Sage Journals: Discover world-class research

Abstract

Background

Tumour biomarkers have become increasingly important in oncology, shaping cancer diagnostics, classification, and patient management. Despite their potential, the use of cancer biomarkers in clinical settings remains limited.

Objective

This paper aims to outline biomarker development, from classical, serum protein markers to emerging tumour biomarkers, including meta-biomarkers, to show their diversity and point out the challenges in their development, reporting, and implementation in clinical practice as well as their relevance in evidence-based pathology and cancer classification.

Methods

A literature-based analysis, incorporating insights from our ongoing research, is presented.

Results

Although numerous potential biomarkers, biomarker signatures, and meta-biomarkers, are being discovered, existing innovations are often not supported by sufficiently rigorous research methodologies and standardised reporting practices to enable their translation into clinical practice.

Conclusions

To ensure that biomarker discoveries are both scientifically sound and clinically useful, improved research and validation methods, along with adherence to established reporting standards, are essential. We propose the use of the Hierarchy of Evidence for Tumour Pathology as a framework to evaluate and map existing evidence and identify knowledge gaps and research priorities.

Keywords

tumour biomarkers cancer diagnostics evidence-based pathology biomarker validation personalised cancer care tumour classification hierarchy of evidence for Tumour pathology (HETP)evidence Gap Maps (EGMs)

Introduction

Recent methodological advancements and new insights have fuelled significant progress in the development of tumour biomarkers. Biomarkers encompass biomolecules, cells, cellular structures, or bioactivities that can be quantified and evaluated as indicators of cancer risk, presence, characteristics, or progression. They can be evaluated in blood, other body fluids or tissues, and help to detect, diagnose, and monitor cancer, as well as predict and assess the effectiveness or toxicity of a treatment. Tumour biomarkers can be categorised according to various criteria, including their nature (e.g. protein, nucleic acid, metabolite, and radiographic biomarker), their function (e.g. receptor and mutation), the type of sample in which they are measured (e.g. blood-derived), their clinical purpose (e.g. screening, early detection, and recurrence monitoring), or the techniques employed for detecting them (e.g. immunoassays and molecular methods).¹

An ideal biomarker should exhibit high specificity and sensitivity, enable minimally invasive or non-invasive testing, and be consistently and accurately measurable across different laboratory settings, using a simple, cost-effective method. Additionally, it should be related to tumour biology, have a clear biological association with tumour characteristics and/or treatment response mechanisms.²

Tumour biomarkers are valuable tools, enhancing the accuracy and reliability of cancer diagnosis, prognosis, and treatment decisions. Therefore, it is essential that tumour biomarkers used in clinical practice are those that have been rigorously evaluated, and whose clinical utility has been demonstrated through validation in well-designed studies. Additionally, standardised procedures with rigorous quality control measures, following respective guidelines, are crucial throughout the biomarker development process, from identification and validation to the implementation of appropriate tests in routine clinical practice.^3,4

Objective

The paper aims to overview the developments in tumour biomarkers, particularly in their application for tumour classification, diagnosis, prognosis, and treatment decisions (prediction). This work examines both classical (serum protein markers, biochemical markers) and emerging biomarkers that have the potential to improve cancer detection, classification, and patient management. A summary of landmark events in tumour biomarker research is shown in Figure 1. The paper also explores the role of study design and scientific rigour in biomarker validation studies.

Figure 1.

Timeline – selected landmarks in tumour biomarker field. Abbreviations: RNA-seq: RNA sequencing; “WCT EVI MAP: Mapping the Evidence for the World Health Organization (WHO) Classification of Tumours: a Living Evidence Gap Map by Tumour Type” project⁸⁹; ISOBM: International Society of Oncology and Biomarkers; MCED: multi-cancer early detection tests; LLMs: large language models.

Classical tumour biomarkers

Historical view

The first biochemical marker was a protein discovered by Henry Bence-Jones in 1847 in the urine of a patient suffering from osteomalacia.⁵ Over a hundred years later, this protein was recognised as the multiple myeloma-associated immunoglobulin light chain and termed the Bence-Jones protein. In the early 20th century, the concept of tumour markers as substances produced by cancer cells or in response to cancer emerged with the observations of higher concentrations of certain substances in the blood of cancer patients, such as adrenocorticotropic hormone (ACTH) in small cell lung carcinoma, and prostatic acid phosphatase (PAP) in prostate cancer. Extensive research on cancer biomarkers began in the 1960s with the discovery of the first clinically useful human tumour markers, alpha-fetoprotein (AFP)⁶ by Abelev, and carcinoembryonic antigen (CEA) discovered by Gold and Freeman.⁷ Based on these discoveries, and to promote research in tumour marker development, the International Research Group for Carcinoembryonic Proteins was founded in 1973. In 1978, it was transformed into a society under the name International Society for Oncodevelopmental Biology and Medicine (ISOBM). The society, later renamed the International Society for Oncology and Biomarkers (ISOBM), included Professors Abelev and Gold as honorary members. Professor Hirai, a co-founder of the society, played a key role in establishing the foundations for the use of tumour markers in oncology, making significant contributions to the development and standardisation of tumour markers, particularly in the area of glycoprotein antigens. Notably, he conducted pioneering research on AFP, helping to elucidate its biochemical properties, improve assay methodologies, and establish its clinical relevance as a key marker in the diagnosis and monitoring of hepatocellular carcinoma and germ cell tumours.^8–10

Further developments

By the 1990s, the use of tumour markers expanded significantly, and significant improvements were made in the standardisation of tumour marker tests. Serum protein markers, often referred to as ‘classical’ tumour markers, became widely used in clinical practice. Other examples include CEA for colorectal cancer, AFP for hepatocellular carcinoma, CA 15-3 for breast cancer, CA125 for ovarian cancer, prostate-specific antigen (PSA) for prostate cancer, CA19-9 for pancreatic cancer, human chorionic gonadotropin (HCG) for trophoblastic tumours, S100-protein for melanoma, and squamous cell carcinoma antigen (SCCA) for cervical and other squamous cell epithelium-derived cancers. Interestingly, the clinical utility of PSA as a screening tool for prostate cancer has been a subject of significant debate, with concerns regarding overdiagnosis and overtreatment discouraging its routine use as a single marker.¹¹ However, the advent and integration of multiparametric magnetic resonance imaging into the diagnostic pathway has markedly improved the specificity of PSA screening. This has led to a re-evaluation of PSA’s role, effectively revitalising its value as a frontline marker in prostate cancer detection.¹² Endocrine markers NSE and ProGRP are used for the diagnosis of small cell lung cancer subtypes, while CEA and CYFRA 21-1 – for non-small cell lung cancer.^13,14

More recent markers include human epididymis protein 4 (HE4), which, in combination with CA125, forms the ROMA (risk of ovarian malignancy algorithm) index for risk assessment of ovarian cancer in women with pelvic masses,¹⁵ as well as PIVKA-II (Protein Induced by Vitamin K Absence or Antagonist-II), which, in combination with AFP, constitutes the GAAD index (combining Gender, Age, AFP, and DCP (PIVKA-II)) to improve hepatocellular carcinoma detection.¹⁶ The issue of meta-biomarkers, which incorporate multiple different molecules and data into diagnostic algorithms, is explored further in a subsequent section of this article.

