Glossary

The open question of whether a second ADC works after the first one fails, especially when both carry the same type of payload.

When a released payload leaks out of the targeted cell and kills its neighbours, including cells that lack the target.

The drug-to-antibody ratio (DAR) is how many payload molecules ride on each antibody, typically 2 to 8.

The linker is the chemical tether between antibody and payload. It must hold in the blood and let go inside the tumour.

The payload is the poison an ADC carries, usually a chemotherapy far too toxic to give on its own.

The hydrazone linker is acid-sensitive and breaks in the lysosome's low pH. It was used in the calicheamicin ADCs, but is too leaky for most modern designs.

Calicheamicin is a DNA-cutting natural product carried by Mylotarg and Besponsa. It is powerful, but liver toxicity limits it.

CL2A is the deliberately fragile linker in Trodelvy: it releases SN-38 in the acidic tumour environment as well as inside cells, feeding the bystander effect.

DM1 is the maytansine-derived payload in Kadcyla: a tubulin blocker held by a non-cleavable linker, so it stays in the cell it entered.

DM4 is a maytansinoid payload released in a form that can cross membranes, giving mirvetuximab soravtansine a bystander effect that DM1 lacks.

DNA alkylator payloads chemically damage DNA regardless of whether the cell is dividing, and are aimed at slow-growing and resistant tumours.

DNA cleaver payloads are natural products that cut both strands of DNA. They are used in the two oldest approved ADCs, both for leukaemias.

PBD dimers are picomolar DNA crosslinkers: enormously potent, so ADCs carry very little, but the narrow safety margin has sunk most programmes.

Duocarmycins are a DNA-alkylating payload family that damages DNA directly rather than through cell division, so they work in slow-growing tumours.

DXd is the topoisomerase poison inside Enhertu and Dato-DXd: about ten times stronger than SN-38, it spreads to neighbouring cells and leaves the body quickly once released.

Exatecan is a very potent camptothecin that is the parent of most next-generation ADC payloads, including those in iza-bren and several Chinese ADCs.

Hydrophilic next-generation linkers have built-in sugars or polyethylene glycol that let ADCs carry more payload without clumping, and they resist efflux pumps.

The non-cleavable maleimidocaproyl (mc) linker is a tether with no cleavage site: the payload is released only when the antibody is fully digested.

The mc-Val-Cit-PABC linker is the workhorse of the vedotin ADCs: a valine-citrulline dipeptide cut by cathepsin B, releasing MMAE with a bystander effect.

MMAE is the most widely used tubulin-blocking ADC payload: it stops cells dividing and leaks into neighbouring cells. Nerve damage is its signature side effect.

MMAF is a charged cousin of MMAE that cannot cross membranes, so it kills only the targeted cell and spares neighbours. Eye side effects are characteristic.

The PBD dimer SG3199 is an extremely potent DNA crosslinker that works at very low doses but has been hard to make tolerable; loncastuximab tesirine is the approved example.

SMCC is the non-cleavable thioether linker in Kadcyla: the payload is freed only when the whole antibody is digested, so nothing leaks to neighbours.

SN-38 is the active form of the chemotherapy irinotecan, carried by sacituzumab govitecan. It is only moderately potent, so the ADC carries a lot of it.

Sulfo-SPDB is a hindered disulfide linker that is stable in blood but is cut by the reducing environment inside cells, then converted to a membrane-crossing form.

T030 is the belotecan-derived topoisomerase payload in sacituzumab tirumotecan, designed to be less affected by the pumps that eject SN-38.

The GGFG tetrapeptide is the four-amino-acid tether in Enhertu and Dato-DXd, cut by enzymes inside the cancer cell and very stable in blood.

Topoisomerase-I inhibitor payloads are the payload class behind Enhertu, Trodelvy and Dato-DXd: poisons that trap the enzyme that untwists DNA, so dividing cells break their own DNA.

Tubulin inhibitor payloads were the first generation of modern ADC payloads: drugs that jam the cell's scaffolding so it cannot divide. Nerve and eye side effects are typical.

The Val-Ala dipeptide linker is easier to manufacture and less prone to aggregation than valine-citrulline, and is used with PBD payloads.

A change in the lining of the lower oesophagus caused by acid reflux that can, in a minority, progress through dysplasia to adenocarcinoma. It is why Western oesophageal cancer is mostly adenocarcinoma.

The tight seal around brain blood vessels that keeps most drugs out, one of the two main reasons brain cancer is so hard to treat.

The dense scar-like tissue that makes up most of a pancreatic tumour, walling off cancer cells from drugs and immune cells.

A marker of stem cells in the gut and of stem-like cells in tumours, used to aim drugs at the cells that regrow a cancer.

When a cancer depends so completely on one mutated gene that blocking it collapses the tumour.

The Reed-Sternberg cell is the giant, often two-nucleus cancer cell of Hodgkin lymphoma; it makes up only about 1% of the tumour, and the rest is immune cells it has recruited.

Where in the colon a tumour starts changes its biology and which drugs work. Left-sided tumours respond to EGFR antibodies; right-sided ones do not.

Two genes where losing either alone is fine but losing both kills the cell. Cancers that have lost one become vulnerable to drugs against the other.

A carcinoma arising from gland-forming cells, the kind that make mucus, milk, digestive juices or hormones. The most common type of breast, lung, colon, prostate, pancreas and stomach cancer.

An agonist switches a receptor on, imitating the natural signal; an antagonist sits in the receptor and blocks it without switching it on. Cancer medicine uses both.

When a cell has many extra copies of a gene instead of the normal two, so it makes far too much of that protein. HER2-positive breast cancer is the classic example.

The growth of new blood vessels. A tumour bigger than a pinhead needs its own blood supply and sends out signals (mainly VEGF) to recruit one; anti-angiogenic drugs cut that supply.

A Y-shaped protein the immune system makes to grab onto one specific target. Manufactured antibodies are now a cornerstone of cancer treatment, either blocking a protein or delivering a payload to it.

Anything an antibody or immune cell can recognise, typically a protein on a cell's surface. In cancer, an antigen is the flag that tells a drug or an immune cell 'this is the cell to attack'.

