Author: Elena Rodríguez
Personalized medicine is on the rise, and correctly identifying the characteristics of each illness in each person is increasingly important. Biomarkers are biological characteristics that provide information about a person’s health status. They are used to estimate the risk of developing a disease, identify it, track its progression, predict the response to a certain treatment or detect its adverse effects. Therefore, a biomarker contributes to clinical decision-making.
What is a biomarker?
Biomarkers have been used almost since the beginning of medicine, although the term appeared for the first time in 1973 (1). The medical definition came in 2001, from the NIH Biomarkers Definitions Working Group, as a characteristic that is measured and evaluated objectively as an indicator of normal biological processes, pathogenic processes or pharmacologic responses to a therapeutic intervention (2). The work was expanded afterwards with the BEST (Biomarkers, EndpointS, and other Tools) glossary, which maintains the same definition and includes two specifications: biomarkers can be molecular, histological, radiographic or physiological characteristics, and a biomarker is a measure of an underlying biological process, not an indicator of how an individual feels, functions or survives (3).
BEST distinguishes different categories that serve to identify who may develop a disease (susceptibility or risk), who has the disease (diagnostic), what can happen when a disease is present (prognostic), and who can benefit from an intervention (predictive), as well as to follow the course of a disease (monitoring), the response to a treatment (response or pharmacodynamic), and the toxicity (safety). Moreover, it is possible to combine several measurements into a composite biomarker, such as the FIB-4 index, which combines age, transaminases and platelets to estimate the risk of advanced hepatic fibrosis before using invasive methods such as biopsies (4).
Although it falls outside the scope of this post, it is worth mentioning that not all biomarkers solely provide information about a person’s biological characteristics: some are also the target on which drugs act.
What makes a good biomarker?
It is normally assumed that a biomarker is good when it discriminates between group A and group B. However, this is not always the case. Many times, the groups A and B belong to homogeneous populations which have been selected with strict criteria, so it is difficult to know if the result will hold when we apply the biomarker to other populations. In addition, even when the discriminatory capacity is maintained, the threshold to change from one group to another may vary.
An illustrative example is troponin as a diagnostic biomarker of acute myocardial infarction. Circulating troponin levels are generally lower in women than in men, so the use of a single threshold can reduce the diagnostic sensitivity in women and contribute to their underdiagnosis. For this reason, the Fourth Universal Definition of Myocardial Infarction recommends using sex-specific thresholds (5). Nevertheless, this recommendation has been adapted differently in the algorithms of the European and American guidelines (6,7). In other words, the same biomarker with the same indication can be handled with different thresholds depending on where the doctor is located.
In contrast, the creatinine-based glomerular filtration rate equation, which traditionally included an adjustment coefficient for black people, could overestimate their glomerular filtration rate. As a consequence, clinical decisions such as the classification of kidney disease and referral to nephrology could be delayed. Removing race as a correction factor in clinical practice has shown that the calibration of a biomarker can change who receives an intervention, and when (8).
Therefore, the validation of a biomarker does not end when an association with a disease is demonstrated, or when it appears in a clinical practice guideline. It must continue to be demonstrated that the biomarker contributes to clinical decisions that improve health outcomes.
How are biomarkers evaluated?
New candidate biomarkers are proliferating in the race towards personalized medicine, but many of them do not reach clinical practice. As has been pointed out, the problem is not identifying new biomarkers, but ensuring that they are relevant to decision-making, which requires analytical and clinical validity, cost-effectiveness and organizational impact (9).
What is more striking in this process is the lack of a regulatory evaluation of the biomarker in clinical practice. Transaminases or troponin are part of the day-to-day practice in the Sanish Health System laboratories because the common portfolio of services covers tests and diagnostic procedures in generic terms, but not because there is a list of approved biomarkers. The biomarkers to use and their thresholds are proposed by scientific societies in their clinical practice guidelines, which are then adopted by the hospital services. Given the lack of an independent organization that systematically evaluates the biomarkers used in clinical practice and their thresholds, there is no standard procedure to identify and correct a biomarker or a threshold that has been inadequately established.
The scenario changes when the biomarker is involved in the development or use of a therapy. The EMA has a biomarker qualification procedure for specific uses, although it is a rarely used route -86 applications were submitted between 2008 and 2020, of which only 13 were qualified (10)- and not equivalent to an authorization for use in clinical practice. In addition, the summaries of product characteristics include information on biomarkers associated with therapies. The AEMPS has created a database with pharmacogenomic biomarkers in order to facilitate access to this information and promote the implementation of pharmacogenetics in the Spanish Health System (11). In any case, inclusion in this database does not entail any validation beyond what has already been done during the evaluation of the therapy.
What is the future of biomarkers?
The capacity to characterize each patient even better has opened the door to the development of many biomarkers. Here, artificial intelligence plays an important role, because it is not only used to discover new biomarkers but especially to integrate and extract information from biomarkers and tests that are already used in clinical practice (12).
One route to optimization is biomarker integration, either of the same or different types -imaging and molecular biomarkers, for example- into a single composite biomarker that improves on the diagnostic or predictive capacity of each one alone. Moreover, there are attempts to extract additional information from diagnostic tests that are carried out routinely, with the final goal of identifying characteristics or patterns that are not seen in traditional analyses. For example, from histological preparations, magnetic resonance images or ultrasound scans.