Accuracy issues

Although tumour marker concentrations vary significantly between cancer patients and healthy individuals, they usually provide insufficient sensitivity and specificity. For instance, many classical biomarkers can be elevated in non-malignant conditions, leading to false-positive results, and reduced diagnostic specificity.¹⁷ Examples include CA125, extensively studied in ovarian cancer, which can also be elevated in benign conditions, such as endometriosis. The low sensitivity of CA125, which is reported to be elevated in up to 50% of stage I cases, limits its effectiveness as a screening tool for early-stage ovarian cancer.¹⁸ Similarly, PSA levels rise in response to benign prostatic hyperplasia. Evidence from secondary care-based studies showed that although PSA has high sensitivity, it has a low diagnostic specificity of 20% in symptomatic prostate cancer patients.¹⁹ Moreover, although PSA-based screening shows a reduction in prostate-cancer-specific mortality, its low specificity results in overdiagnosis and overtreatment.¹¹ Further drawbacks of classical tumour markers include individual variability, which is amplified by factors such as race, age, body mass index, and comorbidities, making the establishment of universal cut-off values for cancer detection challenging. Although of limited use for screening and diagnostic purposes, these markers have proved to be successful in several clinical applications, such as monitoring treatment response and disease progression. This is exemplified by both CA125 and PSA, which are clinically used in the follow-up of ovarian and prostate cancer patients, respectively, to assess treatment response and predict disease recurrence,^18,20 as well as by CEA, which has been identified as a predictor of disease progression in patients with metastatic colorectal cancer receiving chemotherapy.²¹ However, for monitoring of therapy response or surveillance after primary treatment, individual changes in marker levels assessed at defined time intervals are more relevant than absolute cut-off values based on healthy controls or specific cancer cohorts.²²

Novel classes of tumour markers

Molecular histopathological markers

Classical tumour markers remain fundamental to standard clinical practice in oncology. However, advances in biomedical sciences and techniques, such as immunohistochemistry, flow cytometry, polymerase chain reaction, and in situ hybridisation, have introduced a molecular dimension to histopathology, offering new possibilities for determining the origin of a tumour, facilitating differential diagnosis, and improving prediction and prognosis. For example, cytokeratins aid in differentiating carcinomas of various origins, e.g., CK7−/CK20+ is associated with colorectal carcinoma²³; WT1 helps identify ovarian cancer origin in metastases²⁴; HepPar-1 and Arginase-1 assist in confirming hepatocellular carcinoma and excluding metastatic adenocarcinomas in the liver²⁵; Ki67, a proliferation marker, is widely used as a prognostic factor in several cancer types, including breast cancer²⁶; HER2 in breast cancer, EGFR in lung cancer, and CD20 in lymphomas hold both prognostic and predictive significance.^27–31

Immunotherapeutic-context-related tissue markers

Some novel categories of biomarkers validated for clinical use refer not only to different biological levels, but also to a new immunotherapeutic context. Examples include the predictive value of the expression of the immune checkpoint protein, Programmed Cell Death Ligand 1 (PD-L1), in several tumour types, including non-small cell lung cancer³² and microsatellite instability (MSI) in different tumour types, including colorectal cancer,³³ where MSI also serves as a well-established diagnostic and prognostic marker.

Apart from protein biomarkers expressed in tumour cells that aid in classification and prognostication, cells characterising the immune microenvironment may also serve as biomarkers of prognostic value in some cancers. For example, tumour infiltrating lymphocytes (TILs), assessed on routine H&E slides, were shown to be associated with improved prognosis in early-stage HER2-positive breast cancer patients.³⁴

Bacteria-specific markers

Certain tumour markers have recently been associated with bacterial infections, assessed either directly through bacteria-specific molecules or indirectly through host-specific factors triggered by infections. Examples of such markers include gingipains (proteases produced by Porphyromonas gingivalis) and increased MMP-9 levels as host markers in oropharyngeal cancer patients.³⁵ Bacteria-specific markers include various molecules, such as Fusobacterium nucleatum-specific DNA detected in faecal samples, or proteins, such as anti-F. nucleatum antibodies in the serum and saliva of patients with gastrointestinal malignancies.³⁶

Markers derived from high-throughput methods

The advent of ‘omics’ methods, including genomics, transcriptomics, proteomics, epigenomics, and metabolomics, has accelerated the discovery of novel biomarkers within diverse molecules. A seminal transcriptomic study by Perou et al.³⁷ defined molecular subtypes of breast cancer. Whole genome sequencing has led to defining distinct glioma subtypes based on IDH1/2 mutations.^38,39 Hybrid capture-based next generation sequencing (NGS) has shown that tumour mutational burden (TMB) predicts pembrolizumab (anti-PD-1 therapy) response across different tumour types.⁴⁰ Accumulating high-throughput data hold promise for circulating metabolomic signatures to support diagnosis and prognosis, for example, in pancreatic cancer,⁴¹ as well as for circulating microRNA (miRNA) signatures, for example, in colorectal, lung, and breast carcinomas. miRNA-based clinically approved tests are emerging, exemplified by the CE-IVD-certified M371 and GASTROClear tests for early detection of testicular cancer and gastric cancer, respectively.^42,43

Markers generated from single-cell analyses

Novel technologies now enable genomic and transcriptomic characterisation at the single-cell level. This allows the examination of the molecular features of individual cells and the identification of subpopulations with distinct gene expression profiles, genetic alterations, or lineage origins. A seminal example in cancer research is the study by Patel et al.,⁴⁴ in which single-cell RNA sequencing was used to reveal intratumoral heterogeneity in glioblastoma. This approach is particularly valuable for studying tumour immune microenvironment and detecting rare cell types, such as cancer stem cells or therapy-resistant clones, to inform personalised treatment strategies. Further opportunities for tumour biomarker discovery are provided by single-cell spatial omics, offering valuable spatial insight by integrating omics-level data with tissue architecture.⁴⁵

cfDNA/ctDNA and circulating tumour cells as liquid biopsy markers

New classes of circulating tumour markers, at the cellular and molecular levels, referred to as ‘liquid biopsy’, include tumour cells (CTCs) and cell-free tumour DNA (ctDNA) found in the peripheral blood of cancer patients. CTCs are present in very small quantities, even in patients with advanced cancer. They are detected and characterised by surface markers, such as EPCAM or cytokeratins, and have prognostic value, for example, in breast, prostate, and colorectal cancers.⁴⁶ However, to date, CTC detection faces significant challenges due to their low abundance, cellular heterogeneity, and issues related to standardisation of detection methods. Thus, CTCs are often considered overvalued in terms of their clinical utility, currently offering limited practical benefit in most cases.^47,48

ctDNA constitutes only a small fraction of the total cell-free (cf)DNA circulating in the blood stream, sometimes as little as 0.01 to 1%. ctDNA carries various malignancy-associated alterations, including single mutations, deletions, fusions, and amplifications.⁴⁹ These markers can be detected and quantified using highly sensitive technologies such as digital PCR (e.g. for single mutations) or NGS-based deep sequencing (for targeted approaches, i.e. limited to pre-specified genes or untargeted approaches, i.e. genome-/exome-wide).^50,51 ctDNA analysis can be useful even when tumour tissue is unavailable for molecular characterisation (a tumour-agnostic approach), enabling patient stratification for certain targeted or immune therapies based on genetic driver mutations or other molecular indicators. Treatment-relevant ctDNA markers are exemplified by mutations of BRAF in melanoma, KRAS, NRAS and BRAF in colorectal cancer, EGFR in lung cancer, and ESR1 and PIK3CA in breast cancer.^51–54 Pre-analytical factors, including sample collection, processing, storage, and handling, that profoundly affect biomarker stability, accuracy, and reproducibility in both research and clinical applications, are especially critical in ctDNA analysis, though equally important for all cancer biomarkers. Beyond these pre-analytical considerations, obtaining a sufficient amount of cfDNA and employing a standardised, highly sensitive analytical procedure that meets the quality requirements of international (ISO15189) and national institutions^50,55,56 are essential prerequisites for liquid biopsy. Further developments include comprehensive pattern recognition methods for nucleic acid markers in total cfDNA, capturing both tumour characteristics and antitumour responses of immune and other patient cells. These methods encompass epigenetic DNA methylation patterns, DNA fragment patterns, nucleosome positioning patterns, and more.⁵⁷ The development and application of highly sophisticated bioinformatic pipelines are essential for deconvoluting the complex information obtained through untargeted sequencing approaches.⁵⁸

