Validating Bioanalytical Methods for Biomarker Quantitation: A Regulatory Document Review

INTRODUCTION

Biomarkers Definitions Working Group of the Food and Drug Administration (FDA) and National Institutes of Health (NIH, USA) understands biological marker as “a defined characteristic that is measured as an indicator of normal biological processes, pathogenic processes, or biological responses to an exposure or intervention, including therapeutic interventions”¹ [1]. In the interpretation of [2][3] there is a slightly different emphasis on “a characteristic that is objectively measured and evaluated as an indicator of normal biological processes, disease processes, or pharmacologic responses to a therapy”. The authors of [4], with reference to [3], give the following definition of a biomarker: an objectively investigated parameter, measured with high accuracy, reproducibility and reliability, which allows to reflect the intensity of physiological processes, the state of health, the risk degree or the fact of the disease, its stage, and outcome prognosis.

The parameter can be chemical, physical, or biological. Biomarkers are used in assessing body physiological processes; detecting the disease and selecting treatment; predicting the disease course, progression and outcome; developing new medicinal products, establishing their therapeutic and side effects, and identifying the mechanism of action [4]. Given the large number of various biomarkers, a single classification is not applicable. Currently, biomarkers are classified depending on their purpose in diagnosis and treatment, on the study system, type, specificity,etc.; thus, several ways to classify them exist [4]. According to one of classifications [1]:

Type 0 – a marker used to detect presence of a disease and correlating with its clinical signs.

Type I – a marker associated with drug therapeutic effect and mechanism of action.

Type II – (predictor of clinical outcome, surrogate endpoint) a marker that allows to assume the disease outcome and evaluate the treatment effectiveness.

Based on the task, the following biomarker groups are known [2]:

• diagnostic ones allow to establish the presence or absence of a condition and its severity (for example, the amount of glycated haemoglobin and diabetes mellitus);
• prognostic ones allow to assess the risk of a condition or its complications (for example, gene polymorphism of coagulation factor);
• therapeutic ones help to evaluate or predict the outcome of exposure to a drug therapy (for example, the international normalised ratio (INR) is an assessment of warfarin effectiveness, which is a standardised indicator introduced by the World Health Organisation to unify the results of a prothrombin test in different laboratories [5]).

Accepting a parameter as a biomarker is a long multi-level process requiring to analyse the results of various clinical, phylogenetic and other studies [6]. In the review [6], the authors discussed the opportunity of using several parameters (related to genes encoding various enzymes of the cytochrome P450 complex, the uridine diphosphate glucuronosyltransferase family and glycoprotein P) as pharmacogenetic markers of antiepileptic pharmacokinetics to improve the effectiveness and safety of anticonvulsant therapy. However, it was concluded that due to the controversial results and the lack of appropriate randomised prospective placebo-controlled trials, reliable data on this issue do not exist as yet.

Notably, not every laboratory parameter can be called a biomarker. According to the authors [2], the main feature of a biomarker is its statistically valid relationship with the disease, complication, and therapy effect. The authors noted that only 20% of the studies they analysed² showed sensitivity, specificity, and prognostic value of this parameter. In most studies, the indicator is called a biomarker since it significantly increases/decreases/occurs in a certain pathology; this is clearly insufficient to recognise the parameter as a biomarker.

Given the increasing role of biomarkers in the diagnostics, therapy monitoring, and development of modern biological and pharmacological preparations, it is important to review and systematise current regulatory requirements, international, national, and scientific recommendations in order to identify and apply them.

The aim of the study is to review and summarise validation data of bioanalytical methods for biomarker assay.

MATERIALS AND METHODS

Regulatory documents of the International Council for Harmonisation (ICH), Food and Drug Administration (FDA, including Biomarker Assay Collaborative Evidentiary Considerations Writing Group and Critical Path Institute, C-Path), European Medicines Agency (EMA), Eurasian Economic Commission (EEC), and other open scientific sources (including electronic databases PubMed, Web of Science, RSCI (eLibrary), Google Scholar) mostly published over 2014-2024 were used for this study. Keywords used in queries: биомаркеры (biomarkers), аналитические/биоаналитические методы (analytical/bioanalytical methods), количественное определение (assay, quantification), валидация (validation).

