precision

A drink taken before brain surgery that makes high-grade glioma glow under blue light, letting surgeons remove more tumour safely.

People with cancer get blood clots more often than others. A score built from the cancer type and routine blood tests picks out those at higher risk, and in two placebo-controlled trials a low-dose blood-thinning tablet reduced clots but caused more bleeding. United States guidance says it may be offered; surveys in the Netherlands and China found most specialists do not use the score.

The bladder and rectum change size from day to day and move the prostate with them. A randomised trial found that drinking about half a litre of water before each treatment kept the bladder more consistent than drinking a litre, and small trials of anti-wind medicines before prostate radiotherapy found no effect.

Re-shaping the treatment plan to the anatomy of the day, using the images taken on the treatment couch, so the dose follows a shrinking tumour or a moving bladder.

Software that draws organs and tumours on scans automatically, saving hours per patient and making daily plan adaptation practical.

Software that reads digitised biopsy slides to detect cancer, grade it, and score biomarkers such as HER2, PD-L1 and Ki-67 more consistently than the eye alone.

When colorectal cancers labelled mismatch-repair deficient did not respond to immunotherapy, three of five turned out on retesting not to be deficient at all; local results were wrong about one time in ten.

In ten hard-to-treat cancer types, 29 in 100 people whose tumours carried at least 10 mutations per million DNA letters responded to pembrolizumab, against 6 in 100 below that line.

The first randomised comparison of protons with modern photon radiotherapy for lung cancer. Protons spared the heart but not the lung overall, and neither severe lung inflammation nor local failure was reduced: 10.5 per cent against 6.5 per cent for lung inflammation, 10.5 against 10.9 per cent for local failure.

Cancer tablets such as kinase inhibitors are given at one dose for everyone, though the amount in the blood differs widely between people. A Dutch study of 600 patients measured trough levels and adjusted when they were low, cutting the share of people with too little drug by 39 per cent, but no trial has yet shown that this improves survival.

Chemotherapy doses are usually written per square metre of body surface, a convention from 1958 that scales drug clearance between species and people; it is imprecise, and for many newer drugs flat or weight-based doses have replaced it.

Breath-hold is usually taught on the day of the planning scan. Two hospital series found that teaching it at least five days earlier, with practice at home, went with lower heart doses and shorter set-up times, and a survey of American radiation oncologists found interpreters were present at fewer than one in three first treatments for people who do not speak English.

Carboplatin is dosed by kidney function rather than body size, using a formula from 1989 that aims at a set drug exposure. The formula is only as good as the kidney measurement fed into it: in a study of 2,471 people with cancer, the usual blood-test estimates missed the intended dose by more than a fifth in about one person in five to one in four.

Short pieces of DNA floating in the blood, shed by dying cells all over the body. Most is from normal cells; the small tumour-derived fraction (ctDNA) is what liquid biopsies fish out.

Fragmentomics reads the sizes and positions of DNA fragments in blood, not the mutations. Cancer cells die messily and leave a recognisable fragmentation pattern.

Circulating tumour DNA (ctDNA) consists of fragments of DNA shed by tumour cells into the blood, detectable with sensitive sequencing.

How much of the DNA floating in a blood sample came from the cancer. Above a few per cent, a plasma test finds what a biopsy would find; below it, a negative result means the test could not see, not that there is nothing there.

Sequencing hundreds of cancer genes at once from a biopsy to find the mutations a drug can target.

The staging CT scan every cancer patient already has can be measured for muscle and fat at the level of the third lumbar vertebra; low muscle predicts chemotherapy toxicity and shorter survival across cancers, and software now does the measuring automatically.

A ctDNA MRD test looks for the tumour's own mutations in blood after surgery. Detection predicts relapse months before scans, and in 2026 the FDA approved Signatera as the companion test selecting bladder cancer patients for adjuvant atezolizumab.

Blood carries fragments of tumour DNA. This roadmap follows the tests that read them, from the first sighting in 1948 to blood tests that now choose a drug, spare chemotherapy, or screen for many cancers at once, and it lists the readouts to watch next.

A French trial of 470 men compared checking the prostate's position with imaging before every radiotherapy session against checking weekly. Daily imaging lowered rectal side effects and PSA-defined progression, but the main endpoint, recurrence-free survival, did not differ, and overall survival was worse in the daily group.

Taking and holding a deep breath during each radiation beam. For left-sided breast cancer it pulls the heart away from the chest wall and cuts heart dose by about half.

Taking and holding a deep breath during each beam moves the heart away from the treated breast and roughly halves the dose it receives. England's national guideline has recommended it for left-sided breast cancer since 2018, yet surveys in Germany, Austria, Switzerland, Australia and the United States found it applied to a fraction of the people who could have it.

Brain metastases treated with radiosurgery were randomly given a 1 millimetre or a 3 millimetre margin. Control of the treated tumours was similar, the larger margin irradiated more healthy brain, and five of the six biopsy-proven cases of radiation damage were in the 3 millimetre group.

Tumour DNA could be found in the blood of more than three quarters of people with many advanced cancers, but in fewer than half with brain, kidney, prostate or thyroid cancers, and less often when the cancer was localised.

Instead of one uniform dose, the plan gives more to the parts of the tumour that imaging says are most resistant, such as hypoxic or highly active regions on PET.

A one-off blood test done before fluorouracil or capecitabine that finds the roughly one in twenty-five people who break the drug down too slowly. Lowering their starting dose reduces severe and occasionally fatal toxicity, and in the one matched comparison published so far survival was not shortened overall.

Europe's oncology society advised that blood tests for tumour DNA can guide targeted treatment in advanced cancer, that tissue is still preferred for many patients, and that residual disease tests after surgery were not yet ready for routine use.