Multi-omic approaches

High-throughput studies, enabling the investigation of the molecular landscape of multifactorial diseases, such as cancer, have generated vast amounts of data, stored in multiple databases, such as GEO,⁵⁹ Ensembl,⁶⁰ the Cancer Genome Atlas (TCGA),⁶¹ MetaboLights,⁶² and ENCODE,⁶³ just to name a few. A major value of these data lies in their secondary use driven by recent advancements in data science and artificial intelligence (AI). While single-layer omic approaches provide valuable insights, understanding complex, multilayered biological systems requires the integration of genomic, transcriptomic, proteomic, and other ‘omic’ data into multi-omics approaches, which offer new potential for biomarker discovery and are currently being extensively developed.^64–66

The excitement surrounding the discovery of novel markers is tempered by numerous factors that hinder their implementation in clinical practice, and many novel biomarkers still lack sufficient sensitivity and specificity.¹

Improving testing accuracy: Meta-biomarkers

It is unlikely that a single biomarker can reliably support early screening, diagnosis, prognosis, or assess the safety and efficacy of treatment modalities.^67,68 Therefore, new opportunities for precision medicine are emerging through the development of meta-biomarkers incorporating multiple different molecules in biomarker panels or integrating tumour biomarkers with data derived from other diagnostic methods (e.g. endoscopy and radiography), clinical records or exposomes.^69,70

Panels of classical tumour markers

Combining individual classical markers into panels has been shown to improve their effectiveness.⁷¹ Significant progress has been made in the development of blood and urine biomarker panels, which are critical for the early detection and curative interventions in pancreaticobiliary malignancies. For biliary tract cancer, a combination of serum protein biomarkers (CEA, CA 19-9, CYFRA 21-1, and MMP7) shows diagnostic potential. For pancreatic cancer, combinations like CA 19-9 with CFB or IL-8, IL-1β, and CA 19-9 demonstrate high sensitivity and specificity. A urine three-biomarker panel (LYVE-1, REG1B and TFF1) was developed as a non-invasive test for the early detection of pancreatic ductal adenocarcinoma, and its use in the PancRISK score enables the stratification of patients for appropriate management.^72,73 Entire biomarker panels can also be combined. For example, the 4MP protein panel (consisting of the precursor form of surfactant protein B, CA125, CEA and CA 19 fragment) in conjunction with a lung cancer risk prediction model (PLCOm2012) demonstrates high predictive performance for lung cancer screening in high-risk individuals.⁷⁴

Multigene diagnostic tests

Similar to classical tumour markers, integrating molecular markers has proven to be effective. For instance, the Oncotype DX assay, a leading genomic biomarker test for breast cancer, evaluates the expression of 21 genes, including 16 cancer-related genes. By incorporating expression levels of these genes into an algorithm, the assay generates a recurrence score that predicts the benefit of adding chemotherapy to endocrine therapy.⁷⁵ Widely recommended in international guidelines, including those of the American Society of Clinical Oncology (ASCO) and the National Comprehensive Cancer Network (NCCN), Oncotype DX provides both predictive and prognostic insights, guiding adjuvant chemotherapy decisions for women with early-stage oestrogen receptor (ER)-positive, HER2-negative breast cancer. Other innovations include the ColonES assay, a blood-based multigene cell-free DNA (cfDNA) methylation test, which enables early detection of colorectal cancer and prognosis estimation.⁷⁶ Broad implementation of such tests is frequently limited by financial constraints and lack of reimbursement.

Markers combined with clinical data and structural insights

Complementing biomarker-based tests with patient clinical data has a significant potential to improve diagnostic accuracy. For example, combining CA125 and HE4 serum levels with menopausal status resulted in the development of ROMA for ovarian cancer, as mentioned above.¹⁵ Similarly, combining CA 19-9 levels alongside patients’ BMI values into the PancRISK score has been shown to improve its performance.⁷²

Also, tissue tumour markers can be integrated into a meta-biomarker approach, particularly as it is now possible to capture the spatial organisation and structure of the analysed tissues in detail.⁷⁷ Emerging technologies, such as spatial transcriptomics and spatial proteomics, allow researchers to map the relative positions of the investigated molecules within a given sample. This is further enhanced and complemented by the development of digital pathology, image analysis and interpretation, assisted by AI-based tools, such as the FDA-approved Paige AI platform.⁷⁸ This integrative approach is termed ‘pathomics’ and encompasses the entirety of a tissue’s pathological characteristics, the ‘pathome’. When combined with radiomics, pathomics becomes part of ‘morphomics’, the study of morphological features across all levels of achievable resolution within a system or organism. ‘Radiomics’ refers to the extraction of quantitative data from radiological medical images, mostly at a macroscopic scale.⁷⁷ Future research should complement existing knowledge on tumour biology with morphological insights at the ultrastructural, microscopic, and macroscopic scales.

Multi-cancer early detection panels

Multi-cancer early detection (MCED) tests represent a relatively new generation of cancer screening assays that may incorporate meta-biomarkers. These tests are designed to analyse a single specimen, typically blood, for cancer-related biomarkers, most commonly ctDNA. MCED assays aim to detect multiple cancers simultaneously, helping to identify potential malignancies in asymptomatic individuals. MCED platforms include Galleri®, CancerSEEK, SPOT-MAS™, Trucheck™, and Cancer Differentiation Analysis, all of which generally exhibit high specificity. Yet, the sensitivity of these tests remains highly variable depending on the cancer type and stage. Moreover, sensitivity was shown to be influenced by study design, population, follow-up time and the reference standard test used.^79,80 MCED requires further validation in clinical trials before it can be implemented in clinical practice.^68,81 The integration of AI with protein biomarkers and/or cfDNA holds promise for earlier cancer detection.⁸² Machine learning (ML) and AI-based techniques⁸³ hold the potential to establish novel meta-biomarkers by integrating multi-omic datasets (such as genomics or radiomics) with real-world data, while enabling biomarker validation across diverse populations beyond controlled clinical trials.^84,85 In addition, this approach makes it possible to analyse complex, multidimensional data to identify subtle patterns indicative of early disease stages.

Biomarkers for tumour classification

Tumour classification is a complex process that involves identifying and categorising cancer types based on their tissue of origin, histological and molecular characteristics, and clinical behaviour. Historically, the main criterion for tumour classification was microscopic analysis of slides to identify morphological alterations. In recent decades, technological advancements have enabled more in-depth characterisation of tissue and body fluid samples from patients with various tumour types.

The World Health Organisation (WHO) Classification of Tumours (WCT), published by the International Agency for Research on Cancer (IARC), provides globally recognised standards for tumour diagnosis and practical guidance for pathologists and oncologists worldwide. Presented as a series of volumes, known as the WHO Blue Books, and a website (https://tumourclassification.iarc.who.int/welcome/), the WCT provides a comprehensive and standardised classification system, including characteristics of each tumour type, pathological and clinical features, and associated molecular alterations. IARC acknowledges that ‘new technologies are now transforming the field of pathology more rapidly than at any other time in the past 30 years, and it has become increasingly clear that the traditional approach to cancer classification is insufficient’.⁸⁶ Traditionally based on histological and, more recently, molecular characteristics, tumour classification is becoming increasingly multidisciplinary. Currently, the WCT includes additional characteristics, such as imaging and diagnostic molecular pathology data. As a result, the term ‘biomarker’, by referring to diverse data, has gained a broad meaning and has been assigned an increasingly important role in supporting classification. Importantly, while tissue analysis remains essential for tumour classification, chemical pathology (also known as clinical biochemistry), which involves the investigation of bodily fluids, is now reflected in the WCT. Beyond enhancing diagnostic accuracy, biomarkers enable personalised treatment strategies, and the discovery of novel, reliable biomarkers could lead to significant changes in tumour classification. For example, IDH (isocitrate dehydrogenase) mutation and 1p/19q co-deletion status have been integrated into central nervous system tumour classification, refining the categorisation of adult-type diffuse gliomas, to assist in the tailoring of treatment strategies.⁸⁷ Biomarkers are expected to play an increasingly significant role in tumour classification by providing molecular and genetic insights that enhance traditional histopathology. It is highly probable that these biomarkers will include meta-type biomarkers identified through metadata analyses. A notable example is the WHO classification of endometrioid carcinoma of the uterine corpus into four molecular subtypes, largely based on findings from The Cancer Genome Atlas (TCGA).⁸⁸ Such biomarker-driven cancer classifications will pave the way for better personalised and more effective therapeutic strategies.

