Requests for startups

You cannot manage what you do not measure quickly. Publishing stage at diagnosis by cancer and region every quarter, not years later, would show whether detection efforts are working.

Publish a fake but realistic copy of each secure cancer dataset so researchers can write and test their code at home, then run the finished code on the real data.

Frailty is the strongest predictor of who will be harmed by treatment, but it is rarely measured. Software can estimate it automatically from existing records and flag patients who need a closer look.

Whenever an AI tool gives a result about a patient, the hospital system would permanently record what it saw, which version it was, what it said and what the doctor did with it.

Millions of people have had weight-loss surgery or now take weight-loss drugs. Linking those records to cancer registries would show, cancer by cancer, how much reversing obesity prevents, for almost no cost.

Build and certify a single free tool that strips names and identifying marks from cancer scans and pathology slides, so every hospital stops writing its own.

Nobody publishes how much cancer isotope is made, where, or when supply will fall short. A public observatory would let hospitals and investors plan.

Hospitals in poorer countries often run out of basic, cheap chemotherapy for weeks. A shared live map of stock levels would let buyers and donors act before a child's treatment is interrupted.

Combine registry cancer incidence by district with open trial site locations from ClinicalTrials.gov to map the regions where patients live more than an hour from any trial. Sponsors and funders would use it to decide where to open sites and justify site selection in diversity plans.

No one knows exactly how many oncologists, nurses, physicists and pathologists each country has or needs. A public, regularly updated model would let governments plan training and spot shortfalls years ahead.

Every funder that spends more than $50 million a year on cancer research would publish what it funds in a shared, coded database, so gaps and duplication can be seen across the whole system.

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.

Labs usually use whichever tumour models they already have. A searchable index that finds the model closest to a specific patient's tumour would make experiments more relevant.

Publish a simple report card showing how complete, timely and standard each hospital's cancer data are, so poor recording becomes visible and fixable.

Secure online workrooms where approved researchers can analyse cancer records without downloading them, with the data already cleaned and organised for cancer questions.

Most cancer drug approvals depend on a handful of facts already in the chart, such as the diagnosis, biomarker and line of therapy; sending those automatically in a standard format would return most decisions before the patient leaves the room.

Connect the cancer registry to death records, pharmacy records and scan reports automatically every week, so we always know what happened to every patient without anyone filling in a form.

Patients moving between hospitals often carry paper folders or nothing. A standard electronic summary of diagnosis, treatments, and doses that any system can read would stop repeated tests and dangerous gaps.

We know surprisingly little about what happens to cancer survivors twenty years on. Linking their treatment records to later health records would show which treatments cause which problems and who needs watching.

Every cancer trial's anonymised patient-level data would go into one trusted repository within eighteen months of completion, with a single access committee, so researchers can re-analyse, pool and learn from trials that today stay locked up.

Oncologists spend hours a day typing notes. Software that listens to the consultation and drafts the note, the letter and the orders could return that time to seeing patients.

Trial staff still retype data from the hospital record into the trial database, and monitors then check every entry by hand. Piping data directly and checking by risk would cut cost and errors.

Hospitals rarely know what fraction of their patients got the recommended treatment. Software reading the electronic record can show each team, every month, where care deviated from guidelines.

Give each patient a scrambled code that is the same across hospitals, labs and registries, so records can be joined without anyone seeing names.

Hospitals keep their records at home; researchers send in a programme that runs at each hospital and only the summary results come back.

Several big projects have sequenced the same tumours at different times and places, but their data sit apart. Bringing them together with common analysis would show general rules of how cancers evolve.

Pool the side-effect and quality-of-life data patients report in trials into one open database so regimens can be compared honestly and models can be built.

Different cancers favour different organs, and so do different patients. A model that predicts which organ is at risk could target surveillance and prevention.

To prove a leftover-cancer test works you need blood taken years before relapse. Collecting and freezing yearly samples now makes every future test testable.

Dozens of trials have collected immune, genomic and imaging data on the same drugs. Nobody can analyse them together, so the answer stays hidden in fragments.

Join the national list of who got cancer to the genetic profile of each tumour, so we can see for the whole population which mutations matter and which drugs work for them.

Combinations used off-label in over 200 patients in clinico-genomic databases can be analysed by target trial emulation. Emulations cannot replace trials, but they can rule out the pairs with no signal and flag those with large effects before money is spent on randomised studies.

Before spending millions on a randomised trial, analyse existing patient records as if the trial had already happened. If the answer is obvious or the question is unanswerable, skip or redesign the trial.

Millions of people have had weight-loss surgery or now take weight-loss drugs. Linking those records to cancer registries would show, cancer by cancer, how much reversing obesity prevents, for almost no cost.

Use joined-up pharmacy and hospital records to spot when a common everyday medicine makes a cancer drug work worse or cause more harm.

Before a trial is finalised, run its entry rules against records of real patients with that cancer and report what fraction would qualify. If it is under half, explain why.

Some everyday medicines, such as antibiotics or steroids, seem to blunt immunotherapy. Automatically scanning health records for such harmful pairs would catch them years earlier.

Measure and publish, for every practice-changing result, how long it takes before most eligible patients in each country and hospital actually receive it.

Publish a simple report card showing how complete, timely and standard each hospital's cancer data are, so poor recording becomes visible and fixable.

How long people wait between first noticing something wrong and being diagnosed is barely measured. Recording it routinely and publishing it by hospital would expose where the system loses time.

Connect the cancer registry to death records, pharmacy records and scan reports automatically every week, so we always know what happened to every patient without anyone filling in a form.

We know surprisingly little about what happens to cancer survivors twenty years on. Linking their treatment records to later health records would show which treatments cause which problems and who needs watching.

Record, for every patient, the order of treatments and what happened, so that the most common sequences can be compared and the worst ones flagged.

Surgeons' skill affects whether cancer comes back, but nobody measures it. Recording operations and rating them, increasingly with AI, then linking ratings to outcomes, would make surgical quality visible and improvable.

Rare cancer patients are already tracked in registries. Offering randomisation inside the registry makes trials far cheaper and lets almost anyone take part.

Hospitals keep their records at home; researchers send in a programme that runs at each hospital and only the summary results come back.

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.

Millions of people take common drugs and some get cancer. Running standardised analyses across whole-country records could rank which old drugs deserve a real trial.

Scans find unexpected lumps in the adrenal, thyroid, pancreas and lung. Nobody knows how many matter. A national cohort following them for ten years would tell us whom to watch.

Different cancers favour different organs, and so do different patients. A model that predicts which organ is at risk could target surveillance and prevention.

Join the national list of who got cancer to the genetic profile of each tumour, so we can see for the whole population which mutations matter and which drugs work for them.

After surgery or curative treatment, everyone gets regular blood tests for leftover cancer DNA, and the pooled results power forecasts of who will relapse and trials of acting early.