RESULTS AND DISCUSSION

Clinical and analytical validation (qualification) of a biomarker. It was noted in [7] that a biomarker should be selected, studied and validated for its successful application. In this case, validation implies confirming that the selected biomarker fulfils its functions (clinical, diagnostic,etc.), and its use provides the expected results (defining the disease, assessing therapy effect on the disease course, identifying drug targets,etc.). The Points to consider document: scientific and regulatory considerations for the analytical validation of assays used in the qualification of biomarkers in biological matrices³ examines analytical validation (including all factors that belong to the analysis procedure, depend on acceptability of the samples, crucial reagents, and performance of equipment or test system, including sampling, processing, and storage) and clinical validation of a biomarker (to interpret biomarker measurements, assess clinical sensitivity, specificity, and accuracy of the biomarker in predicting the assumed result). Biomarker qualification is another common term in the literature [8], presumably synonymous with the validation; in our opinion, these both terms are attributable to clinical biomarker validation (or clinical qualification). Notably, many papers devoted to the biomarker validation discuss the aspects of clinical validation – prognostic value, biomarker response for various experimental conditions (populations, diseases,etc.) both within clinical and preclinical studies (pre-CS), as well as legal issues of protecting biological information, using biomedical techniques and biomaterials [9-16]. According to the authors [14], due to the lack of large-scale cohort studies that would confirm clinical significance of potential biomarkers, less than a hundred biomarkers out of several thousand discovered over the past decades have found their clinical use in medicine, and a little more than ten biomarkers – in the veterinary medicine. Therefore, clinical validation of biomarkers is certainly noteworthy. Notably, the more generally accepted and clinically widespread an indicator, the more data (using established measurement techniques) are available and assayed in a significant number of patients and populations, the more likely it is to identify the clinical significance and qualification of the indicator as a biomarker.

At the same time, the data underlying clinical validation / qualification of a biomarker should use reliable methods (bioanalytical techniques) to measure a parameter that is a potential biomarker. Since the following descriptions are used for a biomarker: defined, objectively measured, measurement with high accuracy, reproducibility and reliability, it is obviously necessary to develop a methodology for quantifying the selected biomarker and its subsequent validation.

Regulatory aspects of bioanalytical techniques validation and biomarker analysis methods. Validation procedure and scope for bioanalytical techniques is regulated by several documents<⁴. The guidelines aim to describe the validation of bioanalytical techniques used to detect concentrations of chemical and biological products and their metabolites in biological samples (blood, plasma, serum, other bodily fluids or tissues) obtained in toxicokinetic (TK) and pharmacokinetic (PK) studies and at all stages of clinical studies (CS), including comparative bioavailability/bioequivalence studies. All documents show procedural differences for chromatographic and ligand binding methods.

The Guidelines⁵ apply, inter alia, to biomarker analysis techniques. Other documents⁶ state that biomarker analysis and bioanalytical methods used to assess immunogenicity are not within the scope of the guidelines. The discrepancy concerning biomarkers is presumably because a full cycle of validation tests is rather resource-intensive and time-consuming, so its rationale in the early study of a potential biomarker is debatable. According to E.S.Donetal. [7], biomarker validation should first of all include quantification (a bioanalytical technique), or even develop such a technique. The biomarker study has an unknown range of variable possible concentrations, the influence of sampling conditions and storage, matrix components, and reagents used. In addition, the missing well-established reference samples for assay complicate the validation of such methods, causing lack of unified strategy or possibility to validate the method only in certain tasks, thus allowing unsuitability for alternative, other purposes [7].

Regarding analytical techniques for determination of biomarkers in biological matrices, English-language literature establishes the context of use (COU) and fit-for–purpose (FFP)⁷, implying that preliminary results of a potential (prospective) biomarker assessment can be used for exploratory research, while more essential decisions are based on data on a validated (qualified) biomarker [17]. Therefore, it is necessary to perform validation tests to an extent that confirms the compatibility of the technique with its intended use at a specific research stage. According to A.Safavi⁸, it is important to constantly revisit the assay context. The answers to “What will the data be used for?”, “What conclusions will be drawn from the analysis?”, “Are they qualitative or quantitative?”, “Will the biomarker data be used to submit to regulatory authorities or are they only used to better understand the mechanism of action?” form the basis of the COU and FFP concepts. The concept emerged in 2005-2006 [17] and underwent several changes and discussions in ad hoc working groups and during extensive scientific discussions at conferences, in scientific publications,etc. [18-20]; it was documented in detail in 2018-2021⁹ [21]. The legal status of the document¹⁰ entitled “Points to Consider” is not clear – in some papers, it is marked as the final document prepared by Critical Path Institute (C-Path) in USA and Food and Drug Administration (FDA) [22]. However, at the time of this writing, it was not available on the FDA website. Apparently, it is the predecessor of a regulatory guideline, and its recommendations are to be followed in practice. Notably, the concept and the document are subject to further improvement and development of approaches considering different expert opinions in this field [22-24], and the key provisions were included in the FDA document on the use of biomarkers for clinical studies of new veterinary drugs¹¹.