Until now, tumour classification has largely relied on the consensus of the opinions from invited experts. However, the data included have not always been informed by the best representation of the available evidence.⁸⁶ It has become crucial to synthesise and map evidence to ensure guidance is based on the most reliable evidence. It is also critical for identifying knowledge gaps and pockets of low-level evidence in tumour pathology, to guide research prioritisation. To achieve this, the IARC has launched a project entitled ‘Mapping the Evidence for the World Health Organisation (WHO) Classification of Tumours: a Living Evidence Gap Map by Tumour Type (WCT EVI MAP)’,⁸⁹ employing an adapted Evidence Gap Map (EGM) methodology.^90,91 One of the project’s key goals is to integrate the results into the strategic planning for the 6th and subsequent editions of the WCT. The resulting EGMs are also expected to accelerate the discovery of novel clinically applicable tumour markers by describing evidence gaps and pinpointing potential areas of interest for further research.

Biomarker development guidelines

When integrating biomarkers into clinical practice, several challenges arise, including the standardisation of assay methods, validation across diverse populations, and the need for robust clinical trials to establish their efficacy and reliability.⁹² New biomarkers or combinations of biomarkers intended for clinical use, and ultimately to support tumour classification, must then undergo rigorous regulatory evaluation. Regulatory agencies and organisations, such as the International Organization for Standardisation (ISO),⁹³ the International Quality Network for Pathology (IQN Path),⁹⁴ the U.S. Food and Drug Administration (FDA),⁹⁵ the National Institutes of Health (NIH),⁹⁶ the European Medicines Agency (EMA),⁹⁷ and the WHO,⁹⁸ have established guidelines to regulate this process. Biomarkers require validation and must demonstrate their capacity to improve clinical decision-making. Regulatory bodies define the qualification processes for biomarkers, outlining specific performance characteristics, such as precision and quality assurance, that must be met.⁹⁹ Guidelines from organisations, such as the Clinical and Laboratory Standards Institute (CLSI), are essential for standardisation across different laboratories, instruments, and populations to ensure consistent measurements. Unfortunately, the use and assessment of most tumour markers, even the classical ones, are not standardised, and the state of harmonisation remains unknown.^86,100 Detailed guidelines for the evaluation of biomarker studies and strategies to avoid bias are beyond the scope of this article and are accessible through, for example, the EQUATOR (Enhancing the QUAlity and Transparency Of health Research) Network and the GRADE (Grading of Recommendations, Assessment, Development and Evaluation) approach.^101–103

Identification of a tumour biomarker – Study design issues

Results from biomarker discovery studies should be made publicly accessible to foster further research aimed at enhancing their clinical utility. Prior to publication, adherence to biomarker development guidelines is essential. In this field, study design issues require special attention.^99,104 For example, current clinical trial designs often fail to adequately address the requirements for assessing biomarker value.¹⁰⁵ From discovery to establishing the utility and value of biomarkers for clinical decision-making in diagnosis, prognosis, and treatment, multiple types of study designs are employed. These may include laboratory-based exploratory studies, case studies, case–control studies, animal studies, and ultimately, randomised controlled trials. The relevance of study designs to biomarker research is summarised in Table 1. The reliability of medical research for evidence-based medicine (EBM) is stratified by the levels-of-evidence (LoE) hierarchy.¹⁰⁶ Clinical decisions should be guided by an appraisal and synthesis of the highest-level evidence. Similarly, tumour classification, also an important contributor to cancer patient management, should be informed by the best available evidence.

Table 1.

Study designs according to the WCT EVI MAP framework⁸⁹ and their relevance to biomarker research.

Study design	Relevance to biomarker research
Systematic reviews	Rigorously synthesise existing evidence, assess consistency across studies, and increase confidence in a biomarker’s validity while identifying research gaps relevant to the research question
Population-based descriptive studies	Provide epidemiological context, including biomarker prevalence and associated risk factors, supporting population-level relevance
Randomised controlled trials (RCTs)	Critical for validating predictive biomarkers by testing their ability to stratify patients for specific treatments
Observational longitudinal RCT-derived studies	Validate biomarkers for secondary or exploratory endpoints (e.g. prognostic markers) using high-quality datasets
Diagnostic test accuracy studies	Evaluate clinical utility through measures such as sensitivity, specificity, and predictive values, essential for regulatory approval
Diagnostic test agreement/reproducibility studies	Confirm measurement reliability across observers, laboratories, and platforms, supporting clinical adoption
Cohort studies	Validate prognostic biomarkers by establishing temporal relationships with outcomes such as survival or recurrence
Classic and self-matched case–control studies	Identify and validate disease-associated biomarkers that reflect prior exposures and may help explain differences in outcomes. Biomarker levels are compared between individuals with a disease or outcome and independent controls (classic case–control studies), or within the same individuals across different time points or states (self-matched case–control studies). Especially useful for rare conditions and early-stage discovery
Cross-sectional studies	Used in early discovery to explore associations between biomarkers and disease status or tumour characteristics at a single time point
Other longitudinal studies	Provide preliminary performance data in specific subgroups or contexts, refining methodology before large-scale validation
Case series	Generate hypotheses by identifying potential biomarkers from patterns observed in small patient groups
Laboratory test validation studies	Ensure assays reliably detect biomarkers under controlled conditions in early development stages
Case reports	May first reveal biomarker relevance through unique or rare findings, inspiring exploratory research despite limited generalisability

Consequently, a new pathology-relevant hierarchy, the Hierarchy of Evidence for Tumour Pathology (HETP), has been developed via a Delphi-informed process to standardise the evaluation of histopathological research, tumour classification and biomarker studies.¹⁰⁷ The HETP categorises studies based on their design and the associated risk of bias, which can influence the reliability of the results. Studies of higher LoE (e.g. systematic reviews and randomised controlled trials) provide stronger evidence than studies of lower LoE (e.g. case reports). To evaluate the evidence for tumour classification, an updated HETP version¹⁰⁸ aligned with the WCT EVI MAP⁸⁹ framework has been developed. It includes eight categories (‘characteristics’), clinical features, epidemiology, aetiology, pathogenesis, histological diagnostic features, molecular diagnostic features, prognosis, and prediction, most of which accommodate tumour biomarker studies.

In biomarker research, rigorous validation is crucial to ensure clinical utility. The updated HETP¹⁰⁸ organises study designs into five LoE, labelled P1 through P5 (for Pathology). For all categories, the highest level, P1, comprises systematic reviews, while the lowest level, P5, comprises case reports. Levels P2, P3, and P4 vary between ‘Histological and Molecular diagnostic features’, ‘Aetiology (Risk factors)’, ‘Pathogenesis (Mechanisms and pathways)’, ‘Prognosis’, and ‘Prediction’ in terms of the included study designs. A structured LoE system helps to determine which biomarkers are supported by substantial evidence, meeting the precision and reliability required for clinical diagnostics and patient care.

Studies that may lead to identifying new biomarkers can be carried out in nearly all study designs. To ensure a consistent system for classifying study designs of published studies across diverse disciplines, including pathology, epidemiology, oncology, molecular biology, and public health, a Decision Tree (DT) was developed within the WCT EVI MAP project mentioned above.¹⁰⁹ A special diagnostic branch has been included in the DT to address studies evaluating the accuracy and reproducibility of diagnostic tests.

Assigning an adequate LoE according to the study design ensures a rigorous, evidence-based pathway in biomarker development, from discovery to clinical implementation and utility in tumour classification. Levels of evidence of studies reporting on biomarker discovery through to clinical application, stratified by study design and characteristics, are summarised in Table 2.