When a computer suggests a treatment, it should show the doctor the specific trial result and guideline sentence behind the suggestion, so it can be checked and trusted.

When your biopsy or gene test comes back, you get a version written for you, drafted by software and checked and signed by your clinician, alongside the technical report.

Every cancer consultation is recorded with consent and the patient receives the audio plus a checked written summary of what was said and decided.

A simple website that shows, every year, how much research money each cancer receives compared with how many people it kills, so the gaps are impossible to ignore.

Nobody keeps track of how differently the same drug is approved and dosed around the world. A public scoreboard would make the differences visible and push regulators to converge.

Keep a running, well-documented list of cancer practices and approvals that were reversed by later evidence, so the pattern of mistakes is visible and teachable.

Measure and publish, for every practice-changing result, how long it takes before most eligible patients in each country and hospital actually receive it.

When you look up a paper, you should immediately see whether anyone has tried to repeat it and whether they succeeded.

Wherever cancer information appears online, a visible marker shows whether it matches what trusted sources say, with a one-click link to the plain-language evidence.

When a study is retracted or corrected, every guideline and software tool that relied on it would be alerted automatically, so wrong evidence stops influencing care.

Most oncology guidelines exist only in English, and national adaptations lag by years and often diverge. Machine translation checked by a clinician-verifier per language could publish each recommendation update in 20 languages within 30 days, side by side with the source and with local adaptations flagged explicitly.

When an aeroplane crashes, an independent report explains why so it does not happen again. When a cancer drug programme is abandoned, nothing is written. Change that.

Doctors and patients could subscribe to a specific cancer, stage and biomarker and be told, with sources, the moment the recommended treatment changes for that situation.

Hospitals rarely know what fraction of their patients got the recommended treatment. Software reading the electronic record can show each team, every month, where care deviated from guidelines.

For each question patients actually ask, keep a short, current, sourced answer in plain words, updated as evidence changes and available in the languages people speak.

Cancer treatment guidance changes constantly and takes years to reach many clinics. Make guidelines live documents that software can read, updated as evidence arrives and adapted to what each country can afford.

A free web service where any app or hospital system can ask 'what is the recommended treatment for this exact situation today' and get a cited, versioned answer.

Women get more severe side effects than men at identical doses of fluorouracil, several kinase inhibitors and immune checkpoint inhibitors in pooled analyses. Trials should pre-specify sex-stratified drug level and toxicity analyses and, where they differ, run sex-specific dose-finding and label accordingly.

Trials quietly exclude people who do not speak the local language because consent forms and questionnaires exist only in that language. Sponsors should provide validated translations for any language spoken by at least 5 percent of the catchment and fund interpreters; translations of common instruments already exist for dozens of languages.

A majority of people of West African ancestry carry the Duffy-null variant, which lowers baseline neutrophil counts without raising infection risk. Trials apply a single neutrophil cut-off that wrongly labels them unfit, so protocols should use Duffy-specific thresholds; Duffy status is a cheap blood test.

Before a trial is finalised, a paid panel of patients and community members from the groups the trial needs would review it and could require changes to visit schedules, procedures and materials that would deter people like them.

Each trial would set its target mix of patients from cancer registry data on who gets that cancer, by age, sex and ethnicity, and show a public running tally so gaps are visible while there is still time to fix them.

Having had another cancer years ago, or living with controlled HIV or treated hepatitis, still keeps patients out of trials for no scientific reason. Condition-specific rules, as FDA guidance recommended in 2020, would replace blanket bans and widen access in communities where these conditions are more common.

Cancer information tools must meet a tested standard: reading age around 12, main languages of the population, audio versions and clear numbers, or they are not certified for use.

Combine registry cancer incidence by district with open trial site locations from ClinicalTrials.gov to map the regions where patients live more than an hour from any trial. Sponsors and funders would use it to decide where to open sites and justify site selection in diversity plans.

Labs usually use whichever tumour models they already have. A searchable index that finds the model closest to a specific patient's tumour would make experiments more relevant.

People with serious mental illness or dementia are routinely excluded from cancer trials by vague compliance clauses, though they get cancer just as often and do worse. Replacing those clauses with assessable criteria, and funding supported consent and accommodations such as longer visits, would let them take part.

People should not have to pay to be in a trial. Sponsors would routinely cover travel, hotel and food costs, paid up front rather than claimed back, which is already accepted by regulators as fair rather than coercive.

Trials usually take only patients who are up and about most of the day. Those who spend more time resting, a common group in real clinics, are excluded, so nobody knows how to treat them. A dedicated group in each trial would answer that.

Once a patient has safely had the first few doses of a trial drug at the hospital, later doses could be given at home or a local clinic, with blood tests done nearby, so distance no longer decides who can join.

Instead of excluding sicker patients entirely, trials would run a side group for them, receiving the new drug with closer monitoring, so we learn how it behaves in the people who will actually get it.

Most of the world's cancer patients live in countries that host almost no registrational trials. Including sites there, and paying to build them up, would make results apply globally and speed local access.

Build the shared software, legal templates and data standards that let any hospital in the world join a cancer trial in weeks instead of years, the way the internet let any computer join the network.

Trials quietly exclude people who do not speak the local language because consent forms and questionnaires exist only in that language. Sponsors should provide validated translations for any language spoken by at least 5 percent of the catchment and fund interpreters; translations of common instruments already exist for dozens of languages.

A majority of people of West African ancestry carry the Duffy-null variant, which lowers baseline neutrophil counts without raising infection risk. Trials apply a single neutrophil cut-off that wrongly labels them unfit, so protocols should use Duffy-specific thresholds; Duffy status is a cheap blood test.

Each trial would set its target mix of patients from cancer registry data on who gets that cancer, by age, sex and ethnicity, and show a public running tally so gaps are visible while there is still time to fix them.

Before a trial is finalised, run its entry rules against records of real patients with that cancer and report what fraction would qualify. If it is under half, explain why.

Having had another cancer years ago, or living with controlled HIV or treated hepatitis, still keeps patients out of trials for no scientific reason. Condition-specific rules, as FDA guidance recommended in 2020, would replace blanket bans and widen access in communities where these conditions are more common.

Combine registry cancer incidence by district with open trial site locations from ClinicalTrials.gov to map the regions where patients live more than an hour from any trial. Sponsors and funders would use it to decide where to open sites and justify site selection in diversity plans.

People with serious mental illness or dementia are routinely excluded from cancer trials by vague compliance clauses, though they get cancer just as often and do worse. Replacing those clauses with assessable criteria, and funding supported consent and accommodations such as longer visits, would let them take part.

Most trials copy the same kidney and liver cut-offs, such as creatinine clearance above 60, regardless of how the drug is cleared. Setting each threshold from the drug's own clearance route and organ-impairment pharmacokinetic studies would let patients with mild organ impairment join safely instead of being excluded.