The cell's built-in self-destruct programme, which tidily dismantles a damaged or unwanted cell without alarming its neighbours. Cancer cells disable it, and many drugs try to switch it back on.

The immune cells that make antibodies. They matter in cancer twice over: as the source of every therapeutic antibody, and as the cells that go wrong in most lymphomas, leukaemias and myeloma.

A benign tumour grows but stays put and does not invade; a malignant tumour invades surrounding tissue and can spread. Only malignant tumours are cancer.

The soft tissue inside bones where all blood cells are made. Chemotherapy damages it, causing the low blood counts that limit how much treatment a patient can take.

Cancer of the epithelium, the lining tissue that covers surfaces and forms glands. Around 85% of cancers are carcinomas, including breast, lung, colon, prostate and skin cancers.

The smallest living unit of the body. You are made of roughly 30 trillion of them, and cancer begins when one of them starts dividing when it should not.

The ordered sequence a cell goes through to copy its DNA and split in two, with checkpoints along the way where it pauses to check for damage. Cancer cells run through the checkpoints.

One cell splitting into two identical daughters. Tissues need it for growth and repair; cancer is cell division that no longer stops.

A population of cells, usually taken from a tumour decades ago, that keeps dividing indefinitely in the laboratory. The workhorse of cancer research: cheap, fast and infinitely reproducible, but a distant cousin of a real tumour.

The oily outer skin of a cell. Most drugs and antibodies act at this boundary, either by binding to proteins sticking out of it or by slipping through it.

One of the 46 long bundles into which a cell's DNA is packed. Cancer cells often have broken, missing, or extra chromosomes.

Small proteins immune cells use to talk to each other: alarms, recruitment calls, growth orders and stand-down signals. Some are cancer drugs, and a flood of them is the danger of certain immunotherapies.

When a piece of DNA is missing, from a few letters to an entire gene. Deleting a tumour suppressor gene removes one of the cell's brakes.

How closely a cancer cell still resembles the mature, specialised tissue it came from. Well-differentiated tumours look like their parent tissue and tend to behave better; poorly differentiated ones look primitive and grow faster.

The long molecule that stores the instructions for building and running a cell, written in a four-letter chemical alphabet. Cancer is fundamentally a disease of damaged DNA.

Position in a signalling chain: upstream proteins send the message, downstream proteins receive it. A mutation downstream of a drug's target can make the drug useless.

One of the few mutations in a tumour that actually causes it to grow. Everything else is a passenger along for the ride. Drivers are the mutations drugs are aimed at.

A protein that speeds up a specific chemical reaction. Many drug targets are enzymes, because blocking the enzyme stops the reaction.

The exact small patch on an antigen that an antibody or T cell actually grips, like the specific spot on a door handle a hand grabs. Two antibodies can bind the same protein at different epitopes.

A stretch of DNA that holds the recipe for one protein (or one working RNA). Humans have about 20,000, and cancer typically involves a handful of them going wrong.

How much a gene is switched on, and therefore how much of its protein a cell makes. Cancer cells often express the wrong genes at the wrong levels.

A protein released by one cell that tells nearby cells to grow, divide or survive. Cancers often make their own or become hypersensitive to it.

The instruction a cell receives to grow and divide. Normal cells wait for it; cancer cells fake it, so they keep dividing whether or not the body wants more of them.

Brakes on the immune system that stop T cells attacking healthy tissue. Tumours pull these brakes to protect themselves; checkpoint inhibitor drugs release them so T cells can attack the cancer.

The body's defence network of cells and molecules that recognises and destroys infected, foreign and abnormal cells. It kills most would-be cancers before they are ever noticed; the ones that survive have learned to hide from it.

In vitro means 'in glass': experiments on cells or molecules in a dish. In vivo means 'in the living': experiments in animals or people. A drug that kills cancer cells in vitro has cleared only the first and easiest hurdle.

The body's response to injury or infection: immune cells and fluid rush in, causing redness, heat and swelling. Short bursts heal; long-term smouldering inflammation promotes cancer.

A drug that blocks a specific protein from doing its job, usually by lodging in the pocket the protein needs to work. Most targeted cancer drugs are named after what they inhibit.

An enzyme that switches other proteins on by sticking a phosphate group onto them. Cancer signalling runs on kinases, and kinase inhibitors are the biggest class of targeted drugs.

Cancer of the blood-forming cells in the bone marrow, which flood the blood with immature or abnormal white cells and crowd out normal blood production. It has no single lump to remove.

The molecule that fits into a receptor and switches it on, the key to the receptor's lock. Growth factors, hormones and cytokines are all ligands.

Small bean-shaped filters along the lymph vessels where immune cells gather. They are often the first place a cancer spreads, so surgeons check them to see how far the disease has gone.

Cancer of lymphocytes, the white blood cells of the immune system, usually growing as masses in lymph nodes, spleen or other organs. Divided into Hodgkin and the many non-Hodgkin types.

Large immune cells that engulf and digest debris, microbes and dead cells and coordinate inflammation. Tumours recruit and retrain them, so in many cancers macrophages end up helping the tumour rather than fighting it.

Cancer that has spread from where it started to distant parts of the body, travelling through the blood or lymph. Metastasis is what makes most cancers deadly.

A species studied in the lab to understand biology that is shared with humans: mice above all, plus zebrafish, flies, worms and yeast. Mice carrying human tumours are the standard test bed for cancer drugs, and a notoriously imperfect one.

Made from a single clone of cells, so every antibody molecule in the vial is identical and binds exactly the same spot. All therapeutic antibodies are monoclonal.

A change in the DNA sequence, like a typo in a recipe. Most are harmless; a few change a protein in a way that helps a cell grow out of control.

Messy, uncontrolled cell death, where the cell bursts and spills its contents. In tumours it usually means cells have outgrown their blood supply and starved.

Natural killer cells: fast-acting immune cells that destroy abnormal cells without needing to be trained on a specific target. They specialise in killing cells that have hidden themselves from T cells.

The compartment at the centre of a cell that holds the DNA, like a library that keeps the master copies of every instruction the cell might need.