Table 2.

Study designs contributing to biomarker discovery, with their levels of evidence (LoE) for pathology (P), across key biomarker-related characteristics¹⁰⁸

Biomarker discovery characteristic	Study design	LoE
Histological diagnostic features and Molecular diagnostic features	Systematic reviews	P1
	Population-based descriptive studies	P2
	Randomised controlled trials (RCTs)	P2
	Observational longitudinal RCT-derived studies	P2
	Diagnostic test accuracy studies	P2
	Diagnostic test agreement/reproducibility studies	P2
	Cohort studies	P3
	Classic case–control studies	P3
	Cross-sectional studies	P3
	Self-matched case–control studies	P3
	Other longitudinal studies	P4
	Case series/longitudinal case series	P4
	Laboratory test validation studies	P4
	Case reports	P5
Aetiology (Risk factors)	Systematic reviews	P1
	Randomised controlled trials (RCTs)	P2
	Observational longitudinal RCT-derived studies	P2
	Cohort studies	P2
	Other clinical trials	P3
	Other longitudinal studies	P3
	Classic case–control studies	P4
	Case series/longitudinal case series	P4
	Cross-sectional studies	P4
	Self-matched case–control studies	P4
	Ecological studies	P4
	Case reports	P5
Pathogenesis (Mechanisms and pathways)	Systematic reviews	P1
	Observational longitudinal RCT-derived studies	P2
	Cohort studies	P2
	Self-matched case–control studies	P2
	Classic case–control studies	P3
	Cross-sectional studies	P3
	Other longitudinal studies	P3
	Case series/longitudinal case series	P4
	Other mechanistic basic science studies	P4
	Animal studies (mechanistic)	P4
	Case reports	P5
Prognosis	Systematic reviews	P1
	Population-based descriptive studies	P2
	Randomised controlled trials (RCTs)	P2
	Observational longitudinal RCT-derived studies	P2
	Cohort studies	P2
	Other clinical trials	P3
	Other longitudinal studies	P3
	Case series/longitudinal case series	P3
	Cross-sectional studies	P4
	Classic case–control studies	P4
	Ecological studies	P4
	Case reports	P5
Prediction	Systematic reviews	P1
	Randomised controlled trials (RCTs)	P2
	Observational longitudinal RCT-derived studies	P2
	Cohort studies	P2
	Other clinical trials	P3
	Other longitudinal studies	P3
	Cross-sectional studies	P4
	Classic case–control studies	P4
	Case series/longitudinal case series	P4
	Ecological studies	P4
	Case reports	P5

Taken together, studies at levels P3 through P1, with data of progressively higher reliability, potentially provide reduced risk of false-positive or false-negative results that could misguide diagnoses and treatments. Level P5 and P4 studies are used to identify candidate biomarkers. Level P4 studies refine biomarker assays, ensuring their technical reliability. Level P3 and P2 studies evaluate biomarkers in real-world or experimental settings, assessing their predictive, diagnostic or prognostic value. Level P1 studies synthesise findings across multiple studies, providing high-confidence evidence for clinical guidelines or regulatory approvals.

It has become evident that for diagnosis, prediction, and prognosis, studies at levels P1 and P2 have been particularly considered in the definition of clinical guidelines in oncology. To evaluate the efficacy of cancer treatments, study designs such as systematic reviews or RCTs are required, as they provide the highest level of scientific rigour, data collection, and documentation. Since biomarkers do not directly influence patient outcomes and prospective RCTs are not conducted due to financial constraints, biomarker studies have historically struggled to reach these high levels of evidence and have often been excluded from clinical guidelines. However, in our most recent update of the HETP, ‘Observational longitudinal RCT-derived studies’ – that is, translational studies in which biomarkers were evaluated within originally drug-efficacy-focused RCTs – were also classified as level P2. This development opens up the possibility for high-quality biomarker studies to be considered for future clinical guidelines. Furthermore, biomarker evaluations are categorised into different study types: ‘Diagnostic test agreement/reproducibility studies’ for method or investigator comparisons, ‘Laboratory test validation studies’ for purely analytical evaluations in the laboratory, and ‘Diagnostic test accuracy studies’ to assess the clinical utility of biomarkers, which require specific quality criteria, such as sensitivity, specificity, and predictive values. These new categories provide a more differentiated perspective on studies assessing the value of biomarkers in oncological diagnostics.¹⁰⁹

Those studies are paramount for the analytical and clinical accuracy of tumour marker tests when applied in clinical patient care. It is known that different results may be obtained if a tumour marker is quantitatively assessed in blood serum or plasma using different assays or analytical platforms from different manufacturers. A thorough analytical and pre-analytical evaluation, together with head-to-head method comparison using control materials and real patient samples are essential steps for clinical labs. More advanced requirements are better standardisation and harmonisation of the assays already in the development and optimisation phase of the manufacturers. Later, the regular use of internal controls and participation in external quality controls are mandatory measures for clinical labs. Moreover, clinical studies testing the utility of the biomarkers for all relevant clinical questions are necessary to define appropriate decision thresholds and inform about positive and negative predictive values of the markers. All these study types have been covered by the new HETP.¹⁰⁸

Tumour Biomarker Reporting and Evidence Mapping

Adherence to publishing guidelines for clear and transparent reporting of biomarker performance characteristics, including limitations, is essential.⁶⁹ Guidelines, such as STARD (Standards for Reporting Diagnostic Accuracy)¹¹⁰ and REMARK (Reporting Recommendations for Tumour Marker Prognostic Studies),¹¹¹ as well as tumour pathology-specific guidance,¹¹² recommend adhering to standardised reporting frameworks.

The Evidence Gap Maps by tumour type, being developed within the WCT EVI MAP project,⁸⁹ will facilitate evidence-based decision-making in tumour pathology. By integrating multiple published studies, EGMs will provide an overview of the current evidence for biomarkers in cancer classification and help establish research priorities by highlighting areas that require further investigation. With numerous research groups exploring diverse cancer parameters and characteristics for potential biomarker development, the living (regularly updated) EGM tool will dynamically reflect the evolving body of evidence. Designed for open access, this tool will benefit all stakeholders, including the scientific community and diagnostic companies. It will assist researchers, clinicians, and policy makers in identifying biomarkers that meet precision and reliability standards, have high clinical utility, and can be effectively translated into patient care.

Conclusions

Biomarkers hold significant potential for advancing cancer diagnostics and patient management. However, their integration into clinical practice is hindered by scientific, technical, translational, and practical challenges. These difficulties largely stem from the biological complexity of cancer. As a multifactorial and highly heterogeneous disease, cancer is driven by multiple dynamic and context-dependent molecular mechanisms, and influenced by both individual constitution and environmental factors. Our understanding of tumour biology, including the biological role of biomarker molecules and their links to disease mechanisms, is rapidly expanding, driven by advances in molecular profiling technologies at both the tissue and single-cell levels, as well as by the integration of multi-omics approaches supported by data science and artificial intelligence. As a result, countless potential biomarkers, biomarker signatures, and meta-biomarkers are being discovered. It is now possible to develop multiplex panels for multi-cancer early detection, referred to as MCED. Yet, despite this vast potential, a relatively small number of molecular biomarkers and biomarker panels have been successfully translated into clinical practice or incorporated into tumour classification systems. The complexity, inter- and intra-tumour heterogeneity, and the dynamic nature of cancer make the translation of tumour biomarkers into clinical practice exceptionally difficult. First, candidate biomarkers, often identified in small cohorts, may lack reproducibility in independent populations. Second, biomarkers that prove useful in specific contexts may not demonstrate utility across different tumour subtypes or clinical settings. Third, the clinical value of many potential biomarkers is limited by their insufficient specificity and sensitivity, which undermines their diagnostic reliability. This is further confounded by non-biological factors, such as technical variability between biomarker assays, their insufficient sensitivity or inadequate standardisation, and implementation obstacles, such as regulatory and reimbursement barriers. Finally, pre-analytical considerations that affect biomarker stability and reproducibility must not be overlooked. Therefore, to ensure the reliability, reproducibility, and clinical effectiveness of tumour biomarkers, both sample handling and assessment methods must undergo rigorous validation across diverse platforms, patient populations, and clinical contexts. To ensure that biomarker discoveries are scientifically robust, clinically meaningful, and effectively integrated into practice, it is essential to employ study designs appropriate to the specific research objectives, adhere to established reporting and validation guidelines, and adopt evidence-based pathology and the HETP framework. As a result, tumour biomarkers will become increasingly important in cancer classification by enabling more precise stratification of cancer subtypes, which is crucial for personalised treatment approaches.