People should not have to pay to be in a trial. Sponsors would routinely cover travel, hotel and food costs, paid up front rather than claimed back, which is already accepted by regulators as fair rather than coercive.

Thousands of patients have received standard treatment in the control arms of past trials. Pooling their anonymised data would let new trials borrow from them and randomise fewer patients to old treatments.

When the standard treatment has been given to thousands of similar patients in earlier trials, a new trial could randomise fewer people to it and lean on that history, as long as the old data still matches.

Once a patient has safely had the first few doses of a trial drug at the hospital, later doses could be given at home or a local clinic, with blood tests done nearby, so distance no longer decides who can join.

Instead of opening a trial at fifty hospitals and waiting for patients, keep a network of pre-vetted clinics ready and switch a trial on where a matching patient is found.

At each point where treatment is chosen, software checks the patient's record against open trials and the clinician must note which were discussed, so trials stop being something only some people hear about.

Three companies testing three drugs against the same standard treatment each recruit their own control group. Pooling those controls in one shared study would need fewer patients and answer faster.

Build the shared software, legal templates and data standards that let any hospital in the world join a cancer trial in weeks instead of years, the way the internet let any computer join the network.

Rate sponsors on whether they publish their results, share data and register outcomes honestly. Hospitals and patients can then prefer sponsors that behave well.

Trials sometimes quietly swap the outcome they promised to measure for one that looks better. Software can compare the registered plan with the published paper and flag the switch.

Drugs fail for distinct reasons: wrong target, drug never reached it, unacceptable toxicity, unselected population or poor trial design. A shared machine-readable taxonomy applied to every discontinued oncology programme in public pipeline databases would show where the system breaks, as AstraZeneca and Pfizer's own attrition analyses did.

Keep a running, well-documented list of cancer practices and approvals that were reversed by later evidence, so the pattern of mistakes is visible and teachable.

Thousands of cancer compounds that stopped development for portfolio rather than scientific reasons sit unused in university freezers and company archives. A public catalogue listing each asset's mechanism, stage, data, reason for stopping and licensing contact, with a standard research licence and a brokerage function, would let academic and non-profit developers adopt them.

For each drug target, show how many programmes have been tried against it and how many failed, so new teams know what they are up against.

When a study is retracted, everything built on it should get a warning. Today, retracted cancer papers keep being cited and used for years.

Mine grant databases and the literature to find cancer types and questions with heavy burden and zero active projects, then publish the list so funders and scientists can go there.

When an aeroplane crashes, an independent report explains why so it does not happen again. When a cancer drug programme is abandoned, nothing is written. Change that.

For each type of laboratory model, keep a public record of how often its predictions came true in patients, so that researchers know which models to trust for which question.

Large phase 3 trials often continue for years after interim data show the drug is unlikely to work. Pre-registering futility boundaries at 30 to 50 percent of the information, judged by independent committees and reported publicly, would stop them earlier, sparing patients and freeing money for better ideas.

Companies shelve drugs that were safe but failed in the disease they tried. An oncology commons cataloguing discontinued assets with their mechanism, human pharmacokinetics, safety data and reasons for discontinuation, under template access terms, would let others test them where they might work, as NCATS and AstraZeneca schemes have shown.

Every cancer trial's anonymised patient-level data would go into one trusted repository within eighteen months of completion, with a single access committee, so researchers can re-analyse, pool and learn from trials that today stay locked up.

Phase 3 trials are often launched on a small single-arm response-rate study with no comparison group, and most then fail. Requiring a randomised phase 2 with a concurrent control first, with exceptions only for large effects in refractory settings, would filter out weak drugs earlier.

Pumping drugs slowly through fine tubes into a brain tumour can bypass the barrier, but the fluid often leaks away. Robotic placement and live scans would show where it actually goes.

Patients now ask AI assistants about their cancer. Test those assistants regularly on real questions, publish the scores, and certify the ones that meet the bar.

Set common rules for how hospitals check that an AI tool still works as the scanners, patients and practices around it change, and when it must be switched off.

When a computer suggests a treatment, it should show the doctor the specific trial result and guideline sentence behind the suggestion, so it can be checked and trusted.

Whenever an AI tool gives a result about a patient, the hospital system would permanently record what it saw, which version it was, what it said and what the doctor did with it.

Build and certify a single free tool that strips names and identifying marks from cancer scans and pathology slides, so every hospital stops writing its own.

AI systems claim to find new uses for old drugs, but their predictions are rarely tested fairly. Publish their cancer predictions in advance and score them against trial results.

Every staging scan contains a precise measure of muscle mass that nobody looks at. Software could report it automatically and flag patients heading for wasting.

Flu vaccines are chosen by predicting which virus strains will dominate next season. The same forecasting maths could predict which resistance mutation a patient's tumour will develop next.

No one keeps score of which laboratory models actually predicted what happened in patients. A public scoreboard would show which models to trust.

Cancer AI models are usually tested on data from the same hospital they were built on. A registry of independent test datasets, and a rule that every model reports performance on at least one, would show which models really work.

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.

Autologous cell therapy batches fail more often than any other medicine because each patient's starting cells behave differently and the process runs without feedback. Inline sensors for metabolites, cell counts and cytokines, feeding models that adjust feeding and harvest timing in real time, could rescue batches that would otherwise be discarded.

A free web service where any app or hospital system can ask 'what is the recommended treatment for this exact situation today' and get a cited, versioned answer.

Dozens of trials have collected immune, genomic and imaging data on the same drugs. Nobody can analyse them together, so the answer stays hidden in fragments.

Companies with drugs that might work together rarely test them because the legal negotiation takes longer than the trial. A pre-written standard contract would fix that.

Companies fear that testing a combination will hand a competitor a patent. A shared pool where combination patents are cross-licensed by default would remove the fear.

Thousands of cancer compounds that stopped development for portfolio rather than scientific reasons sit unused in university freezers and company archives. A public catalogue listing each asset's mechanism, stage, data, reason for stopping and licensing contact, with a standard research licence and a brokerage function, would let academic and non-profit developers adopt them.

Most of the delay in testing two companies' drugs together is lawyers negotiating from scratch. A single standard agreement, blessed by regulators, would let them sign in weeks.

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.

Companies and public funders would jointly pay for standardised experiments that confirm or refute new cancer targets, sharing all results openly, so nobody wastes years on a target that does not hold up.

Thousands of patients have received standard treatment in the control arms of past trials. Pooling their anonymised data would let new trials borrow from them and randomise fewer patients to old treatments.

Every cancer trial's anonymised patient-level data would go into one trusted repository within eighteen months of completion, with a single access committee, so researchers can re-analyse, pool and learn from trials that today stay locked up.