A gene that, when over-activated by mutation or extra copies, pushes a cell to grow and divide. Think of an accelerator pedal stuck to the floor.

A tiny three-dimensional version of a tissue or tumour grown in the lab from a patient's own cells. Tumour organoids keep more of the original cancer's character than flat cell lines and can be used to test drugs on a patient's own cancer.

When a cell makes much more of a particular protein than normal cells do. Drugs that home in on that protein can then hit the tumour harder than healthy tissue.

Attaching a small phosphate tag to a protein, which changes its shape and switches it on or off. It is the cell's main way of passing signals from one protein to the next.

Everything done to a potential drug before it is given to a human: lab experiments, animal testing, and the safety and manufacturing work regulators require. Most candidates never make it out.

The original tumour where a cancer started. Cancer is named after this site for life: a breast cancer that spreads to the liver is still breast cancer, not liver cancer.

A drug given in an inactive form that the body converts into the active medicine. It can make a drug easier to take, longer-lasting, or active only where it is needed.

Rapid multiplication of cells. A tumour's proliferation rate, often measured by the marker Ki-67, tells you how fast it is growing and how it may respond to treatment.

The molecular machines that do almost everything in a cell: receive signals, copy DNA, build structures, digest food. Most cancer drugs work by binding to one specific protein.

A protein, usually on the cell surface, that catches a specific signal molecule from outside and relays the message inside, like a doorbell wired to the cell's controls.

A working copy of a gene, made when the cell needs to use it. If DNA is the master library, RNA is the photocopy taken to the workshop.

Cancer of the body's connective and supporting tissues: bone, muscle, fat, cartilage, blood vessels, fibrous tissue. Rare (about 1% of adult cancers) but relatively common in children.

A chain of proteins that passes a message from the cell surface to the nucleus, each one switching on the next like a relay race. Cancer drugs try to break a link in the chain.

A carcinoma arising from the flat, layered cells that line surfaces exposed to wear: skin, mouth, throat, oesophagus, cervix, the larger airways. Often linked to tobacco, sun or HPV.

A cell that can both copy itself indefinitely and produce the specialised cells of a tissue. Cancers may be sustained by a small population of stem-like cells that survive treatment and regrow the tumour.

The immune system's trained assassins. Each T cell recognises one specific target displayed on a cell's surface and, if the cell is infected or cancerous, kills it. Most immunotherapy is about unleashing them.

A gene whose normal job is to stop cells dividing or to make damaged cells die. Losing it removes a brake, so the cell can grow unchecked even without a stuck accelerator.

Fragments of Epstein-Barr virus DNA in the blood that measure nasopharyngeal carcinoma: used to screen healthy people in endemic regions, to stage, to decide who needs extra treatment after radiotherapy, and to detect relapse.

A protein made by fetal liver that many liver cancers switch back on; measured in blood for screening, prognosis and to select patients for ramucirumab.

A sugar molecule shed into the blood by most pancreatic cancers; useful to follow treatment, not to screen.

A blood protein that rises in most ovarian cancers; useful for tracking treatment, useless for screening on its own.

Whether a large B-cell lymphoma resembles a germinal-centre B cell or an activated B cell; the activated type does worse and depends on different pathways.

Chromogranin A is a protein released by neuroendocrine cells and measured in blood to follow tumour burden; it is unreliable because acid-reducing drugs and kidney disease also raise it.

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

A PD-L1 score that counts stained tumour cells and immune cells together.

A del(17p) deletion or TP53 mutation means loss or damage of the p53 safety gene in CLL. These patients should never get chemotherapy; they need BTK inhibitors or venetoclax, usually continuously.

Double-hit lymphoma is a large B-cell lymphoma with rearrangements of two oncogenes (MYC plus BCL2 and/or BCL6), which behaves aggressively and often escapes R-CHOP.

EGFRvIII is a mutant, tumour-only version of the EGFR receptor found in about a third of glioblastomas. It is an ideal-looking target, yet every drug against it has failed so far.

A change in the oestrogen receptor gene that lets the cancer grow without oestrogen, so aromatase inhibitors stop working. Found in the blood in about a third of patients after hormone therapy.

FGFR3 is a growth-receptor gene mutated or fused in about a fifth of advanced bladder cancers and most low-grade early ones. It is the only targetable biomarker in bladder cancer so far.

A single change in a histone protein that defines the most lethal childhood brain tumour, and now the target of the first approved drug for it.

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

Chromosome changes such as del(17p), t(4;14), t(14;16) and extra copies of 1q that mark myeloma likely to relapse early.

A tumour that cannot properly repair double-strand DNA breaks, usually because of BRCA or related gene loss.

Throat cancers caused by the human papillomavirus, identified by a p16 stain. They affect younger non-smokers and are far more curable than tobacco-related cancers.

Whether the leukaemia's antibody gene has been 'edited' by the immune system. Unmutated means faster-growing CLL and a bigger benefit from targeted drugs over chemotherapy.

A chemical switch that turns off a DNA-repair gene. When it is off, temozolomide works much better.

Microsatellite instability is the mark of a broken DNA spell-checker that creates thousands of mutations, making the tumour highly visible to immunotherapy.

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

Extra copies of the MYCN oncogene, found in about 20% of neuroblastomas, mark the most aggressive disease and define high risk at any age.

A blood protein made by the prostate; raised levels prompt further tests, and falling levels show treatment is working.

How many mutations a tumour has. More mutations mean more targets for the immune system.

The PD-L1 score used in lung cancer: the percentage of tumour cells that stain positive.

Immune cells that have got inside the tumour. More of them means better outcomes in triple-negative breast cancer.

The scaffolding cells that tumours recruit to build scar-like tissue around themselves. They feed the cancer, block drugs and immune cells, and carry the FAP protein that PET scans can now see.

Abnormal cervical cells caused by HPV that can turn into cancer over 10-20 years. Screening finds them; a five-minute procedure removes them.

Single cancer cells that have already spread to distant organs, often found in bone marrow years before any metastasis appears.

Cancers mutate faster than normal cells because their DNA repair and chromosome segregation are broken. This fuels every other hallmark.

Chronic inflammation supplies growth factors, survival signals, and mutagens that help cancers start and grow.

