Biomarkers are not validated or standardised

A Japanese trial found vitamin D supplements did not help everyone after digestive cancer surgery, but appeared to help a subgroup identified by a tumour marker. That subgroup deserves its own trial.

Before cancer spreads, distant organs are changed to become welcoming. A test for those changes would show which organ is at risk while a person still looks cancer-free.

Volatile compounds in breath differ in cancer. A breath test validated in truly symptomatic patients, not lab volunteers, could tell GPs who needs urgent scans and who can safely wait.

Blood and tissue samples collected in cancer trials are the best material for validating new tests, but most sit unused under contracts that make access impossible. A commons would make them available for approved research.

Completed trials have stored samples. Testing a new biomarker on them with the plan written in advance is nearly as good as a new trial and far cheaper, but regulators have no clear route to accept it.

Tests for 'HRD', which decide who gets PARP inhibitors, rely on genomic scars that reflect the tumour's past, not its present. A test of current DNA-repair function would be better, but needs standardising.

Thousands of tests that could predict who benefits from a treatment are published and never validated. A fund would pay for the boring but essential confirmation studies in independent patient groups.

Thousands of cancer biomarkers are published; almost none reach patients because nobody validates them fairly. Create a public service that tests any candidate blind against stored samples.

Targeted cancer drugs often reach a country years before the test needed to select patients is approved there. Recognising other regulators' test approvals would close the gap.

Between diagnosis and surgery there are usually a few weeks. Giving a new drug in that window and comparing the tumour before and after surgery shows whether it hits its target in real people, quickly and cheaply.

Tumour mutational burden decides who gets immunotherapy in some settings, but every sequencing panel calculates it differently. A shared calibration would make the number mean the same thing everywhere.

Some tumours change fast and escape drugs quickly; others are stable. A single validated score for how evolvable a tumour is would tell doctors how aggressively to combine treatments.

Two people can have identical treatment and only one develops heart failure or a second cancer years later. Collecting blood and genetic data from survivors could reveal who is at risk and who can be reassured.

Some people have only a few sites of spread and can be treated at each one; others have further deposits not yet visible, and counting lesions cannot tell them apart. A signature built from tumour DNA levels, microRNA classifiers and clonal diversity across lesions would define the biology and spare futile ablation.

Once a year, send the same blinded blood samples to every company selling a tumour-DNA test and publish how each performed.

Trials often measure a stand-in for survival, such as time until the cancer grows on scans. An independent body would test, for each cancer and treatment type, whether the stand-in actually predicts survival, and publish the answer.

Three genetic classifications of the commonest aggressive lymphoma exist and none of them yet decides anyone's treatment. The trial that would change that has not been run.

Whether a breast cancer counts as HER2-low, and so qualifies for trastuzumab deruxtecan, turns on the least reproducible step of the HER2 stain, score 1+ versus 0. Cell-line microarrays with a known quantity of HER2 protein per cell, run alongside clinical slides, would anchor every laboratory to a physical standard.

Dozens of companies sell blood tests for tumour DNA and they report different results on the same sample. Government-issued reference samples with known amounts of tumour DNA would expose the differences.

Some tumours have broken the machinery that displays their identity to immune cells. Those patients cannot benefit from most immunotherapy and should be routed elsewhere.

Cancer cells travelling in the blood can be counted. If a drug clears them, that is an early sign it may stop spread, and it reads out in weeks rather than years.

After pre-surgery chemotherapy leaves tumour behind, a blood test for tumour DNA triples the risk of distant relapse when positive. The one trial that acted on it found the DNA usually appeared too late. A trial that tests at surgery with a tumour-informed assay, escalates positives to an antibody-drug conjugate and observes negatives has not been run.

A drug label says 'for PD-L1 positive patients' but does not say that other tests give different answers. Labels should list the validated tests and how much they disagree.

Several agents have been tried for thymic recovery after transplant and none is in routine use. The blocker is as much the missing endpoint as the missing drug.

If a blood test reliably shows whether cancer will come back after surgery, trials could use it instead of waiting years for relapse. Regulators have a process to bless such a test; oncology should use it.

Three PD-L1 tests run on the same tumours called 46, 75 and 73 percent of them positive and agreed only 69 percent of the time. With atezolizumab withdrawn, pembrolizumab's test (22C3, combined positive score of 10) is the only one that matters, yet laboratories still run whichever kit they have. One assay, one score and a proficiency scheme would end answers that vary by postcode.