India's 2012 compulsory licence on sorafenib cut its price by about 97% and its 2019 cap on trade margins lowered the shelf price of 42 cancer drugs; voluntary licences through a patent pool would achieve the same without a fight.

Companies hold thousands of well-characterised drugs that could help rare cancers, but each request takes a year of legal negotiation. One standing agreement would unblock it.

Companies would put their cancer drugs into a shared licensing pool so that any qualified investigator can test combinations of drugs from different owners under one standard agreement, with royalties split by a fixed formula.

Academic labs find new tumour targets but cannot turn an antibody into an antibody-drug conjugate or a bispecific without licensed linker and payload technology. A shared platform would provide that at no cost for first trials.

Three companies testing three drugs against the same standard treatment each recruit their own control group. Pooling those controls in one shared study would need fewer patients and answer faster.

Companies and public funders would pool money and scientists to crack the hardest cancer proteins, such as MYC and mutant p53, sharing everything openly until there is a real drug candidate, then competing on the final product.

Trials sometimes quietly swap the outcome they promised to measure for one that looks better. Software can compare the registered plan with the published paper and flag the switch.

Before spending millions on a randomised trial, analyse existing patient records as if the trial had already happened. If the answer is obvious or the question is unanswerable, skip or redesign the trial.

The traditional way of finding a dose looks only at severe side effects in the first month. Modern statistical designs can use all patients' data and count the grumbling, long-lasting problems that make people quit months later.

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.

Large phase 3 trials often continue for years after interim data show the drug is unlikely to work. Pre-registering futility boundaries at 30 to 50 percent of the information, judged by independent committees and reported publicly, would stop them earlier, sparing patients and freeing money for better ideas.

Thousands of patients have received standard treatment in the control arms of past trials. Pooling their anonymised data would let new trials borrow from them and randomise fewer patients to old treatments.

When the standard treatment has been given to thousands of similar patients in earlier trials, a new trial could randomise fewer people to it and lean on that history, as long as the old data still matches.

Trial staff still retype data from the hospital record into the trial database, and monitors then check every entry by hand. Piping data directly and checking by risk would cut cost and errors.

Rare cancer patients are already tracked in registries. Offering randomisation inside the registry makes trials far cheaper and lets almost anyone take part.

Phase 3 trials are often launched on a small single-arm response-rate study with no comparison group, and most then fail. Requiring a randomised phase 2 with a concurrent control first, with exceptions only for large effects in refractory settings, would filter out weak drugs earlier.

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.

Three companies testing three drugs against the same standard treatment each recruit their own control group. Pooling those controls in one shared study would need fewer patients and answer faster.

Instead of stopping after the mid-sized trial, waiting a year, then starting the big one, run them as one study with a clear pre-agreed rule for continuing. This saves a year or more per drug.

Some cancers reach the brain in up to a quarter of patients. Prevention trials could aim specifically at stopping that, instead of treating it once it has happened.

Hospitals in poorer countries often run out of basic, cheap chemotherapy for weeks. A shared live map of stock levels would let buyers and donors act before a child's treatment is interrupted.

Most oncology guidelines exist only in English, and national adaptations lag by years and often diverge. Machine translation checked by a clinician-verifier per language could publish each recommendation update in 20 languages within 30 days, side by side with the source and with local adaptations flagged explicitly.

A drug approved in the US may take five years to reach a patient in Poland or never reach Nigeria. A live public tracker would show exactly where and why it is stuck.

In much of Africa and South Asia there is about one pathologist per million people, so the pathologist, not the tissue sample, is the diagnostic bottleneck. A slide scanner in each district lab, routing images to a pooled roster of pathologists including diaspora volunteers, could give a diagnosis in days instead of months; the gap is scale and financing.

Cancer treatment guidance changes constantly and takes years to reach many clinics. Make guidelines live documents that software can read, updated as evidence arrives and adapted to what each country can afford.

Cervical cancer cannot be cured by external radiotherapy alone; it needs brachytherapy, internal radiation that LMIC radiotherapy centres frequently lack or cannot keep running because iridium-192 sources must be replaced every three months. Cobalt-60 sources last about five years and give equivalent doses, so funding cobalt units for every regional centre closes a well-understood cure gap.

Medical physicists, who keep radiotherapy machines accurate and safe, are in even shorter supply than oncologists in under-resourced systems. Routine linac quality assurance is now largely automated and log-file based, so one physicist could review it remotely while trained radiation therapists take the measurements, covering three to five machines once regulators define the supervision standard.

Radioactive cancer drugs decay while they travel and get stuck at borders. Regional production and simpler transport rules would get more doses to patients on time.

Hospitals without enough physicists could upload scans to a shared planning centre, where AI drafts the treatment plan and remote experts finish and check it within a day.

Cancer units in poorer countries lose treatment days, spoil drugs and damage machines during power cuts. Solar panels with batteries sized for the cancer unit would remove that failure point.

India's 2012 compulsory licence on sorafenib cut its price by about 97% and its 2019 cap on trade margins lowered the shelf price of 42 cancer drugs; voluntary licences through a patent pool would achieve the same without a fight.

In places where sophisticated machines break down, a modern version of the older cobalt radiotherapy unit, upgraded with image guidance, could treat more people reliably while infrastructure catches up.

Most of the world's cancer patients live in countries that host almost no registrational trials. Including sites there, and paying to build them up, would make results apply globally and speed local access.

Surgeons often cannot see where a tumour ends. Fluorescent dyes and AI-read imaging in the operating theatre can show them, cutting repeat operations. Make this routine everywhere.

Companies with drugs that might work together rarely test them because the legal negotiation takes longer than the trial. A pre-written standard contract would fix that.

Combinations used off-label in over 200 patients in clinico-genomic databases can be analysed by target trial emulation. Emulations cannot replace trials, but they can rule out the pairs with no signal and flag those with large effects before money is spent on randomised studies.

Some everyday medicines, such as antibiotics or steroids, seem to blunt immunotherapy. Automatically scanning health records for such harmful pairs would catch them years earlier.

Companies fear that testing a combination will hand a competitor a patent. A shared pool where combination patents are cross-licensed by default would remove the fear.

Most of the delay in testing two companies' drugs together is lawyers negotiating from scratch. A single standard agreement, blessed by regulators, would let them sign in weeks.

Ask experts and models to predict, in public, which registered combination trials will meet their endpoint. Track who is right, and use the best forecasters to decide what to fund.

Record, for every patient, the order of treatments and what happened, so that the most common sequences can be compared and the worst ones flagged.

If most tumours escape a drug by the same back-up route, blocking that route from the start may prevent resistance rather than chase it.

A rice-grain-sized implant releases microdoses of up to 20 drugs into separate spots of a tumour for one to three days, then is removed so pathologists can see which drug worked in that person's own tumour. First-in-human studies have been done in breast, sarcoma and brain tumours.