Four groups defined by a few tests that predict outcome better than the microscope: POLE-mutated (excellent), mismatch-repair deficient, p53-abnormal (worst), and 'no specific profile'.

Cancers can change behaviour, including becoming drug-tolerant, without any new mutation, by rewriting the chemical tags that control which genes are read.

The bacteria living in and around us differ from person to person and influence cancer risk, progression, and how well treatments work.

Cells that have permanently stopped dividing but stay alive and secrete inflammatory signals, helping nearby cancer cells grow and resist treatment.

Cancer cells escape the normal rule that mature cells stay what they are: they dedifferentiate, transdifferentiate, or refuse to mature.

Cancer cells detach, invade, travel, and colonise other organs, the process responsible for most cancer deaths.

Cancers hide from or switch off the immune system, by losing the molecules that display their antigens or by raising checkpoint brakes.

Cancer cells bypass the division limit set by shortening telomeres, usually by reactivating telomerase.

Evading growth suppressors means cutting the brakes that normally stop division, p53 and RB above all.

Tumours grow their own blood supply, or co-opt existing vessels, to get oxygen and nutrients.

Cancer cells rewire their metabolism to build biomass fast, burning glucose inefficiently and gorging on glutamine and lipids.

Resisting cell death is the hallmark by which cancer cells disable their self-destruct programmes, chiefly apoptosis.

Cancer cells keep telling themselves to divide, by making their own growth signals or jamming the receptors on.

The Hallmarks of Cancer is the most-cited framework in cancer biology: a short list of the capabilities every cancer must acquire, from unlimited growth to hiding from the immune system. The 2022 update lists eight hallmarks, two enabling characteristics, and four new dimensions.

High-grade serous ovarian carcinoma is the most common and most lethal type of ovarian cancer; it actually starts in the fallopian tube and almost always has a broken TP53 gene.

A tumour where the immune cells arrive but are stopped at the edge, held back by scar tissue and signals such as TGF-β.

Low-grade serous ovarian cancer is a slow-growing, RAS-driven type of ovarian cancer that resists chemotherapy but responds to hormone blockers and MEK-pathway drugs.

Immature immune cells that tumours summon to switch off T cells. Their numbers in blood predict worse immunotherapy outcomes.

Tumour-associated macrophages are immune cells that should eat cancer cells but are re-educated by the tumour to protect it instead. They are often the most abundant immune cell in a tumour.

Cancer cells burn glucose into lactate even when oxygen is plentiful, an inefficient but fast way to make building blocks. Otto Warburg described it in the 1920s.

Whole-genome doubling is a single event in which a cancer cell duplicates its entire genome, becoming tetraploid. It happens in about a third of cancers and buffers the chaos that follows.

Anything measurable in the body or tumour that tells you something useful: what kind of cancer it is, how it is likely to behave, or which drug is likely to work.

Taking a small piece of the suspicious tissue so a pathologist can examine it under the microscope. It is the only way to confirm a cancer diagnosis and to learn its type and markers.

Keeping patients (single-blind) or patients and doctors (double-blind) from knowing who is on which treatment, so that hopes and expectations cannot colour the results.

Drugs that kill rapidly dividing cells by damaging their DNA or the machinery of division. Because cancer cells divide fast they are hit hardest, but so are hair, gut lining and bone marrow, which is where the side effects come from.

A research study that tests a treatment in volunteers under strict rules, to find out whether it is safe and whether it works. Every approved cancer drug went through several.

All detectable signs of the tumour have vanished on scans and examination after treatment. The best result a trial can record for an individual patient, though microscopic disease may remain.

The range of values consistent with the trial's data, usually given at 95%. A hazard ratio of 0.70 with an interval of 0.55 to 0.89 means the true effect is probably somewhere in that range; if the range crossed 1.0 the result would not be significant.

How much of a drug is given, how often, and for how long. Cancer drugs have long been given at the highest dose a patient can stand, and regulators are now pushing for doses chosen for benefit rather than tolerability.

How long a tumour stays shrunk once it has responded to a drug. A response that lasts two years is worth far more than one that lasts two months, so this is reported alongside response rate.

Finding a cancer while it is still small and confined, when it is most often curable. Five-year survival for most cancers is several times higher at stage I than at stage IV, which is why so much effort goes into detecting them sooner.

The specific outcome a trial is designed to measure, fixed in advance: for example how long patients live, or how long before their cancer grows. A trial 'meets its endpoint' when the new treatment beats the comparison on that measure.

Reading the DNA (and sometimes RNA) of a tumour to list the mutations it carries and match them to drugs. Now routine in lung, colorectal, breast, prostate and many other cancers.

How abnormal the cancer cells look under the microscope, from grade 1 (close to normal, slow) to grade 3 or 4 (wildly abnormal, fast). Grade is about behaviour; stage is about extent.

The time it takes for the amount of a drug in the blood to fall by half. It determines how often a drug must be given: hours for many chemotherapies, weeks for antibodies.

The study of tissue under the microscope. A cancer's histology is what kind of cells it is made of and how they are arranged, which is how a pathologist names it and grades it.

Treatment that starves cancers which grow in response to a hormone, mainly oestrogen in breast cancer and testosterone in prostate cancer, by lowering the hormone or blocking its receptor.

Any treatment that works by getting the patient's own immune system to attack the cancer, rather than attacking the cancer directly. It can produce responses that last for years, but only in some patients.

Latin for 'in place': abnormal cells that look like cancer but have not yet broken through the layer they started in. Stage 0. Not yet able to spread, and usually curable by removing it.

Incidence is how many new cases occur in a year; prevalence is how many people are living with the disease at a given moment. A curable or fast-killing cancer can have high incidence but low prevalence.

Cancer that has grown through the boundary of the tissue it started in and into the surrounding tissue. This is the step that separates true cancer from in situ disease and gives it access to blood and lymph vessels.

A doctor's neutral word for any abnormal patch of tissue seen on a scan or examination: a lump, a nodule, a shadow. It may or may not be cancer until a biopsy says so.

Ongoing, gentler treatment given after the main course has shrunk the cancer, to hold it in check for as long as possible rather than to shrink it further.