Footnotes

ORCID iDs

Kateryna Maslova

Magdalena Chechlinska

Irmina Maria Michalek

Lukasz Taraszkiewicz

Paulina Kober

Inga Trulson

Karolina Worf

Sophie Gabriel

Laura Knoblauch

Fiona Campbell

Stefan Holdenrieder

Magdalena Kowalewska

Collaborators

WCT EVI MAP, including: Anne-Sophie Bres, Richard Colling, Beatriz Pérez Gómez, Elena Plans-Beriso, Ester García-Ovejero, Oana M. Craciun, Joanna Didkowska, Christine Giesen, Iciar Indave, Alex Inskip, Ramon Cierco Jimenez, Dilani Lokuhetty, Nur Diyana Md Nasir, Cécile Monnier, Marina Montes-Mota, Marina Pollán Santamaría, and Puay Hoon Tan.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This publication, through the WCT EVI MAP project, is funded by the European Union (HORIZON EUROPE Grant No. 101057127).

Declaration of conflicting interest

The authors declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: SH has received research funding or honoraria from Roche, Sysmex, Thermo, Guardant, Novartis, Trillium, Medica, and Instand and is founder of SFZ BioCoDE and CEBIO. MCh, SH, and MK are editorial board members of Tumour Biology journal but had no involvement in the peer review process of this article. The other authors declared no potential conflicts of interest.

Disclaimer

The content of this article represents the personal views of the authors and does not represent the views of the authors’ employers and associated institutions. Where authors are identified as personnel of the International Agency for Research on Cancer/World Health Organization, the authors alone are responsible for the views expressed in this article, and they do not necessarily represent the decisions, policy, or views of the International Agency for Research on Cancer/World Health Organization.

References

Zhou

Tao

Qiu

, et al. Tumor biomarkers for diagnosis, prognosis and targeted therapy. Signal Transduct Target Ther 2024; 9: 132.

Biomarker qualification: Evidentiary framework . Draft guidance for industry and FDA staff. https://www.fda.gov/regulatory-information/search-fda-guidance-documents/biomarker-qualification-evidentiary-framework (2018, accessed 4 October 2025).

Understand translational research tools – biomarkers. https://toolkit.ncats.nih.gov/module/discovery/developing-translational-research-tools/biomarkers/ (accessed 4 October 2025).

Biomarker qualification program. https://www.fda.gov/drugs/drug-development-tool-ddt-qualification-programs/biomarker-qualification-program (accessed 4 October 2025).

Bence-Jones

On a new substance occurring in the urine of a patient with mollities ossium. Philosophical Transactions of the Royal Society. 1848; 138: 55–62.

Abelev

Perova

Khramkova

, et al. [Embryonal serum alpha-globulin and its synthesis by transplantable mouse hepatomas]. Biokhimiia 1963; 28: 625–634.

Primus

Freeman

Goldenberg

. Immunological heterogeneity of carcinoembryonic antigen: purification from meconium of an antigen related to carcinoembryonic antigen. Cancer Res 1983; 43: 679–685.

Taketa

Endo

Sekiya

, et al. A collaborative study for the evaluation of lectin-reactive α-fetoproteins in early detection of hepatocellular carcinoma. Cancer Res 1993; 53: 5419–5423.

Taga

Hirai

Ishizuka

, et al. Early diagnosis of hepatocellular carcinoma with lectin electrophoresis of serum alpha-fetoprotein. Tumor Biol. 1988; 9: 110–115.

10.

Taketa

Hirai

. Lectin affinity electrophoresis of alpha-fetoprotein in cancer diagnosis. Electrophoresis 1989; 10: 562–567.

11.

Kang

. The effectiveness and harms of PSA-based prostate cancer screening: a systematic review. Healthcare 2025; 13: 1381.

12.

National Institute for Health and Care Excellence . Prostate cancer: diagnosis and management. https://www.nice.org.uk/guidance/ng131 (2019, accessed 4 October 2025).

13.

Trulson

Klawonn

Von Pawel

, et al. Improvement of differential diagnosis of lung cancer by use of multiple protein tumor marker combinations. Tumor Biol. 2024; 46: S81–S98.

14.

Yao

Wang

, et al. Multi-biomarkers panel in identifying benign and malignant lung diseases and pathological types of lung cancer. J Cancer 2023; 14: 1904–1912.

15.

Davenport

Rai

Sharma

, et al. Menopausal status, ultrasound and biomarker tests in combination for the diagnosis of ovarian cancer in symptomatic women. Cochrane Database Syst Rev 2022; 7: CD011964.

16.

Piratvisuth

Tanwandee

Thongsawat

, et al. Multimarker panels for detection of early stage hepatocellular carcinoma: a prospective, multicenter, case-control study. Hepatol Commun 2022; 6: 679–691.

17.

Duffy

. Clinical uses of tumor markers: a critical review. Crit Rev Clin Lab Sci 2001; 38: 225–262.

18.

Charkhchi

Cybulski

Gronwald

, et al. CA125 and ovarian cancer: a comprehensive review. Cancers 2020; 12: 3730.

19.

Merriel

SWD

Pocock

Gilbert

, et al. Systematic review and meta-analysis of the diagnostic accuracy of prostate-specific antigen (PSA) for the detection of prostate cancer in symptomatic patients. BMC Med 2022; 20: 54.

20.

Urabe

Kimura

Tashiro

, et al. Prognostic value of PSA bounce in prostate cancer following definitive radiation therapy: a systematic review and meta-analysis. Prostate Cancer Prostatic Dis 2021; 24: 976–985.

21.

Kang

Kim

Hong

, et al. CEA dynamics for predicting response after anti-EGFR monoclonal antibody treatment in metastatic colorectal cancer. Sci Rep 2023; 13: 6735.

22.

Di Gioia

Stieber

Schmidt

, et al. Early detection of metastatic disease in asymptomatic breast cancer patients with whole-body imaging and defined tumour marker increase. Br J Cancer 2015; 112: 809–818.

23.

Bayrak

Yenidunya

Haltas

. Cytokeratin 7 and cytokeratin 20 expression in colorectal adenocarcinomas. Pathol Res Pract 2011; 207: 156–160.

24.

Zhao

Guo

Sneige

, et al. Value of PAX8 and WT1 immunostaining in confirming the ovarian origin of metastatic carcinoma in serous effusion specimens. Am J Clin Pathol 2012; 137: 304–309.

25.

Timek

Shi

Liu

, et al. Arginase-1, HepPar-1, and Glypican-3 are the most effective panel of markers in distinguishing hepatocellular carcinoma from metastatic tumor on fine-needle aspiration specimens. Am J Clin Pathol 2012; 138: 203–210.

26.

Finkelman

Zhang

Hicks

, et al. The evolution of Ki-67 and breast carcinoma: past observations, present directions, and future considerations. Cancers 2023; 15: 808.

27.

Robbins

Bates

Rimm

. HER2 testing: evolution and update for a companion diagnostic assay. Nat Rev Clin Oncol 2025; 22: 408–423.

28.

Zhao

, et al. The role of EGFR mutation as a prognostic factor in survival after diagnosis of brain metastasis in non-small cell lung cancer: a systematic review and meta-analysis. BMC Cancer 2019; 19: 145.