This would be a great advance in medicine, not just because of a better characterization of patients, but also because, by extracting new information from tests which are already done in routine care, more equitable access is promoted. Nevertheless, most of the algorithms developed with artificial intelligence are sensitive to the population on which they have been trained and can perform differently when they are validated in other populations, even if they are apparently similar. Moreover, their complexity can make the evaluation of their calibration and their subsequent recalibration challenging (13).
Going back to our premise, the challenge is not to extend the catalogue of biomarkers, but to determine which of them help make decisions that improve patient management without introducing bias between populations or subgroups. This is not just a clinical decision: it is also an economic and social one. A well-chosen and well-calibrated biomarker avoids unnecessary tests, ineffective treatments and late diagnoses, while a poorly implemented biomarker misuses resources and may prevent certain subgroups from accessing those decisions. All of this demands a systematic evaluation of their utility before and during their implementation in clinical practice, considering not just their diagnostic and predictive performance, but also their impact on health outcomes, resources, and equity in access to the associated decisions.
References
1. Rho JH, Bauman AJ, Boettger HG, Yen TF. A search for porphyrin biomarkers in Nonesuch Shale and extraterrestrial samples. Space Life Sci. 1973;4(1):69-77. doi:10.1007/BF02626343
2. Definitions Working Group. Biomarkers and surrogate endpoints: preferred definitions and conceptual framework. Clin Pharmacol Ther. 2001;69(3):89-95. doi:10.1067/mcp.2001.113989
3. FDA-NIH Biomarker Working Group. BEST (Biomarkers, EndpointS, and other Tools) Resource [Internet]. Silver Spring (MD): Food and Drug Administration (US); 2016 [citado 24 de agosto de 2026]. Disponible en: http://www.ncbi.nlm.nih.gov/books/NBK326791/ PubMed PMID: 27010052.
4. Association for the Study of the Liver (EASL), European Association for the Study of Diabetes (EASD), European Association for the Study of Obesity (EASO). EASL-EASD-EASO Clinical Practice Guidelines on the management of metabolic dysfunction-associated steatotic liver disease (MASLD). J Hepatol. 2024;81(3):492-542. doi:10.1016/j.jhep.2024.04.031
5. Thygesen K, Alpert JS, Jaffe AS, Chaitman BR, Bax JJ, Morrow DA, et al. Fourth universal definition of myocardial infarction (2018). Eur Heart J. 2019;40(3):237-69. doi:10.1093/eurheartj/ehy462
6. Gulati M, Levy PD, Mukherjee D, Amsterdam E, Bhatt DL, Birtcher KK, et al. 2021 AHA/ACC/ASE/CHEST/SAEM/SCCT/SCMR Guideline for the Evaluation and Diagnosis of Chest Pain: Executive Summary: A Report of the American College of Cardiology/American Heart Association Joint Committee on Clinical Practice Guidelines. Circulation. 2021;144(22):e368-454. doi:10.1161/CIR.0000000000001030
7. Byrne RA, Rossello X, Coughlan JJ, Barbato E, Berry C, Chieffo A, et al. 2023 ESC Guidelines for the management of acute coronary syndromes. Eur Heart J. 2023;44(38):3720-826. doi:10.1093/eurheartj/ehad191
8. Disease: Improving Global Outcomes (KDIGO) CKD Work Group. KDIGO 2024 Clinical Practice Guideline for the Evaluation and Management of Chronic Kidney Disease. Kidney Int. 2024;105(4S):S117-314. doi:10.1016/j.kint.2023.10.018
9. AR, Lord SJ, StJohn A, Sandberg S, Cobbaert CM, Lorenz S, et al. From biomarkers to medical tests: the changing landscape of test evaluation. Clin Chim Acta. 2014;427:49-57. doi:10.1016/j.cca.2013.09.018
10. E, Hendrikse NM, Ehmann F, van der Meer DS, Llinares Garcia J, Vetter T, et al. Biomarker Qualification at the European Medicines Agency: A Review of Biomarker Qualification Procedures From 2008 to 2020. Clin Pharmacol Ther. 2022;112(1):69-80. doi:10.1002/cpt.2554
11. de datos de biomarcadores farmacogenómicos | AEMPS. Agencia Española de Medicamentos y Productos Sanitarios [Internet]. [citado 27 de agosto de 2026]. Disponible en: https://www.aemps.gob.es/medicamentos-de-uso-humano/base-de-datos-de-biomarcadores-farmacogenomicos/
12. G, Naimimohasses S, Rambhatla S, Komorowski M, Ferro D, Lewis PR, et al. Responsible adoption of multimodal artificial intelligence in health care: promises and challenges. Lancet Digit Health. 2025;7(12):100917. doi:10.1016/j.landig.2025.100917
13. GS, Moons KGM, Dhiman P, Riley RD, Beam AL, Van Calster B, et al. TRIPOD+AI statement: updated guidance for reporting clinical prediction models that use regression or machine learning methods. BMJ. 2024;385:e078378. doi:10.1136/bmj-2023-078378