The time by which half the patients in a group have died (or, for other endpoints, progressed). It is the midpoint of a spread, not a prediction for anyone: half live longer, some much longer.

The number of deaths from a disease in a population per year. Cancer mortality has fallen about a third since 1991 in the US, mostly from less smoking, earlier detection and better treatment.

Prescribing an approved drug for a use not listed on its official label, for example a different cancer or an earlier stage. Legal and common in oncology, but insurers may refuse to pay and the evidence is often thinner.

A drug for a rare disease (in the US, fewer than 200,000 patients) that gets extra incentives, tax credits, fee waivers and seven years of market exclusivity, to make development worthwhile. Most cancers qualify, so most cancer drugs are orphans.

The probability of seeing a difference at least this large if the treatment actually did nothing. Below 0.05 (a 1 in 20 chance) is the conventional threshold for calling a result 'statistically significant'. It measures surprise, not importance.

Treatment aimed at relieving symptoms and improving quality of life rather than curing the disease. Not the same as end-of-life care: palliative treatments are given alongside anticancer drugs at every stage.

The tumours have shrunk substantially (by at least 30% in total diameter) but not disappeared. Together with complete responses it makes up a trial's response rate.

The three stages a new drug passes through: phase 1 finds a safe dose in a few dozen patients, phase 2 looks for signs of activity in a hundred or so, and phase 3 compares it with the standard treatment in hundreds or thousands.

An inactive look-alike treatment given to the comparison group in a trial so that neither patients nor doctors can tell who is on the real drug. In cancer trials it is added on top of standard treatment, never given instead of it.

The expected course of a disease: how likely it is to be cured, how long a person is likely to live, and how they are likely to feel. Always an estimate based on groups of similar patients, never a prediction for one person.

The cancer is growing or spreading despite treatment, or after it. Progression usually means the current drug has stopped working and it is time to switch to the next line.

How a patient actually feels and functions day to day: symptoms, energy, mood, ability to work and live normally. Measured with questionnaires, it is the outcome that matters most alongside survival, and the one trials have historically neglected.

Using high-energy X-rays or particles to damage the DNA of cancer cells in a precisely aimed volume of the body. It cures many localised cancers and relieves symptoms in advanced ones, and about half of all cancer patients receive it.

A trial in which a coin toss (done by computer) decides which treatment each patient gets, so the groups are alike in every way except the treatment. It is the only reliable way to prove a treatment causes a benefit.

A hazard ratio of 0.70 means that at any given moment patients on the new treatment had 30% less risk of the event (death, progression) than those on the comparison. It says nothing about how many months that is worth.

The cancer has come back after a period in which it could not be detected. 'Recurrence' is the usual word for solid tumours and 'relapse' for blood cancers; they mean the same thing.

Cancer that does not respond to a treatment at all, or grows straight through it. Distinct from relapse, where the cancer responded first and came back later.

When the signs of cancer have shrunk (partial remission) or disappeared entirely (complete remission). It is not the same as cure, because cells too few to detect may remain.

Anything that raises the chance of developing a cancer: smoking, alcohol, obesity, sunlight, certain infections, inherited genes, age. Having a risk factor does not mean getting cancer, and many cancers occur without any known one.

Testing people who have no symptoms to catch cancer, or its precursors, early enough to cure. Proven for breast, cervical, colorectal and (in smokers) lung cancer; every test also finds some cancers that would never have caused harm.

An adverse event is anything bad that happens to a patient during a trial, whether or not the drug caused it; a side effect is a harm the drug is known to cause. Trial tables list adverse events and then judge which were treatment-related.

The tumours have neither shrunk enough to count as a response nor grown enough to count as progression. For a slow cancer or a well-tolerated drug, months of stability can be a real benefit.

How far a cancer has spread, from stage I (small and confined) to stage IV (spread to distant organs). Stage is the single strongest guide to treatment and prognosis.

A quicker, easier measurement used as a stand-in for what really matters. Tumour shrinkage or delayed growth stands in for living longer, on the assumption, not always true, that one leads to the other.

Local therapy (surgery, radiotherapy) treats one place in the body; systemic therapy (drugs given by mouth or vein) travels through the bloodstream and treats the whole body, including cancer cells too small to see.

Drugs designed to hit a specific molecule the cancer depends on, usually a protein made by a mutated or amplified gene, while leaving normal cells relatively alone. The tumour is tested first to see whether it carries the target.

A 1 to 5 severity scale for side effects: grade 1 mild, grade 2 moderate, grade 3 severe and needing intervention, grade 4 life-threatening, grade 5 fatal. Trials report the percentage of patients with grade 3 or worse.

A substance, usually a protein, that a tumour releases into the blood in measurable amounts. Rising or falling levels track whether the cancer is growing or responding, though they are rarely good enough for diagnosis on their own.

About 1 in 1,000 pregnancies is complicated by cancer, most often breast, cervical, lymphoma, melanoma or leukaemia. Most chemotherapy is safe after the first trimester, surgery is safe throughout, and ending the pregnancy does not improve the mother's outcome.

The harm caused to patients by the cost of cancer care: depleted savings, debt, skipped medication and worse survival. Measured like any other side effect and, increasingly, treated like one.

A US law that lets certain hospitals buy outpatient drugs, including cancer drugs, at deep discounts. Controversial because hospitals may bill insurers full price.

The FDA's definition of early bladder cancer that has failed adequate BCG treatment, the population in which most new bladder drugs are first approved.

The liver-cancer staging system that combines tumour size, liver function and fitness to recommend treatment: ablation or surgery, transplant, TACE, or drugs.

PSA rising again after surgery or radiation, usually years before anything shows on a scan.

Flushing, diarrhoea and wheezing caused by hormones (mostly serotonin) released by some neuroendocrine tumours; over years it can scar the heart valves.

Prostate cancer that keeps growing even though testosterone has been reduced to castrate levels.

Scores for how well the liver still works; in liver cancer they decide whether a patient can tolerate treatment at all.

The three-tier system (favourable, intermediate, adverse) that decides how aggressively an adult with AML is treated and whether a transplant is recommended.