Counting every drug approved for a gene change up to January 2018, about 8 in 100 Americans with advanced cancer were eligible for one and about 5 in 100 were estimated to benefit.

Tiny gold seeds or electromagnetic beacons placed in or near a tumour so the treatment machine can see exactly where it is each day.

Delivering an entire dose in under a second, which in animals spares healthy tissue while still killing the tumour.

Machines built or modified to deliver a whole radiation dose in a fraction of a second, so scientists can test whether ultra-fast dosing spares healthy tissue in people the way it does in animals.

Injecting a dye that makes tumour glow so the surgeon can see exactly where to cut.

Fluorouracil is dosed by body size, yet the amount of drug in the blood varies widely between people given the same dose. Measuring the level during the infusion and adjusting the next dose raised the response rate from 18 to 34 per cent in one randomised trial and is recommended by an international pharmacology society, but it is not routine care and no survival gain has been established.

A trial of 571 men with intermediate or high-risk prostate cancer gave half of them an extra dose, up to 95 Gy, to the tumour visible on MRI inside the prostate. After five years 92 per cent of the boosted group were free of PSA relapse against 85 per cent, with no measurable increase in side effects and no difference in survival.

Chemotherapy doses used to be capped for larger people for fear of overdose. The American Society of Clinical Oncology guideline found that up to 40 per cent of people with obesity were given reduced doses, that the feared extra toxicity is unfounded, and recommends full doses based on actual body weight.

Reading the whole genome of DNA in the blood, rather than a few genes very deeply, let researchers detect tumour DNA at one part in 100,000 by adding up thousands of mutations.

Reading the DNA (and sometimes RNA) of a tumour, usually with a panel of a few hundred genes, to list its mutations, fusions and amplifications and match them to approved drugs or trials. Guidelines now require it before first-line treatment in lung, colorectal, breast and prostate cancer, though in several cancer types most tumours still show no actionable finding.

A short structured check of memory, mobility, nutrition and medicines before treatment cuts serious side effects in older patients without reducing benefit. It should be automatic, not optional.

The nested outlines a radiation plan is built on: the visible tumour (GTV), the tumour plus the tissue likely to hold microscopic spread (CTV), and a further margin for daily movement and setup error (PTV).

Tumours with a little HER2 (IHC 1+ or 2+ without amplification), or a trace (ultralow), which older HER2 drugs ignored but Enhertu can attack.

Pooling more than 10,000 tumours, a high mutation count predicted response to immunotherapy in melanoma, lung and bladder cancers but not in breast, prostate or brain tumours.

Sequencing blood plasma and white blood cells side by side showed that most mutations found in plasma came from ageing blood-forming cells, not from the cancer.

IMRT and IGRT shape the radiation beam to the tumour's outline from multiple angles and check the patient's position with a scan before every session, so surrounding organs receive less dose. Fewer, larger doses are now standard in breast and prostate cancer, but a low-dose bath still spreads across normal tissue.

Radiation beams shaped and varied in intensity from multiple angles, or while the machine rotates around the patient (VMAT), so the dose hugs the tumour and misses healthy tissue, checked against imaging before each session so it lands where planned. It cut dry mouth in head and neck cancer and rectal toxicity in prostate cancer.

Ways of checking, before the operation ends, whether cancer reaches the edge of the breast tissue removed, so the surgeon can take more there and then instead of operating again days later. In randomised trials a radiofrequency probe cut second operations from 25.8 to 19.8 per cent, and routinely shaving an extra layer from the cavity cut them from 21 to 10 per cent.

During breath-hold radiotherapy a camera or sensor measures how deep the breath is, but often only the radiographers see the trace. Showing it to the patient, on a screen or in video glasses with a target band to stay inside, made holds steadier and more repeatable in the clinical comparisons OnCo found, though none has yet measured whether it changes side effects.

A blood test that reads fragments of DNA shed by the tumour, so you can genotype or monitor cancer without a needle in the tumour.

A yearly low-radiation CT scan for current and former heavy smokers that finds lung cancer early enough to cure it.

The scoring systems that turn a screening CT into an action. Each nodule gets a category from the scan, and the category dictates the next step: rescan in a year, rescan in three months, or go to a lung cancer clinic. They exist so that the same scan gets the same answer in every centre.

After conserving surgery a pathologist measures how close the cancer came to the edge of the tissue removed. Cancer at the inked edge means another operation is offered; within 1 mm it is considered; and the threshold was lowered in 2024 specifically to spare people a second operation.

Cancer still present after treatment but too small to see on scans, detected by blood or marrow tests.

An MR-linac is a radiation machine with an MRI scanner built in, so images taken during setup let the plan be re-optimised in minutes to that day's anatomy. It allows tighter margins and higher doses in pancreatic and prostate cancer, but treatment is slow and costly, and whether daily adaptation improves cure rates rather than only toxicity is unproven.

An ultra-sensitive blood test after surgery that detects leftover cancer months before a scan would.

Tests that show whether a tumour has lost its DNA spell-checker; if so, immunotherapy works unusually well and an inherited syndrome may be present.

A single blood test intended to screen for dozens of cancers at once, including ones with no screening today.

NCI-MATCH was the largest trial to sequence tumours from thousands of patients across the United States and give each a drug matched to the mutation rather than to the organ. It proved the system could work at national scale, found real activity for a handful of matches, and showed that most single drugs given for a single mutation do little.

Sequencing the unique genetic barcode of a patient's leukaemia to find one cancer cell in a million.

People having radiotherapy for head and neck cancer were randomly given the standard closed face mask or one of two open masks, with cameras tracking the face during treatment. The open masks were set up as accurately and were significantly better tolerated.