What Are Cancer Biomarkers and Why Do They Matter?

A biomarker, short for biological marker, is anything that can be objectively measured and used as an indicator of a normal process, a disease process, or a response to treatment. In oncology the term usually refers to something measured inside the tumour itself or in a body fluid that reflects the tumour.

Biomarkers exist at several biological levels. At the DNA level, they include mutations (a spelling error in a gene, such as KRAS G12C), amplifications (too many copies of a gene, such as HER2 amplification), deletions, and fusions or rearrangements in which two genes join abnormally, as happens with ALK. At the protein level, biomarkers include the amount of a protein displayed on the surface of cancer cells, measured for example as a PD-L1 expression score or a HER2 immunohistochemistry score. At the level of broader signatures, biomarkers include microsatellite instability (MSI), mismatch repair deficiency (dMMR), tumour mutational burden (TMB) and homologous recombination deficiency (HRD) — patterns that summarise how a tumour's DNA repair machinery behaves rather than pointing at one single gene.

An important distinction is where the change came from. Somatic biomarkers are acquired inside the tumour during a person's life and are present only in cancer cells; they cannot be passed on to children. Germline biomarkers are inherited variants present in every cell of the body, which is why they matter to blood relatives as well as to the patient. BRCA1 and BRCA2 can appear in either category, which is one reason biomarker reports must be read carefully rather than skimmed.

For you as a patient, biomarker testing has three practical consequences. First, access: a positive predictive marker can unlock a targeted tablet or an immunotherapy infusion that would otherwise never be offered, sometimes with better tolerability than chemotherapy. Second, avoidance: a result showing that a drug class will not work — for example RAS mutation in colorectal cancer ruling out anti-EGFR antibodies — protects you from side effects and lost time. Third, options: many clinical trials, including early-phase studies of the newest agents, select participants by biomarker status, so a molecular profile is often the key that makes a trial possible.

It is worth knowing that testing is not yet uniform. Coverage, panel size, turnaround time and reimbursement differ substantially between hospitals and between countries, and re-testing after progression is not automatic everywhere. Asking direct questions about what was tested is not second-guessing your team; it is standard practice in modern oncology, and it is the single most useful thing many patients can do to broaden their own options.

Biomarker testing usually begins with a sample of tumour tissue obtained by biopsy or during surgery. The tissue is fixed in paraffin and stored, which means archived material from an earlier procedure can often be re-tested without a new operation, provided enough good-quality tumour remains. When tissue is insufficient, unsafe to obtain, or out of date because the disease has evolved, a liquid biopsy may be used instead: a blood sample that captures circulating tumour DNA shed into the bloodstream.

Several laboratory methods are used, each suited to different questions. Immunohistochemistry (IHC) stains a tissue slide with antibodies to show how much of a protein is present, and is the standard first step for HER2 and PD-L1. Fluorescence in situ hybridisation (FISH) counts gene copies or detects rearrangements and is often used to confirm equivocal HER2 results or ALK fusions. Polymerase chain reaction (PCR) detects a defined list of known mutations quickly and cheaply. Next-generation sequencing (NGS) reads many genes in parallel and is the basis of comprehensive genomic profiling; a broad panel can report mutations, fusions, amplifications, MSI and TMB from one sample.

Turnaround time is typically one to three weeks, longer for large panels or when samples must be shipped to a reference laboratory. Results arrive as a molecular pathology report listing each variant, its precise notation, its clinical significance, and often the therapies and trials associated with it. Variants of uncertain significance are common and do not mean something is wrong with the report; they simply mean the evidence is not yet sufficient to act on that finding.

Some biomarkers matter mainly in one disease, while others are tumour-agnostic and relevant wherever they appear. The examples below are the ones patients encounter most often.

KRAS is one of the most frequently mutated genes in cancer. It acts as an on/off switch for growth signalling, and a mutation locks the switch on. KRAS mutations occur in roughly nine out of ten pancreatic ductal adenocarcinomas, in about a third to a half of colorectal cancers, and in around a quarter of non-small cell lung cancers. The exact variant matters: G12C is targetable with approved inhibitors in several settings, while other variants such as G12D are the focus of intense trial activity.

EGFR, the epidermal growth factor receptor, drives growth when mutated. Activating EGFR mutations — most commonly exon 19 deletions and the L858R point mutation — are found in a substantial minority of non-small cell lung adenocarcinomas and are markedly more common in patients of East Asian background and in those who never smoked. They predict strong benefit from oral EGFR tyrosine kinase inhibitors.

HER2 is a growth-signalling receptor that becomes overactive when the gene is amplified or the protein is overexpressed. It is classically associated with about 15 to 20 percent of breast cancers and a subset of gastric and oesophageal cancers, and HER2-directed antibodies and antibody-drug conjugates have transformed outcomes in these groups. HER2 alterations also occur in lung, colorectal, biliary and other cancers, and low-level HER2 expression has recently become clinically actionable in breast cancer.

BRCA1 and BRCA2 are DNA repair genes. When either stops working, cells lose an important mechanism for fixing double-strand DNA breaks. Inherited faults raise lifetime risk of breast, ovarian, prostate, pancreatic and some other cancers, and they also occur as somatic changes inside tumours. The same repair defect makes these cancers unusually sensitive to platinum chemotherapy and to PARP inhibitors — a strategy known as synthetic lethality.

PD-L1 is a protein some tumours display to switch off the immune cells that would otherwise attack them. It is measured by immunohistochemistry and reported as a score such as TPS or CPS, with thresholds that differ by cancer type and by drug. High PD-L1 expression increases the likelihood of benefit from immune checkpoint inhibitors, particularly in lung cancer, head and neck cancer, gastric cancer and several others, although it is an imperfect predictor and some low-expressing tumours still respond.

Alongside these, MSI-high and mismatch repair deficient status deserve mention because they are tumour-agnostic: wherever they occur, in colorectal, endometrial, gastric or other cancers, they predict a meaningful chance of benefit from immunotherapy.

Biomarker results influence treatment in four recurring ways. They can select a targeted therapy, as when an EGFR mutation leads to an oral tyrosine kinase inhibitor as first-line treatment rather than chemotherapy. They can select an immunotherapy, as when high PD-L1 expression or MSI-high status supports a checkpoint inhibitor. They can exclude a therapy, as when a RAS mutation in colorectal cancer shows that anti-EGFR antibodies would not help. And they can open a clinical trial, since molecularly selected studies are now the norm in early-phase oncology.

Sequencing of therapies also depends on biomarkers. In many targeted settings, treatment continues until resistance emerges, at which point repeat testing on new tissue or on circulating tumour DNA may reveal a resistance mechanism and point to a next-generation drug rather than a return to chemotherapy. This is why biomarker testing is not a single event at diagnosis but a process that can be repeated at key decision points.

A biomarker result should always be interpreted alongside the rest of the clinical picture: stage, organ function, previous treatments, performance status, symptoms and personal priorities. No marker on its own dictates a plan. Discussion in a multidisciplinary tumour board, and a second opinion at a centre experienced with the specific alteration, are both reasonable steps when a result is unusual or when the recommended option is not available locally.