The pathologist's 1-to-5 grade of how abnormal prostate cancer looks, which drives most treatment decisions.

Up to half of women with metastatic HER2-positive breast cancer develop brain metastases, because antibodies control the body but historically not the brain.

Histotype-tailored therapy means choosing treatment by the specific sarcoma subtype (there are more than 70) rather than treating all sarcomas alike.

The IMDC score uses six factors to sort metastatic kidney cancer into three risk groups. It decides whether dual immunotherapy or immunotherapy plus a targeted pill is offered first.

INRG staging is the international system that sorts neuroblastoma into very-low, low, intermediate and high risk using age, spread, MYCN status and tumour biology.

A five-point score (age, stage, performance status, LDH, extranodal sites) that predicts how risky a lymphoma is before treatment.

Bile duct cancers are named by where they start: inside the liver, at the hilum where the ducts join, in the lower duct near the pancreas, or in the gallbladder. Each behaves and mutates differently.

Hormone-driven breast cancer can come back 10 or even 20 years after treatment, unlike most cancers, which is why hormone therapy lasts so long.

Small-cell lung cancer is split into disease that fits in one radiation field (limited) and disease that has spread beyond it (extensive). The first is treated to cure, the second to control.

First-line is the first treatment for advanced cancer; second-line is what comes after it fails, and so on.

Small thyroid cancers confined to the gland with no spread. They are almost always cured by surgery alone and no longer need radioactive iodine.

The Lugano classification is the lymphoma staging system: stage I to IV by how many lymph node regions and organs are involved, with PET-based response criteria.

Neoadjuvant, adjuvant and perioperative therapy mean treatment given before surgery, after surgery, or both, respectively.

Bladder cancer is divided by whether it has grown into the bladder's muscle wall. Before that, it is treated inside the bladder; after, the bladder is usually removed or irradiated.

Cancer that has spread to only a few places, which may still be curable by treating each spot.

When only one or two spots grow on an otherwise working targeted therapy; treat the spots and keep the pill.

Temporarily switching off the ovaries with injections (or removing them) so a premenopausal woman's cancer is starved of oestrogen.

Spread across the lining of the abdomen, the most common way stomach cancer recurs and the hardest to treat.

Ph-positive ALL is acute lymphoblastic leukaemia carrying the Philadelphia chromosome. It was once the worst kind and is now often the best-controlled, thanks to TKIs and blinatumomab.

R-ISS is the myeloma staging system, combining blood markers with high-risk chromosome changes to predict outcome.

Thyroid cancer that no longer takes up radioactive iodine, or keeps growing despite it. This is when kinase inhibitor pills come in.

When slow CLL suddenly turns into an aggressive lymphoma. Rare, hard to treat, and the focus of new immunotherapy trials.

A way of classifying cancers at the junction of the oesophagus and stomach by where their centre sits, which decides whether they are treated as oesophageal or gastric.

Early plasma-cell conditions with no organ damage; most never progress, but high-risk smouldering disease is now sometimes treated.

The standard of care is the treatment that guidelines and experts currently consider the best-proven option.

TNM staging is the universal system describing tumour size (T), lymph node spread (N), and distant metastasis (M).

Giving slightly more thyroid hormone than the body needs after thyroid cancer surgery, to switch off the pituitary signal that could feed leftover cancer cells.

The Bethesda category is a six-step scale, from 'not enough cells' to 'cancer', that pathologists use to report a thyroid needle biopsy.

A stool test for hidden blood, done at home every year. Positive results are followed by colonoscopy.

Staining a tissue slice with antibodies so a protein shows up in colour under the microscope.

A standard checklist of cancer data every electronic record should capture the same way, so data can flow between hospitals, registries, and research.

A shared format for patient records that lets hospitals around the world run the same study on their data without sharing it.

Positive predictive value (PPV) is the chance that a test saying 'cancer' is right.

Stage shift means finding more cancers early and fewer late; it is the measurable goal of screening.

No sign of tumour on examination, endoscopy, and MRI after treatment, without surgery to confirm it. It is the basis of organ-preservation strategies.

In early-stage cancer: how long patients stay free of recurrence, progression, or death.

A hazard ratio is a number comparing the rate of bad events in two groups. An HR of 0.5 means the risk is halved at any moment.

When, after pre-surgery treatment, the removed tumour contains little or no living cancer: 10% or less viable cells.

MRD negativity means no detectable myeloma cell among 100,000 or a million marrow cells. It is the best predictor of long survival and, since 2024, an accepted endpoint for accelerated approval.

MRD-negative complete remission means not just no leukaemia visible under the microscope, but none detectable with tests a thousand times more sensitive. It is the goal of modern leukaemia treatment.

Objective response rate (ORR) is the percentage of patients whose tumours shrink by at least 30%.

Overall survival (OS) is how long patients live, full stop. It is the gold-standard endpoint.

No invasive cancer left in the breast and lymph nodes when the surgeon removes the tissue after pre-surgery treatment.

Progression-free survival (PFS) is how long patients live without their cancer growing.

PSA50 is the share of patients whose PSA falls by at least half on treatment, and PSA90 the share whose PSA falls by 90%.

RECIST is the rulebook for measuring whether tumours have grown or shrunk on scans.

A score for how much cancer is left after pre-surgery treatment, from 0 (none) to III (a lot).

Cancer in people aged 15-39, about 90,000 US cases a year, with a distinct mix of cancers, slower survival improvement than children or older adults, and specific needs: fertility, education and work, psychosocial support and trial access.

Systematic differences in who gets cancer, how early it is found and who survives, driven by race, income, geography, insurance and structural racism rather than biology alone.

Rare cancers are those with fewer than about 6 new cases per 100,000 people per year. Individually rare, together they are a quarter of all cancers and have worse survival because of late diagnosis, few trials and scattered expertise.

About 5-10% of cancers arise from an inherited gene fault. Recognising the syndromes (BRCA, Lynch, Li-Fraumeni, VHL, MEN, FAP, retinoblastoma and dozens more) changes screening, surgery and treatment for the patient and their relatives.