Stages and key elements of biomarker analytical validation. Biomarker analysis according to the fit-for-purpose concept (FFP) is a process that accompanies drug development using a biomarker (regardless of the role of the biomarker in this process) and includes (according to [18]) four continuous stages: methodology development and pre-validation, research methodology validation, complete or extensive methodology validation, and methodology validation in a study (Figure1). Measurements subsequently used to establish and confirm decision points in the biomarker clinical validation (qualification) should be sufficiently and rigorously validated to ensure that the analysis is sufficiently effective for application of the biomarker¹². In this paper, we focus specifically on the analytical part – measuring methodology/procedure for a parameter.

Most studies point out two main stages or directions – biomarker analysis in the drug development and for diagnosis¹³ [7][17][18][22][23]; they determine the scope of validation tests. Preliminary (screening) stages [17] are now also included (translational/partial validation, not found in earlier works [18][19] and in Figure 1, but present in Points to Consider¹⁴) as well as comparison with the scope of validation tests for methods evaluating BE, PK, and TK of drug substances¹⁵ [7][17]. The recommendations are summarised in Table1.

Table 1 is based on amended and changed materials¹⁶ of [7][17][18]; however, the key document was the preferred source of classification/validation stages in the table¹⁷, which does not include pre-validation stage of the technique. Considering pre-analytical issues and method development, this stage was mentioned in [17] and included elements such as describing reagent sources and reference standards; assessing lower limit of quantification (LLOQ) and the preliminary analytical range; method accuracy and precision (without recommended number of concentration levels and repetitions); calibration function; reagent specificity (according to supplier’s information from the literature), matrix effect; arranging feasible conditions for collection, processing, storage of reagents and samples, and recommendations on the number of concentration levels and repetitions for validated diseases and bioanalytical methods,etc. This should be called a preparatory and exploratory stage that does not end up with analysing experimental samples to obtain values for decision-making, even within pre-CS. Later, the authors¹⁸ also noted the importance of pre-analytical factors (biomaterial, interference, sampling procedure, sampling tubes and instruments, time and conditions, primary sample processing, storage and further logistics, freezing/thawing¹⁹) to be outlined while the technique is being developed and validated, standardised and defined for applicability (acceptability) of variations.

The early biomarker research (screening, biomarker selection, effectiveness evaluation) will only require a simple and minimally validated analysis (Table 1, Research validation). However, the clinical qualification of a biomarker (assessment of the therapeutic effect of the created drug, diagnostics,etc., both within pre-CS and CS) will require a confirmed analytical result. At the same time, according to different sources, the required tests may vary, and the choice of the stage/task to attribute the validated method during biomarker analysis are not complete or ambiguous, which certainly challenges development of a single consequent algorithm for analytical biomarker validation.

Analysing the combined recommendations (Table 1), noteworthy is the similarity of the key indicators used to validate a bioanalytical technique, regardless of the task. Analytical validation of the biomarker in general should include an assessment of seven parameters: accuracy (precision), range of analytical measurements (including LLOQ and the upper limit of quantification, ULOQ), parallelism (minimum required dilution and dilution linearity, if applicable), precision (within and between cycles, operators, days, and reagent lots, if applicable), selectivity, specificity, and stability (under operating conditions, short-term, long-term, during freezing/thawing). Some cases require additional analytical parameters, including stability, assessment of drug interactions,etc.²⁰. The authors [22] emphasised that the document²¹ highlighted the important assessment of parallelism and introduced the concept of an acceptable total analytical error (TAE). Notably, the concept of total analytical error sums up total error in determining the accuracy (relative systematic error, δ) and precision (CV), not exceeding 30% for ULOQ and 40% for LLOQ), and is present in the current documents regarding ligand binding methods²². Validation parameters, procedures, and acceptance criteria for biomarker analysis are similar, though they may not be identical to those used to obtain data on BE, FK, and TK²³. This document notes that method validation for biomarker analysis should address the same issues as the validation of a method for drug analysis. ... The approach used for drug analysis should be the starting point for validating biomarker analysis, although FDA understands that some characteristics may not apply or require various considerations²⁴.