Simulate how two drugs interact in the body and the tumour to pick a starting dose and schedule, instead of guessing from single-drug data.

Use existing cell-line and organoid data to score thousands of drug pairs, publish the ranking openly, and only test the top of the list in people.

Companies would put their cancer drugs into a shared licensing pool so that any qualified investigator can test combinations of drugs from different owners under one standard agreement, with royalties split by a fixed formula.

When two cancer drugs are combined, each is usually given at its full single-agent dose, which often proves too toxic. Testing a grid of dose pairs would find combinations that work with tolerable side effects.

Software can already spot impossible statistics, mismatched p-values and duplicated images in a paper. Journals should run these checks on every submission, as spell-check runs on every document.

Each batch of cells used in an experiment would carry a small digital record showing when it was authenticated, tested for contamination, and how many times it had been grown, attached to the published result.

When you look up a paper, you should immediately see whether anyone has tried to repeat it and whether they succeeded.

When a study is retracted, everything built on it should get a warning. Today, retracted cancer papers keep being cited and used for years.

When a study is retracted or corrected, every guideline and software tool that relied on it would be alerted automatically, so wrong evidence stops influencing care.

Results can change when a supplier changes a batch of serum, antibody or growth factor. Recording which batch was used in each experiment, in a shared ledger, would let these effects be spotted.

No one keeps score of which laboratory models actually predicted what happened in patients. A public scoreboard would show which models to trust.

Electronic notebooks record when each experiment was done and what the raw result was. Submitting them with the paper would show whether the analysis was planned or fitted after the fact.

Secure online workrooms where approved researchers can analyse cancer records without downloading them, with the data already cleaned and organised for cancer questions.

Companies and public funders would jointly pay for standardised experiments that confirm or refute new cancer targets, sharing all results openly, so nobody wastes years on a target that does not hold up.

Cancer AI models are usually tested on data from the same hospital they were built on. A registry of independent test datasets, and a rule that every model reports performance on at least one, would show which models really work.

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.

Large drugs such as antibody-drug conjugates must cross vessel walls and travel through dense tissue. A chip with flowing channels and human tissue can measure how far they get.

Each batch of cells used in an experiment would carry a small digital record showing when it was authenticated, tested for contamination, and how many times it had been grown, attached to the published result.

Labs usually use whichever tumour models they already have. A searchable index that finds the model closest to a specific patient's tumour would make experiments more relevant.

No one keeps score of which laboratory models actually predicted what happened in patients. A public scoreboard would show which models to trust.

For each type of laboratory model, keep a public record of how often its predictions came true in patients, so that researchers know which models to trust for which question.

A rice-grain-sized implant releases microdoses of up to 20 drugs into separate spots of a tumour for one to three days, then is removed so pathologists can see which drug worked in that person's own tumour. First-in-human studies have been done in breast, sarcoma and brain tumours.

Dormant cancer cells hide in bone marrow. A lab-built model of that hiding place would let us watch them sleep and wake, and test drugs on them.

Use existing cell-line and organoid data to score thousands of drug pairs, publish the ranking openly, and only test the top of the list in people.

Drugs that reach brain tissue may still fail to reach the fluid where cancer spreads along the linings. A lab model of that second barrier would let us screen for drugs that cross it.

Damage to the lungs, heart or liver is a common reason cancer drugs fail. Connected chips of human tissue may spot this earlier than animal tests.

Cancer usually kills by spreading to bone, liver, lung or brain. Almost all laboratory models grow tumours under the skin instead, where the surroundings are nothing like those organs.

After surgery, a tumour with its blood vessels can be connected to a pump and kept alive for hours or days, allowing drugs to be tested in genuinely human tissue.

Most patients take supplements and rarely tell their oncologist. Ask routinely and check automatically for interactions with chemotherapy and targeted drugs.

Women get more severe side effects than men at identical doses of fluorouracil, several kinase inhibitors and immune checkpoint inhibitors in pooled analyses. Trials should pre-specify sex-stratified drug level and toxicity analyses and, where they differ, run sex-specific dose-finding and label accordingly.

Use joined-up pharmacy and hospital records to spot when a common everyday medicine makes a cancer drug work worse or cause more harm.

Every staging scan contains a precise measure of muscle mass that nobody looks at. Software could report it automatically and flag patients heading for wasting.

Children treated for cancer are followed for decades in a study that has changed how they are treated. Adults have nothing similar. Build it.

Antibody-drug conjugates deliver chemotherapy payloads into tumours but cause lung, eye and nerve damage that tracks total exposure, and several were approved without an optimised dose. Randomised comparisons of lower doses, longer intervals and capped cumulative payload could keep the benefit while cutting these harms.

Instead of starting everyone on the full dose and cutting back after side effects, start lower and increase in patients who tolerate it. This keeps more people on treatment and is how blood-pressure drugs are given.

TGF-beta is a signal that keeps immune cells out of tumours, but blocking it throughout the body caused bleeding and heart toxicity and sank bintrafusp alfa. Tethering the blocker to tumour stroma with a FAP anchor, a collagen-binding domain or a protease-activated mask could give the benefit without the harm.

Drugs that grab T cells and drag them onto tumours work well in blood cancers. The same trick aimed at tumour-eating cells might work where T cells are absent.

Damage to the lungs, heart or liver is a common reason cancer drugs fail. Connected chips of human tissue may spot this earlier than animal tests.

Pool the side-effect and quality-of-life data patients report in trials into one open database so regimens can be compared honestly and models can be built.

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.

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.

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.

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.

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.

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.

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.

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.

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.

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.

There are far too few genetic counsellors. A validated chatbot can do the standard pre-test education, leaving people with complex needs for humans.

No one knows exactly how many oncologists, nurses, physicists and pathologists each country has or needs. A public, regularly updated model would let governments plan training and spot shortfalls years ahead.

Hospitals rarely know which step, the scanner, the biopsy, the pathologist or the clinic slot, is causing the queue. Modelling the pathway like a factory line shows where a small change would remove weeks of waiting.

Most cancer drug approvals depend on a handful of facts already in the chart, such as the diagnosis, biomarker and line of therapy; sending those automatically in a standard format would return most decisions before the patient leaves the room.

In much of Africa and South Asia there is about one pathologist per million people, so the pathologist, not the tissue sample, is the diagnostic bottleneck. A slide scanner in each district lab, routing images to a pooled roster of pathologists including diaspora volunteers, could give a diagnosis in days instead of months; the gap is scale and financing.

Oncologists spend hours a day typing notes. Software that listens to the consultation and drafts the note, the letter and the orders could return that time to seeing patients.

Medical physicists, who keep radiotherapy machines accurate and safe, are in even shorter supply than oncologists in under-resourced systems. Routine linac quality assurance is now largely automated and log-file based, so one physicist could review it remotely while trained radiation therapists take the measurements, covering three to five machines once regulators define the supervision standard.