29.

Fang

Wang

. EGFR mutations as a prognostic and predictive marker in non-small-cell lung cancer. Drug Des Devel Ther 2014; 8: 1595–1611.

30.

Prevodnik

Lavrencak

Horvat

, et al. The predictive significance of CD20 expression in B-cell lymphomas. Diagn Pathol. 2011; 6: 33.

31.

Suzuki

Yoshida

Wang

, et al. Association of CD20 levels with clinicopathological parameters and its prognostic significance for patients with DLBCL. Ann Hematol 2012; 91: 997–1005.

32.

Davis

Patel

. The role of PD-L1 expression as a predictive biomarker: an analysis of all US Food and Drug Administration (FDA) approvals of immune checkpoint inhibitors. J Immunother Cancer. 2019; 7: 278.

33.

Diaz

LAJ

Shiu

Kim

, et al. Pembrolizumab versus chemotherapy for microsatellite instability-high or mismatch repair-deficient metastatic colorectal cancer (KEYNOTE-177): final analysis of a randomised, open-label, phase 3 study. Lancet Oncol 2022; 23: 659–670.

34.

Loi

Michiels

Adams

, et al. The journey of tumor-infiltrating lymphocytes as a biomarker in breast cancer: clinical utility in an era of checkpoint inhibition. Ann Oncol 2021; 32: 1236–1244.

35.

Kylma

Sorsa

Jouhi

, et al. Prognostic role of porphyromonas gingivalis gingipain Rgp and matrix metalloproteinase 9 in oropharyngeal squamous cell carcinoma. Anticancer Res. 2022; 42: 5415–5430.

36.

Xin

, et al. Application of fusobacterium nucleatum as a biomarker in gastrointestinal malignancies. World J Gastrointest Oncol. 2024; 16: 2271–2283.

37.

Perou

Sorlie

Eisen

, et al. Molecular portraits of human breast tumours. Nature 2000; 406: 747–752.

38.

Parsons

Jones

Zhang

, et al. An integrated genomic analysis of human glioblastoma multiforme. Science 2008; 321: 1807–1812.

39.

Moller

Schoof

Ligon

, et al. Integration of omics data in the diagnosis and therapy of glioblastoma. Brain Pathol 2025; 36: e70027.

40.

Goodman

Kato

Bazhenova

, et al. Tumor mutational burden as an independent predictor of response to immunotherapy in diverse cancers. Mol Cancer Ther 2017; 16: 2598–2608.

41.

Mayerle

Kalthoff

Reszka

, et al. Metabolic biomarker signature to differentiate pancreatic ductal adenocarcinoma from chronic pancreatitis. Gut 2018; 67: 128–137.

42.

The M371-Test. https://www.mirdetect.de/Produkte/ (accessed 4 October 2025).

43.

The GASTROClear test. https://mirxes.com/oncology/ (accessed 4 October 2025).

44.

Patel

Tirosh

Trombetta

, et al. Single-cell RNA-seq highlights intratumoral heterogeneity in primary glioblastoma. Science 2014; 344: 1396–1401.

45.

Lewis

Asselin-Labat

Nguyen

, et al. Spatial omics and multiplexed imaging to explore cancer biology. Nat Methods 2021; 18: 997–1012.

46.

Kelley

Pantel

. A new era in liquid biopsy: from genotype to phenotype. Clin Chem 2020; 66: 89–96.

47.

Rusnakova

Aziri

Dubovan

, et al. Detection, significance and potential utility of circulating tumor cells in clinical practice in breast cancer (review). Oncol Lett 2025; 29: 10.

48.

Rapanotti

Cenci

Scioli

, et al. Circulating tumor cells: origin, role, current applications, and future perspectives for personalized medicine. Biomedicines 2024; 12: 2137.

49.

Bronkhorst

Ungerer

Oberhofer

, et al. New perspectives on the importance of cell-free DNA biology. Diagnostics 2022; 12: 2147.

50.

Bronkhorst

Holdenrieder

. The changing face of circulating tumor DNA (ctDNA) profiling: factors that shape the landscape of methodologies, technologies, and commercialization. Med Genet 2023; 35: 201–235.

51.

Martinez-Vila

Teixido

Aya

, et al. Detection of circulating tumor DNA in liquid biopsy: current techniques and potential applications in melanoma. Int J Mol Sci 2025; 26: 861.

52.

Garcia-Alfonso

Valladares-Ayerbes

Munoz Martin

, et al. State of the art of the molecular hyperselection to guide treatment with anti-EGFR antibodies in RAS WT mCRC: implications for clinical practice and future perspectives. Expert Opin Biol Ther. 2025; 25: 413–423.

53.

Zhang

Zhu

Liu

, et al. Prognostic and predictive molecular biomarkers in colorectal cancer. Front Oncol 2025; 15: 1532924.

54.

Van der Leest

Rozendal

Rifaela

, et al. Detection of actionable mutations in circulating tumor DNA for non-small cell lung cancer patients. Commun Med 2025; 5: 204.

55.

Ungerer

Bronkhorst

Holdenrieder

. Preanalytical variables that affect the outcome of cell-free DNA measurements. Crit Rev Clin Lab Sci 2020; 57: 484–507.

56.

Parviz Ahmad-Nejad

Bauersfeld

Baum

, et al. Revision of the “Guideline of the German Medical Association on Quality Assurance in Medical Laboratory Examinations – rili-baek”. J Lab Med 2023; 48: 263–306.

57.

Oberhofer

Bronkhorst

Uhlig

, et al. Tracing the origin of cell-free DNA molecules through tissue-specific epigenetic signatures. Diagnostics 2022; 12: 1834.

58.

Bronkhorst

Holdenrieder

. A pocket companion to cell-free DNA (cfdna) preanalytics. Tumor Biol. 2024; 46: S297–S308.

59.

Gene Expression Omnibus. https://www.ncbi.nlm.nih.gov/geo/ (accessed 4 October 2025)

60.

Ensembl. https://www.ensembl.org/index.html (accessed 4 October 2025)

61.

The cancer genome atlas program (TCGA). https://www.cancer.gov/ccg/research/genome-sequencing/tcga (accessed 4 October 2025)

62.

MetaboLights. https://www.ebi.ac.uk/metabolights/ (accessed 4 October 2025)

63.

ENCODE. https://www.encodeproject.org/ (accessed 4 October 2025)

64.

Heo

Hwa

Lee

, et al. Integrative multi-omics approaches in cancer research: from biological networks to clinical subtypes. Mol Cells 2021; 44: 433–443.

65.

Chakraborty

Sharma

Karmakar

, et al. Multi-omics approaches in cancer biology: new era in cancer therapy. Biochim Biophys Acta Mol Basis Dis 2024; 1870: 167120.

66.

Smelik

Zhao

Mansour Aly

, et al. Multiomics biomarkers were not superior to clinical variables for pan-cancer screening. Commun Med 2024; 4: 234.

67.

Cree

. Improved blood tests for cancer screening: general or specific? BMC Cancer 2011; 11: 499.

68.

Cree

Uttley

Buckley Woods

, et al. The evidence base for circulating tumour DNA blood-based biomarkers for the early detection of cancer: a systematic mapping review. BMC Cancer 2017; 17: 697.

69.

Vargas

Harris

. Biomarker development in the precision medicine era: lung cancer as a case study. Nat Rev Cancer 2016; 16: 525–537.

70.

Das

Dey

Devireddy

, et al. Biomarkers in cancer detection, diagnosis, and prognosis. Sensors 2023; 24: 37.

71.

Holdenrieder

Pagliaro

Morgenstern

, et al. Clinically meaningful use of blood tumor markers in oncology. BioMed Res Int 2016; 2016: 9795269.

72.

Debernardi

Blyuss

Rycyk

, et al. Urine biomarkers enable pancreatic cancer detection up to 2 years before diagnosis. Int J Cancer 2023; 152: 769–780.