Opinion of A. Safavi²⁵, which presents differences/similarities between biomarker validation and qualification, deserves to be highlighted, apparently from the purely analytical perspective of these processes (without affecting clinical validation/qualification of the biomarker). In his opinion, both processes are aimed at proving that the test is suitable for its intended purpose but differ in the scope and reliability of the estimated parameters and the number of replicates per each parameter. The scope of qualification procedures recommended by the author is comparable to the research and/or partial validations (Table 1). Notably, qualification of analytical procedure is a term hardly found in the scientific literature and regulatory documents.

Features of an analytical biomarker validation due to the used methods. When validating biomarker analysis methodology, some points to consider originate from the underlying method, depending on the measured biomarker (protein, lipid,etc.), availability of current methods, reagent kits, and measurement features (small sample volume, laboratory or field conditions, personnel qualifications,etc.), sensitivity and selectivity requirements, as well as method availability²⁶. Typically, evaluation can include plate-based analysis and various detection methods (fluorescence, chemiluminescence, electrochemiluminescence, chromogenic method, mass spectrometric assessment, and relatively new acoustic detection systems²⁷). According to [23], most biomarker analysis techniques are based on one of three methods:

– chromatographic techniques to be developed de novo;

– ligand binding techniques to be developed de novo;

– ligand binding techniques using commercial reagent kits.

Commercial reagent kits may be authorised for in vitro diagnostics, clinical laboratory testing, or for research only. Authors [25] examined some aspects of commercial kits used in the biomarker analysis. Commercial reagent kits should be tested on each analysed object (matrix) [22]. According to the authors [22], clinical laboratory kits may differentiate the need for validation and validation parameters on a case-by-case basis, depending on the intended use or characteristics of biomarkers. In this regard, it appears correct to discuss verification of methods aimed to confirm that the requirements used at its earlier validation (for example, by a manufacturer of reagent kits) are fulfilled in a specific laboratory (using a set from a certain batch and equipment available in the laboratory). These aspects seem to be similar to those discussed in [26] in relation to the activities of clinical diagnostic laboratories that perform biomedical biomarker tests as well. According to the publication based on analysis of relevant international regulatory documents, a minimum list of tests verifying clinical and laboratory studies should include intermediate precision, accuracy, linearity and range of calibration function, calibration verification, and confirmation of applicability for biological reference intervals specified in the manufacturer's manuals (if applicable) [26].

Research kits (intended for research only) are fully validated regardless of validation information in the operating manuals [22]. It is recommended to choose a specific kit when fully understanding whether it corresponds to the intended use of the relevant biomarker (such as calibration range or specificity). The reference standard included in a kit is evaluated using other commercially available products, paying attention to the differences from the endogenous substance, which is the target analyte. When the kit lot changes, make sure to confirm that the differences in the measured analyte concentration of the same sample are acceptable. Since research kits may become unavailable (discontinued production, disrupted economic contacts or supply), alternative analytic approaches should be considered (for example, separate use of reagents, not as part of the kit, if applicable).

Choosing the level of biomarker analytical validation depends on the method and the tasks. Pre-CS conducted in vivo using various animal species and in vitro using buffer media, cell lines, isolated organs,etc., may use clinical and research kits both specific to the measured marker, animal species or type of biological matrix, and with a different initial purpose. Kit application range beyond the manufacturer's recommendations for assessment of BE, PK, and TK requires full validation of the bioanalytical methodology, otherwise (including PD assessment) depending on the task. Thus, at the stage of biomarker selection, research level of validation is sufficient; when switching to model development using the selected biomarker, it is recommended to increase the validation level of the bioanalytical procedure to a partial one, similar to the full validation of the biomarker analysis method used to solve pre-CS/CS tasks, such as diagnosing abnormalities, assessing therapy effect,etc. (Table 1). A similar scheme can be recommended in some other cases (Figure 2).

Thus, only using clinical reagent kits as intended will minimise method validation procedures; in all other cases, relatively extensive validation tests will be required.

CONCLUSION

In conclusion, we would like to emphasize that extended use of biomarkers for drug development necessitates the improved methods of analytical quantification. Obviously, use of a biomarker should be preceded by a confirmation that the procedure is suitable for the planned goals, similar to a common procedure validating bioanalytical methods. The authors have compared key validation elements of the biomarker analysis and considered the level of analytical validation based on the method and the tasks. A more detailed scientific discussion of biomarker analysis validation is warranted, as there is still no joint scheme for this complex procedure. The discussion will allow using biomarkers more reliably, improving the quality of the accompanying research, and ultimately help introduce new effective and safe medicines into civil commerce.