Li-Fraumeni syndrome is an inherited fault in the TP53 gene giving a lifetime cancer risk near 100% in women and ~75% in men, with sarcomas, breast cancer, brain tumours, adrenal cancer and leukaemias often in childhood. Whole-body MRI surveillance saves lives.

Von Hippel-Lindau disease is an inherited condition causing kidney cancers, adrenal tumours, and blood-vessel tumours of the brain, spine, eye and pancreas from early adulthood. It taught us how cells sense oxygen and gave rise to the drug belzutifan, the first medicine for VHL tumours.

Lynch syndrome is the most common inherited cancer syndrome: a faulty mismatch-repair gene raises lifetime bowel cancer risk to 40-80% and also endometrial and other cancers.

The consensus molecular subtypes are four gene-expression groups of bowel cancer: immune (CMS1), canonical (CMS2), metabolic (CMS3), and mesenchymal (CMS4), with different prognoses.

Exon 19 deletion and L858R are the two common EGFR mutations, together ~85% of EGFR-mutant lung cancer, and both respond to EGFR pills.

A rarer EGFR mutation (~2% of lung cancers) that does not respond to standard EGFR pills and needs its own drugs.

A broken-and-rejoined FGFR2 gene that drives about one in eight intrahepatic bile duct cancers and can be switched off with pills.

How much of the FLT3 gene in the leukaemia carries the internal duplication. Once used to grade risk; now less important because FLT3 inhibitors help regardless.

A gene fusion is two genes broken and joined together, creating a hybrid protein that can drive cancer. Fusion-driven cancers are often exquisitely drug-sensitive.

Germline mutations are inherited and in every cell; somatic mutations arise in the tumour only.

Inherited conditions (MEN1, VHL, NF1, tuberous sclerosis) that cause neuroendocrine tumours, often multiple and at a young age, so families need genetic testing and surveillance.

A characteristic pattern of mutations that reveals what caused them: tobacco, UV, a broken repair gene.

Reading millions of DNA fragments in parallel, the engine behind every modern genomic test.

A 50-gene test that sorts breast cancers into luminal A, luminal B, HER2-enriched, and basal-like.

A mutation that keeps the cell's immortality enzyme switched on. In thyroid cancer, having it alongside BRAF marks the tumours most likely to spread and resist iodine.

The fraction of DNA reads carrying a mutation. Tells you how much of the sample is mutant and how clonal it is.

A variant of uncertain significance (VUS) is a genetic change that has been found but nobody yet knows whether it matters.

A 1-to-5 score for how bright a lymphoma looks on PET compared with the liver; 1-3 is considered a complete metabolic response.

The standardised uptake value (SUV) is a number for how brightly a spot lights up on a PET scan.

When irradiating one tumour causes untreated tumours elsewhere to shrink, via the immune system.

In ADCC, an antibody flags a cell, and NK cells recognise the flag and kill it.

Tweaking the antibody's tail to make it recruit immune cells more strongly, or not at all.

Some T-cell-receptor drugs only work in people with a particular immune 'tissue type'. About half of people of European ancestry have it; far fewer in some other populations.

'Hot' tumours are full of immune cells and respond to immunotherapy; 'cold' tumours have kept the immune system out.

Immunogenic cell death is a way of dying that alerts the immune system, unlike quiet apoptosis.

A protein fragment created by a tumour mutation that the immune system has never seen before, so it can attack it without harming normal cells.

The share of cancers caused by drinking: about 4% of all new cancers worldwide, roughly 740,000 a year, of which a meaningful fraction come from light and moderate drinking.

What a person's weight is made of: muscle, fat under the skin, fat around the organs. It predicts cancer outcomes far better than weight or BMI, and can be read from routine CT scans.

Severe loss of weight and muscle in advanced cancer that eating more cannot reverse on its own. It affects up to eight in ten patients with advanced disease and contributes to a fifth of cancer deaths.

Scores that grade a whole diet against a healthy pattern instead of counting single foods. They predict cancer risk and survival better than any individual nutrient.

The umbrella term researchers use for the combined effect of what you eat, how much you move and how much fat you carry. It is the framework linking diet, exercise and obesity research to cancer.

How fast a food raises blood sugar (index) and how much, given the portion (load). High-glycaemic-load diets are weakly linked to some cancers, largely via obesity and insulin.

How many different kinds of bacteria live in the gut and which ones dominate. Higher diversity and certain species are linked with better immunotherapy response; antibiotics and poor diet reduce both.

Immunonutrition means nutritional drinks enriched with specific nutrients (arginine, omega-3 fats, nucleotides) meant to strengthen immune function, mostly given before major surgery.

Quick questionnaires that flag who is at risk of malnutrition: recent weight loss, low BMI, poor appetite, illness severity. Anyone flagged should see a dietitian.

The unit used to measure how much exercise someone does. One MET is the energy of sitting quietly; brisk walking is about 4 METs, so 45 minutes of brisk walking is 3 MET-hours. Guidelines aim for around 10 MET-hours a week.

Metabolic syndrome is a cluster of central obesity, high blood pressure, high blood sugar and abnormal blood fats. It raises the risk of several cancers and worsens outcomes after diagnosis, largely through high insulin levels.

The side-effects of cancer and its treatment that stop people eating: nausea, mouth soreness, taste changes, difficulty swallowing, early fullness, constipation, pain and low mood. Treating them is often the most effective nutrition intervention.

Excess body fat is an established cause of at least thirteen cancers, including womb, oesophagus, kidney, liver, bowel, pancreas and postmenopausal breast cancer. It is the second largest preventable cause of cancer after smoking in many countries.

Using the weeks between diagnosis and surgery or chemotherapy to get fitter, better nourished and psychologically prepared, so treatment goes better and recovery is faster.

Sarcopenia is loss of muscle mass and strength. In cancer it predicts worse chemotherapy side-effects, more surgical complications and shorter survival, and it can hide in people who look a normal weight or overweight.

Industrially formulated products made mostly from extracted or synthesised ingredients (soft drinks, packaged snacks, reconstituted meats, many ready meals). High intake is linked with obesity and, in cohorts, with more cancer.

Diets marketed as cancer cures or preventives with no supporting evidence: alkaline diets, juice cleanses, Gerson therapy, apricot kernels and the like. Some are merely useless; several have caused harm or led people to delay effective treatment.