Hospitals without enough physicists could upload scans to a shared planning centre, where AI drafts the treatment plan and remote experts finish and check it within a day.

Women get more severe side effects than men at identical doses of fluorouracil, several kinase inhibitors and immune checkpoint inhibitors in pooled analyses. Trials should pre-specify sex-stratified drug level and toxicity analyses and, where they differ, run sex-specific dose-finding and label accordingly.

The traditional way of finding a dose looks only at severe side effects in the first month. Modern statistical designs can use all patients' data and count the grumbling, long-lasting problems that make people quit months later.

Most trials copy the same kidney and liver cut-offs, such as creatinine clearance above 60, regardless of how the drug is cleared. Setting each threshold from the drug's own clearance route and organ-impairment pharmacokinetic studies would let patients with mild organ impairment join safely instead of being excluded.

If most tumours escape a drug by the same back-up route, blocking that route from the start may prevent resistance rather than chase it.

Antibody-drug conjugates deliver chemotherapy payloads into tumours but cause lung, eye and nerve damage that tracks total exposure, and several were approved without an optimised dose. Randomised comparisons of lower doses, longer intervals and capped cumulative payload could keep the benefit while cutting these harms.

Simulate how two drugs interact in the body and the tumour to pick a starting dose and schedule, instead of guessing from single-drug data.

Instead of starting everyone on the full dose and cutting back after side effects, start lower and increase in patients who tolerate it. This keeps more people on treatment and is how blood-pressure drugs are given.

When two cancer drugs are combined, each is usually given at its full single-agent dose, which often proves too toxic. Testing a grid of dose pairs would find combinations that work with tolerable side effects.

Every patient finishing treatment should get a clear document listing what they had, what to watch for, and when to be checked. Software can write it from the record so it actually happens.

Patients wait weeks for a manufacturing slot while other slots go unused when a patient drops out. A shared booking system would match spare slots to waiting patients.

How long people wait between first noticing something wrong and being diagnosed is barely measured. Recording it routinely and publishing it by hospital would expose where the system loses time.

Hospitals rarely know which step, the scanner, the biopsy, the pathologist or the clinic slot, is causing the queue. Modelling the pathway like a factory line shows where a small change would remove weeks of waiting.

Most cancer drug approvals depend on a handful of facts already in the chart, such as the diagnosis, biomarker and line of therapy; sending those automatically in a standard format would return most decisions before the patient leaves the room.

Patients moving between hospitals often carry paper folders or nothing. A standard electronic summary of diagnosis, treatments, and doses that any system can read would stop repeated tests and dangerous gaps.

Hospitals rarely know what fraction of their patients got the recommended treatment. Software reading the electronic record can show each team, every month, where care deviated from guidelines.

Cancer treatment guidance changes constantly and takes years to reach many clinics. Make guidelines live documents that software can read, updated as evidence arrives and adapted to what each country can afford.

Drugs fail for distinct reasons: wrong target, drug never reached it, unacceptable toxicity, unselected population or poor trial design. A shared machine-readable taxonomy applied to every discontinued oncology programme in public pipeline databases would show where the system breaks, as AstraZeneca and Pfizer's own attrition analyses did.

Thousands of cancer compounds that stopped development for portfolio rather than scientific reasons sit unused in university freezers and company archives. A public catalogue listing each asset's mechanism, stage, data, reason for stopping and licensing contact, with a standard research licence and a brokerage function, would let academic and non-profit developers adopt them.

Companies shelve drugs that were safe but failed in the disease they tried. An oncology commons cataloguing discontinued assets with their mechanism, human pharmacokinetics, safety data and reasons for discontinuation, under template access terms, would let others test them where they might work, as NCATS and AstraZeneca schemes have shown.

Companies and public funders would jointly pay for standardised experiments that confirm or refute new cancer targets, sharing all results openly, so nobody wastes years on a target that does not hold up.

Companies hold thousands of well-characterised drugs that could help rare cancers, but each request takes a year of legal negotiation. One standing agreement would unblock it.

Academic labs find new tumour targets but cannot turn an antibody into an antibody-drug conjugate or a bispecific without licensed linker and payload technology. A shared platform would provide that at no cost for first trials.

Personalised cancer vaccines and cell therapies need a manufacturing run for each patient, and universities cannot afford their own plants. Regional closed, automated, modular GMP facilities offering slots to academic trials at cost, with common release testing and a shared quality system, modelled on the UK Cell and Gene Therapy Catapult centre, would let academic groups run these trials.

Investors will not back a single university drug because most fail. A fund that finances fifty of them at once in exchange for a small slice of each one's future royalties spreads the risk enough to attract capital.

Rate sponsors on whether they publish their results, share data and register outcomes honestly. Hospitals and patients can then prefer sponsors that behave well.

When your biopsy or gene test comes back, you get a version written for you, drafted by software and checked and signed by your clinician, alongside the technical report.

Before a trial is finalised, a paid panel of patients and community members from the groups the trial needs would review it and could require changes to visit schedules, procedures and materials that would deter people like them.

Every cancer consultation is recorded with consent and the patient receives the audio plus a checked written summary of what was said and decided.

Cancer information tools must meet a tested standard: reading age around 12, main languages of the population, audio versions and clear numbers, or they are not certified for use.

Doctors and patients could subscribe to a specific cancer, stage and biomarker and be told, with sources, the moment the recommended treatment changes for that situation.

At each point where treatment is chosen, software checks the patient's record against open trials and the clinician must note which were discussed, so trials stop being something only some people hear about.

Nobody publishes how much cancer isotope is made, where, or when supply will fall short. A public observatory would let hospitals and investors plan.

Patients wait weeks for a manufacturing slot while other slots go unused when a patient drops out. A shared booking system would match spare slots to waiting patients.

Radioactive cancer drugs decay while they travel and get stuck at borders. Regional production and simpler transport rules would get more doses to patients on time.

The poisons carried by antibody-drug conjugates are so toxic that only a few factories can make them, and they are booked years ahead. Making them in small continuous reactors would ease the bottleneck.

Autologous cell therapy batches fail more often than any other medicine because each patient's starting cells behave differently and the process runs without feedback. Inline sensors for metabolites, cell counts and cytokines, feeding models that adjust feeding and harvest timing in real time, could rescue batches that would otherwise be discarded.

Each cell-therapy machine uses its own proprietary process and cartridges. A common standard would let a process run on any machine, like a document opening in any word processor.

Personalised cancer vaccines and cell therapies need a manufacturing run for each patient, and universities cannot afford their own plants. Regional closed, automated, modular GMP facilities offering slots to academic trials at cost, with common release testing and a shared quality system, modelled on the UK Cell and Gene Therapy Catapult centre, would let academic groups run these trials.