Lab-grown mini-tumours are already being sold to guide treatment, but the tests are not validated like other medical tests. They should be.

Drugs are labelled phase 1, 2 or 3 so everyone knows how proven they are. Biomarkers should carry a comparable grade, from B1 discovery to B5 utility shown in a randomised trial, so that a marker with only discovery-stage evidence is not mistaken for a validated one in guidelines and papers.

A drug aimed at a mutation present in every tumour cell works differently from one aimed at a mutation in only some cells. Test reports should say which is which.

The detailed molecular maps of tumours being built today mostly lack information on what happened to the patient. Require every atlas sample to carry consented outcome data.

Companies sometimes choose the biomarker threshold that makes their trial look best after seeing the data. Requiring the threshold to be fixed and published before the big trial starts prevents this.

When a screening test misses a cancer, we should know. Linking every negative result to the cancer registry and publishing what was missed, by stage, should be a condition of use.

The support cells that build a tumour's scaffolding come in several types: some protect the tumour, others restrain it. Treating all of them the same way explains past failures.

Screening finds cancers that would never have caused harm alongside dangerous ones. Pair every screening test with a test that says which is which, so people with harmless findings can safely watch and wait.

If one lab suddenly reports twice the rate of 'positive' biomarker results that other labs report, its assay has probably drifted. Pooling anonymised positivity rates by laboratory, assay and version, with automated outlier detection and case-mix adjustment, would catch reagent lot problems and protocol drift within weeks rather than at occasional proficiency runs.

Thousands of papers extract 'radiomic' features from scans to predict outcomes, but the features change with scanner settings. Journals should require standard compliance before any clinical claim is made.

There are several different PD-L1 tests, each tied to a different drug, and they disagree. A single digitally calibrated scale would let any lab's result be translated into any drug's cut-off.

A substantial share of patients with advanced non-small-cell lung cancer start first-line treatment before their biomarker results arrive, or without full testing, and so miss targeted therapy. Letting the pathologist order the full genomic panel the moment a cancer of a defined type and stage is confirmed shortens time to result and makes testing complete.

Most genomic and liquid biopsy tests are reimbursed on analytical validity and association with outcome, not on proof that they improve care. Paying for new oncology biomarker tests only inside registries or randomised studies, as Medicare did for PET, would sort the useful from the useless.

Drug trials must be registered before they start so results cannot be hidden or reshaped. Studies that claim a biomarker predicts outcome should be registered too.

Anyone building a new test for HER2, PD-L1 or tumour DNA should be able to check it against the same public reference set. Today each developer validates on private data nobody can inspect.

Prostate MRI is good at telling you whether a man has a serious cancer and poor at telling you where all of it is. It missed at least one significant tumour in a third of men in the study that checked it against the whole removed prostate. Services that treat part of the gland, or follow men on imaging alone, are relying on the number they do not measure.

Labs already get tested on whether they score biomarkers correctly, but the results are private. Publishing them would let hospitals and patients avoid labs that get it wrong.

Trials often have independent radiologists check whether tumours grew. How often they disagree with the treating doctors is a measure of how trustworthy the result is, and it is rarely published.

Current blood tests for leftover cancer track a few dozen mutations and miss low-level disease. Whole-genome and error-corrected methods integrate signal across thousands of tumour-specific sites plus methylation and fragment features, reaching detection near one part per million in research settings; cost, turnaround and reproducibility are the barriers.

Epigenetic clocks and senescence markers run fast after cancer treatment, and nobody has shown that moving one matters. The way to find out is to make the clock a secondary endpoint in a trial whose primary endpoint is physical function.

PSMA scans could measure prostate cancer burden and response far better than PSA, but no one has done the standardisation work to make the measurement trustworthy across scanners.

Patients are denied a drug when a test says they will not benefit, but that restriction is usually inferred from enrichment trials rather than tested. For high-stakes markers with weak evidence in the negative group, such as PD-L1 and HER2 0, randomised trials in biomarker-negative patients should test the assumption itself.

Fluid taken from the lower back contains DNA from brain tumours. Testing it can diagnose, monitor and detect resistance without brain surgery.

How a tissue sample is handled before it reaches the lab changes the results of biomarker tests. That handling time is almost never recorded, so nobody can tell a true negative from a spoiled sample.

Different companies' leftover-cancer blood tests disagree, and there is no shared yardstick. Public reference samples would let anyone check which test actually works.