Because cancer cells burn a lot of glucose, many people conclude that cutting sugar starves tumours. The biology is real; the conclusion is not. Blood glucose is tightly regulated and no trial shows sugar restriction improves cancer outcomes.

How deep a melanoma has grown into the skin, in millimetres. The single strongest predictor of whether it will spread.

Mesothelioma comes in a slower 'epithelioid' form and an aggressive 'sarcomatoid' form. Chemotherapy works better in the first; immunotherapy helps most in the second.

The FNCLCC grade is a 1-to-3 score for soft-tissue sarcomas based on how abnormal, how fast-dividing, and how much dead tissue the tumour shows; grade drives whether chemotherapy is considered.

Stomach cancers come in two main shapes: intestinal (gland-forming, linked to H. pylori and HER2) and diffuse (scattered cells, linked to CDH1 loss and worse outcomes).

How fast the tumour cells are dividing, measured by Ki-67 staining, separates slow-growing neuroendocrine tumours from aggressive neuroendocrine carcinomas and decides the treatment.

A spindle-cell change found in about 10% of kidney cancers that makes them aggressive and, unexpectedly, unusually responsive to immunotherapy.

Oesophageal cancer is two different diseases in one organ: squamous cell carcinoma (upper/mid oesophagus, tobacco and alcohol, dominant in Asia) and adenocarcinoma (lower oesophagus, reflux and obesity, dominant in the West).

Loss of the skin surface over a melanoma under the microscope; a sign of aggressive biology that raises the stage.

A company that makes drugs for other companies. Most ADCs, biologics, and cell therapies are made by a handful of them.

Using two HER2 antibodies (trastuzumab and pertuzumab) at once, which works better than one.

An oral SERD is a pill that destroys the oestrogen receptor rather than just blocking it, replacing the monthly fulvestrant injection.

New cancer drugs commonly launch above $150,000-$250,000 per year in the US, with prices set without reference to benefit and hidden behind rebates; other countries negotiate or assess cost-effectiveness. Pricing policy determines who gets treated.

There are not enough oncologists, radiation therapists, pathologists, oncology nurses and pharmacists for the growing number of patients, in rich countries and especially in poor ones; burnout, geography and training pipelines shape who receives care.

The World Health Organization's list of medicines every health system should provide, which now includes about 60 cancer drugs from cyclophosphamide and cisplatin to trastuzumab, imatinib and PD-1 inhibitors, guiding procurement and pricing in lower-income countries.

Two viruses cause most liver cancer worldwide. One is preventable by vaccine, the other curable with pills, which makes liver cancer one of the most preventable cancers.

Beta particles (lutetium-177) travel millimetres and are good for bulky disease; alpha particles (actinium-225) travel a few cells' width and kill with far higher energy.

Measuring how much radiation dose each organ and tumour actually received from a radioactive drug.

Using the same targeting molecule for a diagnostic scan and a therapy, so you treat exactly what you can see.

FDA approval based on early evidence (like tumour shrinkage) on condition that a confirmatory trial follows.

Proposed US legislation restricting federally funded work with certain Chinese biotech service companies, which would affect where ADCs and sequencing get done.

A biosimilar is a copy of a biologic drug such as trastuzumab, shown to be as safe and effective as the original once its patent expires, usually at a lower price.

Breakthrough Therapy, Priority Review and the Priority Voucher are FDA fast lanes for drugs that look substantially better than existing options.

A companion diagnostic is a test the FDA requires before a particular drug can be prescribed.

Independent bodies that judge whether a new cancer drug's benefit is worth its price; decisive for coverage outside the US and increasingly inside it.

Lists of recommended off-label and on-label drug uses that US insurers, including Medicare, use to decide what to pay for.

An FDA initiative encouraging companies to test new cancer drugs earlier in the disease, in first-line or curative settings, rather than only after everything else has failed.

An FDA programme pushing companies to find the best dose of a cancer drug, not just the highest tolerable one.

Project Orbis is a scheme where the FDA and partner regulators (Australia, Canada, UK, Switzerland, Singapore, Brazil, Israel) review a cancer drug at the same time.

FDA processes that let reviewers start on trial data before the full application is filed, speeding approvals by months.

A tumour-agnostic approval lets a drug be used for any cancer carrying a specific molecular feature, regardless of where it started.

A truncated form of the androgen receptor that is permanently switched on and ignores hormone-blocking pills.

Molecular pumps on cancer cells that eject chemotherapy and ADC payloads before they can act.

Why cancer drugs stop working: the tumour either never depended on the target or evolves around the block.

A mutation that stops osimertinib from binding to EGFR; the main on-target way lung cancers escape it.

When hormone therapy stops controlling a hormone-driven breast cancer, either quickly (primary) or after years (acquired).

When a lung adenocarcinoma escapes targeted therapy by turning into a different cell type, usually small-cell.

When lung cancer switches on the MET receptor to bypass a blocked EGFR pill.

A flood of inflammatory signals when immune cells are activated en masse, causing fever, low blood pressure, and sometimes organ failure.

When a targeted drug makes leukaemia cells mature all at once, causing fever, fluid in the lungs, and weight gain. Treatable with steroids if caught early.

ICANS is confusion, speech difficulty, and rarely seizures after CAR-T or bispecific therapy.

Immune-related adverse events (irAEs) are the autoimmune side effects of checkpoint inhibitors: colitis, thyroid problems, rash, hepatitis, pneumonitis.

Interstitial lung disease (ILD) is lung inflammation, a serious side effect of some ADCs (especially Enhertu) and immunotherapy.

HER2 drugs can weaken the heart's pumping, usually reversibly, so heart function is checked every three months during treatment.

When a treatment kills so many cancer cells so fast that their contents overwhelm the kidneys and heart. The reason venetoclax starts at a tiny dose and ramps up.

Trial designs that test one drug across many cancers (basket), many drugs in one cancer (umbrella), or keep adding arms over time (platform).

Having outside radiologists who do not know the treatment assignment re-read every scan in a trial.

Data from routine care rather than clinical trials, used to check whether results hold outside the trial population.