Pancreatic tumours are packed with a type of white blood cell that shuts down the immune attack. Blocking the signal that recruits them may open the tumour to immunotherapy.

A rice-grain-sized implant releases microdoses of up to 20 drugs into separate spots of a tumour for one to three days, then is removed so pathologists can see which drug worked in that person's own tumour. First-in-human studies have been done in breast, sarcoma and brain tumours.

Tumours fill with immune cells that protect them. Earlier drugs tried to remove those cells and failed. Newer ones aim to switch them to the attacking side.

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.

TGF-beta is a signal that keeps immune cells out of tumours, but blocking it throughout the body caused bleeding and heart toxicity and sank bintrafusp alfa. Tethering the blocker to tumour stroma with a FAP anchor, a collagen-binding domain or a protease-activated mask could give the benefit without the harm.

Drugs that grab T cells and drag them onto tumours work well in blood cancers. The same trick aimed at tumour-eating cells might work where T cells are absent.

Some harmless bacteria naturally grow in the low-oxygen core of tumours. Engineering them to produce immune-activating drugs turns them into tiny factories inside the tumour.

Before spending millions on a randomised trial, analyse existing patient records as if the trial had already happened. If the answer is obvious or the question is unanswerable, skip or redesign the trial.

A simple website that shows, every year, how much research money each cancer receives compared with how many people it kills, so the gaps are impossible to ignore.

Mine grant databases and the literature to find cancer types and questions with heavy burden and zero active projects, then publish the list so funders and scientists can go there.

Ask experts and models to predict, in public, which registered combination trials will meet their endpoint. Track who is right, and use the best forecasters to decide what to fund.

Every funder that spends more than $50 million a year on cancer research would publish what it funds in a shared, coded database, so gaps and duplication can be seen across the whole system.

Large phase 3 trials often continue for years after interim data show the drug is unlikely to work. Pre-registering futility boundaries at 30 to 50 percent of the information, judged by independent committees and reported publicly, would stop them earlier, sparing patients and freeing money for better ideas.

For each drug target, show how many programmes have been tried against it and how many failed, so new teams know what they are up against.

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.

Cancers with faulty DNA proof-reading depend on one particular unwinding enzyme to survive. Blocking it kills them and spares normal cells.

Some harmless bacteria naturally grow in the low-oxygen core of tumours. Engineering them to produce immune-activating drugs turns them into tiny factories inside the tumour.

The brain imports iron through the transferrin receptor. Antibody shuttle domains that bind that receptor raise brain exposure roughly ten to fifty-fold in primates and are already used in clinical Alzheimer's antibodies; the same engineering could carry antibody-drug conjugates or T-cell engagers to brain metastases.

Companies and public funders would pool money and scientists to crack the hardest cancer proteins, such as MYC and mutant p53, sharing everything openly until there is a real drug candidate, then competing on the final product.

You cannot manage what you do not measure quickly. Publishing stage at diagnosis by cancer and region every quarter, not years later, would show whether detection efforts are working.

Millions now wear continuous glucose monitors. A sudden, unexplained worsening of glucose control in a middle-aged wearer could be flagged as a possible early sign of pancreatic cancer.

How long people wait between first noticing something wrong and being diagnosed is barely measured. Recording it routinely and publishing it by hospital would expose where the system loses time.

A platelet count that is normal but rising year on year, or haemoglobin drifting down, can signal cancer. Records already hold these trends; software could use them.

To prove a leftover-cancer test works you need blood taken years before relapse. Collecting and freezing yearly samples now makes every future test testable.

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.

Pancreatic tumours are packed with a type of white blood cell that shuts down the immune attack. Blocking the signal that recruits them may open the tumour to immunotherapy.

Tumours fill with immune cells that protect them. Earlier drugs tried to remove those cells and failed. Newer ones aim to switch them to the attacking side.

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.

Dozens of trials have collected immune, genomic and imaging data on the same drugs. Nobody can analyse them together, so the answer stays hidden in fragments.

Flu vaccines are chosen by predicting which virus strains will dominate next season. The same forecasting maths could predict which resistance mutation a patient's tumour will develop next.

Blood tests tell you which tumour sub-populations are growing; scans tell you which lesions are growing. Joining the two would tell you where to biopsy or irradiate.

Blood tests can already detect tumour DNA. Reporting which sub-populations of the tumour are growing or shrinking, cycle by cycle, would turn the test into an evolution monitor.

Several big projects have sequenced the same tumours at different times and places, but their data sit apart. Bringing them together with common analysis would show general rules of how cancers evolve.

Subsets of lung, bladder and breast cancers carry raised APOBEC enzyme activity that keeps generating new mutations, feeding resistance. Blocking APOBEC3 alongside a targeted drug aims not to kill cells but to slow the rate at which resistant variants arise; the inhibitors are still in discovery.

Flu vaccines are chosen by predicting which virus strains will dominate next season. The same forecasting maths could predict which resistance mutation a patient's tumour will develop next.

Blood tests can already detect tumour DNA. Reporting which sub-populations of the tumour are growing or shrinking, cycle by cycle, would turn the test into an evolution monitor.

If most tumours escape a drug by the same back-up route, blocking that route from the start may prevent resistance rather than chase it.

Cancers with faulty DNA proof-reading depend on one particular unwinding enzyme to survive. Blocking it kills them and spares normal cells.

Subsets of lung, bladder and breast cancers carry raised APOBEC enzyme activity that keeps generating new mutations, feeding resistance. Blocking APOBEC3 alongside a targeted drug aims not to kill cells but to slow the rate at which resistant variants arise; the inhibitors are still in discovery.

Blood tests tell you which tumour sub-populations are growing; scans tell you which lesions are growing. Joining the two would tell you where to biopsy or irradiate.

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.

Different cancers favour different organs, and so do different patients. A model that predicts which organ is at risk could target surveillance and prevention.

Cancer usually kills by spreading to bone, liver, lung or brain. Almost all laboratory models grow tumours under the skin instead, where the surroundings are nothing like those organs.

Some cancers reach the brain in up to a quarter of patients. Prevention trials could aim specifically at stopping that, instead of treating it once it has happened.

A small ultrasound device implanted in the skull can be switched on at each chemotherapy visit to briefly open the brain barrier, letting drugs in every cycle.

Pumping drugs slowly through fine tubes into a brain tumour can bypass the barrier, but the fluid often leaks away. Robotic placement and live scans would show where it actually goes.

Drugs that reach brain tissue may still fail to reach the fluid where cancer spreads along the linings. A lab model of that second barrier would let us screen for drugs that cross it.

The brain imports iron through the transferrin receptor. Antibody shuttle domains that bind that receptor raise brain exposure roughly ten to fifty-fold in primates and are already used in clinical Alzheimer's antibodies; the same engineering could carry antibody-drug conjugates or T-cell engagers to brain metastases.