Between one in eleven and one in five gallbladder cancers is HER2-positive and two HER2 drugs are now approved, but testing still happens only when someone asks. Making it automatic on every advanced biliary diagnosis, on resection tissue where it exists, would find the patients the trials were built for.

About a third of triple-negative tumours are HER2-low and so eligible for trastuzumab deruxtecan, but the difference between a HER2 score of 0 and 1+ is the one pathologists agree on least, and most triple-negative tumours were scored before the label mattered. Re-scoring archived slides with digital help when a patient relapses would find the eligible third.

A test approved to select patients for a drug in the US must go through separate approval in Europe, Japan and elsewhere, delaying the drug. Accepting each other's test approvals would fix the delay.

Trials that fail still collected thousands of blood and tissue samples. Instead of being destroyed, they should be pooled so scientists can learn who might have benefited.

About 1 in 100 people who develop diabetes after 50 has a pancreatic cancer behind it. A score built from weight change, blood sugar change and age, calculable from records already in primary care, picks out a group where the rate is nearer 1 in 30; the proposal is to scan that group rather than wait for symptoms.

Instead of testing a drug in lab models first and hoping the results carry over, build the same models from trial participants and run both experiments in parallel to see how well the models predict.

Wasting has several causes. Measuring the specific hormone in each patient's blood would put the right patients into the right trial instead of mixing everyone together.

Ninety-five percent of advanced bowel cancers ignore immunotherapy, and the only real signal so far came in patients without active liver secondaries. Trials keep enrolling by treatment line rather than by immune biology, which guarantees the responders are diluted away.

Growing a patient's own tumour-fighting cells only works if those cells are there to start with. A test for them would spare futile treatment.

New microscopes can measure dozens of proteins at once and map where immune cells sit in a tumour. These readouts could predict immunotherapy response, but every lab does it differently.

For each cancer type, agree the set of stains and tests that are always needed, and have the lab run them automatically on diagnosis rather than waiting for someone to ask.

Cancers driven by MYC need to make proteins at an unusually fast rate. Slowing the cell's protein factory hits them harder than it hits normal cells.

A blood test after surgery already lets stage II colon cancer patients skip chemotherapy safely. The same test in stage III would spare far more people, and the unproven half, giving more chemotherapy to those who test positive, has so far changed nothing.

Cancer cells enter the blood mostly during rest, so a morning blood test may miss them. Sampling and dosing at the right hour may be a free improvement.

For rare cancers there is often no genetic clue and no trial to join. Testing drugs directly on the patient's fresh tumour cells has guided treatment with reported benefit in a randomised haematology study and in paediatric case series; turnaround and tissue quality are the barriers.

About one man in eight with prostate cancer that has spread carries an inherited DNA repair fault, and about one in five has one in the tumour. The drugs for those faults have moved to the beginning of treatment, but the test is still usually done near the end, when it is too late to use the result.

Genetic testing often returns a variant of uncertain significance that cannot be acted on, most often in people of non-European ancestry. Saturation genome editing has already classified nearly all BRCA1 single-nucleotide variants; a consortium doing the same for the roughly 30 actionable hereditary cancer genes would end most uncertain results.

Trials that only enrol patients with a positive biomarker can never prove the biomarker matters. Including a smaller biomarker-negative group would show whether the test is really needed.

Whether immune cells are next to cancer cells matters more than how many there are. Turning that spatial picture into a reliable, standardised test would predict response better.

Tumours starved of oxygen produce a chemical that switches immune cells off. A scan can show which tumours are starved, and those are the ones to treat with blockers.

A blood test detects lymphoma left behind better than a scan does, and no trial has yet used it to change anyone's treatment.

A single protein predicts whether a tumour will respond to DNA-damaging drug payloads. Measuring it could stop patients receiving a second drug of the same kind that will not work.

Pragmatic trials want to use the progression dates recorded in ordinary clinic notes instead of expensive protocol scans. Checking how well those routine records match formal trial measurements would show when that shortcut is safe.

AI is starting to decide which patients get which cancer drug. Every change to the software should be tested against a fixed public set of cases before it is used on patients.

In a minority of men, prostate cancer escapes hormone drugs by becoming a different kind of cell that no longer needs the androgen receptor. By the time a biopsy shows it, the treatment options are almost gone. The genetic changes that allow the switch are detectable years earlier, and nobody is looking for them.

When immunotherapy works, specific immune cell families multiply in the blood within weeks. Tracking that could tell patients early whether to continue.