessibility of samples and data may also facilitate biomarker development. Ultimately, addressing the challenges and maximizing the effective utilization of biomarkers throughout the regulatory process requires multidisciplinary collaboration across stakeholders. Early consultation with regulatory agencies is encouraged to ensure alignment on biomarker utilization strategies and their intended applications.

Limitations of the Analysis

This analysis relied primarily on publicly available review documents of approved NMEs, which do not fully capture all instances of biomarker-related discussions that occurred during the preclinical or investigational new drug (IND) stages, biomarkers used for exploratory purposes, biomarker utilization in NDA/BLA supplements (e.g., approval for additional route of administration²²/dosing frequency²³, another indication²⁴) or biosimilar products²⁵. In addition, it may not include all the neurological conditions by using FDA divisions (DN1 and DN2) as the filter. A further survey on additional data sources including literature may provide a more comprehensive overview on the use of biomarkers.

Conclusion

Biomarker data have been increasingly integrated in drug development and regulatory submissions, supporting the regulatory decision-making as surrogate endpoints, confirmatory evidence, and basis for dose selection. By addressing current challenges such as bioanalytical validation, clinical relevance, and biomarker data quality, stakeholders can further leverage biomarkers to expedite drug development and approval, not only for neurological diseases but also potentially for other therapeutic areas with high unmet medical needs.