73.

Debernardi

O'Brien

Algahmdi

, et al. A combination of urinary biomarker panel and pancrisk score for earlier detection of pancreatic cancer: a case-control study. PLoS Med 2020; 17: e1003489.

74.

Fahrmann

Marsh

Irajizad

, et al. Blood-based biomarker panel for personalized lung cancer risk assessment. J Clin Oncol 2022; 40: 876–883.

75.

Jarzab

Litwiniuk

Innis

, et al. The utility of the 21-gene Oncotype DX breast recurrence score® assay in node-negative breast cancer patients - the final analysis of the Polish real-life survey PONDx. Contemp Oncol 2024; 28: 245–252.

76.

Dai

Wang

, et al. Early detection and prognosis prediction for colorectal cancer by circulating tumour DNA methylation haplotypes: a multicentre cohort study. eClinicalMedicine 2023; 55: 101717.

77.

Bulow

Holscher

Costa

, et al. Extending the landscape of omics technologies by pathomics. NPJ Syst Biol Appl 2023; 9: 38.

78.

AI-Assisted diagnostic & biomarker AI. https://www.paige.ai/ (accessed 4 October 2025).

79.

LeeVan

Pinsky

. Predictive performance of cell-free nucleic acid-based multi-cancer early detection tests: a systematic review. Clin Chem 2024; 70: 90–101.

80.

Wade

Nevitt

Liu

, et al. Multi-cancer early detection tests for general population screening: a systematic literature review. Health Technol Assess 2025; 29: 1–105.

81.

Miyo

Chatanaka

EPD

. Implosion of grail’s galleri cancer screening test? eJIFCC 2025; 36: 7–8.

82.

Savage

Messenger

Neal

, et al. Development and validation of multivariable machine learning algorithms to predict risk of cancer in symptomatic patients referred urgently from primary care: a diagnostic accuracy study. BMJ Open 2022; 12: e053590.

83.

Prelaj

Miskovic

Zanitti

, et al. Artificial intelligence for predictive biomarker discovery in immuno-oncology: a systematic review. Ann Oncol 2024; 35: 29–65.

84.

Liu

Panagiotakos

. Real-world data: a brief review of the methods, applications, challenges and opportunities. BMC Med Res Methodol 2022; 22: 287.

85.

European Medicines Agency . Real-world evidence. https://www.ema.europa.eu/en/about-us/how-we-work/big-data/real-world-evidence (accessed 4 October 2025).

86.

Tumor Marker Harmonization (WG-TMH). https://ifcc.org/ifcc-scientific-division/sd-working-groups/tumor-marker-harmonization-wg-tmh/ (accessed 4 October 2025)

87.

Brat

Ellison

Figarella-Branger

, et al. Gliomas, glioneuronal tumours, and neuronal tumours. In: WHO Classification of Tumours. Central Nervous System Tumours. International Agency for Research on Cancer, 2021, vol 6.

88.

Bosse

TDB

Euscher

, et al. Female Genital Tumours. Endometrioid Carcinoma of the Uterine Corpus. WHO Classification of Tumours. IARC, 2020, vol 4, pp. 252–255.

89.

Mapping the evidence for the World Health Organization classification of tumours: a living evidence gap map by tumour type. https://wct-evi-map.iarc.who.int/ (accessed 4 October 2025).

90.

Del Aguila Mejia

Armon

Campbell

, et al. Understanding the use of evidence in the WHO classification of tumours: a protocol for an evidence gap map of the classification of tumours of the lung. BMJ Open. 2022; 12: e061240.

91.

Md Nasir

Koh

VCY

Cree

, et al. Phyllodes tumour evidence gaps mapped from the 5th edition of the WHO classification of tumours of the breast. Histopathology 2023; 82: 704–712.

92.

Orive

Echepare

Bernasconi-Bisio

, et al. Protein biomarkers in lung cancer screening: technical considerations and feasibility assessment. Arch Bronconeumol 2024; 60(Suppl 2): S67–S76.

93.

The International Standards Organization. https://www.iso.org/home.html (accessed 4 October 2025).

94.

The International Quality Network for Pathology. https://www.iqnpath.org/ (accessed 4 October 2025).

95.

U.S. Food and Drug Administration . https://www.fda.gov/ (accessed 4 October 2025).

96.

National Institutes of Health . https://www.nih.gov (accessed 4 October 2025).

97.

European Medicines Agency . https://www.ema.europa.eu/en/homepage (accessed 4 October 2025).

98.

World Health Organization . https://www.who.int (accessed 4 October 2025).

99.

Michiels

Shyr

, et al. Biomarker discovery and validation: statistical considerations. J Thorac Oncol 2021; 16: 537–545.

100.

Van Rossum

Holdenrieder

Ballieux

, et al. Investigating the current harmonization status of tumor markers using global external quality assessment programs: a feasibility study. Clin Chem 2024; 70: 669–679.

101.

Savva

Kawka

Vadhwana

, et al. The biomarker toolkit – an evidence-based guideline to predict cancer biomarker success and guide development. BMC Med 2023; 21: 383.

102.

Enhancing the QUAlity and Transparency Of health Research (EQUATOR) . https://www.equator-network.org/ (accessed 2 December 2025).

103.

GRADE handbook. https://gdt.gradepro.org/app/handbook/handbook.html (accessed 2 December 2025).

104.

Hall

Brown

Paul

. An exploration into study design for biomarker identification: issues and recommendations. Cancer Genom. Proteom. 2007; 4: 111–119.

105.

Ruppert

, et al. Discussion of trial designs for biomarker identification and validation through the use of case studies. JCO Precis Oncol 2019; 3: PO.19.00051.

106.

Vatkar

Kale

Shyam

, et al. Understanding the levels of evidence in medical research. J Orthop Case Rep 2025; 15: 6–9.

107.

Colling

Indave

Del Aguila

, et al. A new hierarchy of research evidence for tumor pathology: a Delphi study to define levels of evidence in tumor pathology. Mod Pathol 2024; 37: 100357.

108.

Chechlinska

Kowalewska

Plans-Beriso

, et al. The adapted “hierarchy of evidence in tumour pathology” for the classification of biomarker studies. Tumor Biol. 2025; 47: P71.

109.

Craciun

García-Ovejero

Campbell

, et al. Development of a decision tree diagram for classifying study designs in pathology research: a multidisciplinary approach. J Pathol Clin Res 2026; 12: e70056.

110.

Standards for reporting diagnostic accuracy. Available at: https://www.equator-network.org/reporting-guidelines/stard/ (accessed 4 October 2025).

111.

McShane

Altman

Sauerbrei

, et al. REporting recommendations for tumour MARKer prognostic studies (REMARK). Br J Cancer 2005; 93: 387–391.

112.

Colling

Indave

Del Aguilla

, et al. Moving forward on tumor pathology research reporting: a guide for pathologists from the World Health Organization classification of tumors living evidence gap map by tumour type group. Mod Pathol. 2024; 37: 100515.

Tumour markers and evidence-based pathology

Abstract

Background

Objective

Methods

Results

Conclusions

Keywords

Introduction

Objective

Classical tumour biomarkers

Historical view

Further developments

Accuracy issues

Novel classes of tumour markers

Molecular histopathological markers

Immunotherapeutic-context-related tissue markers

Bacteria-specific markers

Markers derived from high-throughput methods

Markers generated from single-cell analyses

cfDNA/ctDNA and circulating tumour cells as liquid biopsy markers

Multi-omic approaches

Improving testing accuracy: Meta-biomarkers

Panels of classical tumour markers

Multigene diagnostic tests

Markers combined with clinical data and structural insights

Multi-cancer early detection panels

Biomarkers for tumour classification

Biomarker development guidelines

Identification of a tumour biomarker – Study design issues

Tumour Biomarker Reporting and Evidence Mapping

Conclusions

Footnotes

ORCID iDs

Collaborators

Funding

Declaration of conflicting interest

Disclaimer

References