Some cancers reach the brain in up to a quarter of patients. Prevention trials could aim specifically at stopping that, instead of treating it once it has happened.

Surgeons' skill affects whether cancer comes back, but nobody measures it. Recording operations and rating them, increasingly with AI, then linking ratings to outcomes, would make surgical quality visible and improvable.

Cervical cancer cannot be cured by external radiotherapy alone; it needs brachytherapy, internal radiation that LMIC radiotherapy centres frequently lack or cannot keep running because iridium-192 sources must be replaced every three months. Cobalt-60 sources last about five years and give equivalent doses, so funding cobalt units for every regional centre closes a well-understood cure gap.

Pumping drugs slowly through fine tubes into a brain tumour can bypass the barrier, but the fluid often leaks away. Robotic placement and live scans would show where it actually goes.

In places where sophisticated machines break down, a modern version of the older cobalt radiotherapy unit, upgraded with image guidance, could treat more people reliably while infrastructure catches up.

Surgeons often cannot see where a tumour ends. Fluorescent dyes and AI-read imaging in the operating theatre can show them, cutting repeat operations. Make this routine everywhere.

Generic imatinib and abiraterone can cost patients hundreds of dollars a month through insurance yet be sold at cost plus a fixed markup by transparent pharmacies; making that the default channel for oral cancer generics would save patients and plans money.

AI systems claim to find new uses for old drugs, but their predictions are rarely tested fairly. Publish their cancer predictions in advance and score them against trial results.

Preparing a regulatory filing requires expensive specialist software and consultants. Free, open tools would let universities and small generic firms file in more countries.

Millions of people take common drugs and some get cancer. Running standardised analyses across whole-country records could rank which old drugs deserve a real trial.

Patients now ask AI assistants about their cancer. Test those assistants regularly on real questions, publish the scores, and certify the ones that meet the bar.

Most patients take supplements and rarely tell their oncologist. Ask routinely and check automatically for interactions with chemotherapy and targeted drugs.

Wherever cancer information appears online, a visible marker shows whether it matches what trusted sources say, with a one-click link to the plain-language evidence.

For each question patients actually ask, keep a short, current, sourced answer in plain words, updated as evidence changes and available in the languages people speak.

Frailty is the strongest predictor of who will be harmed by treatment, but it is rarely measured. Software can estimate it automatically from existing records and flag patients who need a closer look.

Most trials copy the same kidney and liver cut-offs, such as creatinine clearance above 60, regardless of how the drug is cleared. Setting each threshold from the drug's own clearance route and organ-impairment pharmacokinetic studies would let patients with mild organ impairment join safely instead of being excluded.

Trials usually take only patients who are up and about most of the day. Those who spend more time resting, a common group in real clinics, are excluded, so nobody knows how to treat them. A dedicated group in each trial would answer that.

Instead of excluding sicker patients entirely, trials would run a side group for them, receiving the new drug with closer monitoring, so we learn how it behaves in the people who will actually get it.

Every patient finishing treatment should get a clear document listing what they had, what to watch for, and when to be checked. Software can write it from the record so it actually happens.

Children treated for cancer are followed for decades in a study that has changed how they are treated. Adults have nothing similar. Build it.

Patients moving between hospitals often carry paper folders or nothing. A standard electronic summary of diagnosis, treatments, and doses that any system can read would stop repeated tests and dangerous gaps.

We know surprisingly little about what happens to cancer survivors twenty years on. Linking their treatment records to later health records would show which treatments cause which problems and who needs watching.

Nobody keeps track of how differently the same drug is approved and dosed around the world. A public scoreboard would make the differences visible and push regulators to converge.

A drug approved in the US may take five years to reach a patient in Poland or never reach Nigeria. A live public tracker would show exactly where and why it is stuck.

Preparing a regulatory filing requires expensive specialist software and consultants. Free, open tools would let universities and small generic firms file in more countries.

Build the shared software, legal templates and data standards that let any hospital in the world join a cancer trial in weeks instead of years, the way the internet let any computer join the network.

Instead of opening a trial at fifty hospitals and waiting for patients, keep a network of pre-vetted clinics ready and switch a trial on where a matching patient is found.

Rare cancer patients are already tracked in registries. Offering randomisation inside the registry makes trials far cheaper and lets almost anyone take part.

Companies hold thousands of well-characterised drugs that could help rare cancers, but each request takes a year of legal negotiation. One standing agreement would unblock it.

Join the national list of who got cancer to the genetic profile of each tumour, so we can see for the whole population which mutations matter and which drugs work for them.

Millions of people have had weight-loss surgery or now take weight-loss drugs. Linking those records to cancer registries would show, cancer by cancer, how much reversing obesity prevents, for almost no cost.

Air pollution causes lung cancer in people who never smoked. Clean air zones and coal phase-outs should be tracked against never-smoker lung cancer rates.

Precancer is treated by cutting away part of the cervix, which raises pregnancy risks. A vaccine that makes the immune system clear the infected cells would avoid surgery.

Dormant cancer cells hide in bone marrow. A lab-built model of that hiding place would let us watch them sleep and wake, and test drugs on them.

To prove a leftover-cancer test works you need blood taken years before relapse. Collecting and freezing yearly samples now makes every future test testable.

After surgery or curative treatment, everyone gets regular blood tests for leftover cancer DNA, and the pooled results power forecasts of who will relapse and trials of acting early.

Generic imatinib and abiraterone can cost patients hundreds of dollars a month through insurance yet be sold at cost plus a fixed markup by transparent pharmacies; making that the default channel for oral cancer generics would save patients and plans money.

India's 2012 compulsory licence on sorafenib cut its price by about 97% and its 2019 cap on trade margins lowered the shelf price of 42 cancer drugs; voluntary licences through a patent pool would achieve the same without a fight.

There are far too few genetic counsellors. A validated chatbot can do the standard pre-test education, leaving people with complex needs for humans.

Doctors rarely take a full family history, so eligibility for genetic testing goes undetected. An app that gathers the history from the patient, feeds it into the record and checks it against NCCN or NICE criteria would identify several-fold more eligible people and, the proposal estimates, roughly double the number tested.

Every staging scan contains a precise measure of muscle mass that nobody looks at. Software could report it automatically and flag patients heading for wasting.

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.

Precancer is treated by cutting away part of the cervix, which raises pregnancy risks. A vaccine that makes the immune system clear the infected cells would avoid surgery.

Scans find unexpected lumps in the adrenal, thyroid, pancreas and lung. Nobody knows how many matter. A national cohort following them for ten years would tell us whom to watch.

Investors will not back a single university drug because most fail. A fund that finances fifty of them at once in exchange for a small slice of each one's future royalties spreads the risk enough to attract capital.