Wrong doses
Most drug doses were chosen as the highest a person can tolerate, which is often more than they need.
Oncology inherited from cytotoxic chemotherapy the assumption that more drug is better, so phase 1 trials escalate to the maximum tolerated dose over a few weeks in a few dozen patients and that dose becomes the label. Targeted agents, antibodies, ADCs and immunotherapies saturate their targets far below toxicity, and chronic dosing exposes patients to months of grade 2 toxicity that the 28-day dose-limiting-toxicity window never measured. The consequences are dose reductions and discontinuations in a third or more of patients on many oral targeted agents, avoidable cost, and occasionally drugs that fail because nobody can stay on them. The FDA's Project Optimus (2021) and its 2024 guidance now require randomised comparison of at least two doses before registration, and trials such as PERSEPHONE show that duration can be halved without loss of efficacy. Dosimetry-based individualisation is the parallel challenge in radioligand therapy.
- Phase 1 designs optimise for speed and identify the maximum tolerated dose rather than the optimal biological dose.
- Dose-limiting toxicity is assessed in one cycle, missing cumulative and chronic toxicities.
- Sponsors have no commercial incentive to study lower doses or shorter durations after approval.
- Regulators historically accepted a single dose in registration trials.
- Pharmacokinetic variability between patients is large but dosing is rarely individualised.
- FDA Project Optimus and the 2024 final guidance on dose optimisation require sponsors to compare multiple doses before registration.
- PERSEPHONE, SOLD and Short-HER established that shorter adjuvant trastuzumab is non-inferior for many patients.
- The CodeBreaK 100 randomised 960 mg vs 240 mg sotorasib comparison, required by the FDA, is the model for post-approval dose re-evaluation.
- Dosimetry-personalised PRRT and radioligand therapy trials replace fixed activity with absorbed-dose-based dosing.
- Intermittent and stop-and-restart dosing is being tested for nirogacestat in desmoid tumours and for venetoclax in AML.
- Finaldose and academic pharmacometrics groups apply model-informed precision dosing to oncology drugs.
A drug's label should say whether its dose was chosen by comparing several doses or simply by finding the most patients could tolerate. Doctors could then know how much room there is to reduce the dose safely.
Most people with cancer are over 65, but trials mostly enrol younger, fitter people. Requiring a group of older patients, assessed for frailty, in every big trial would show whether the drug works and is safe for those most likely to receive it.
Instead of hitting a tumour with the maximum dose until it stops working, adjust the dose to keep the tumour small and let drug-sensitive cells suppress resistant ones. Test this properly across several cancers.
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.
Women experience more severe side effects from many chemotherapy, targeted and immune drugs than men at identical doses. Trials should analyse drug levels and side effects separately by sex and test whether women need different doses.
Drugs are processed differently by people with different genetic backgrounds, but doses are set mostly in white and East Asian populations. Every new drug should be studied for how it behaves across ancestries, with dosing advice by genotype rather than by race.
Rather than giving the same dose until the cancer grows, measure tumour DNA in blood every few weeks and let a validated algorithm raise, lower, pause or switch drugs to keep the cancer suppressed for longer.
A trial in India found that adding a very small dose of an immunotherapy drug, about a twentieth of the usual amount, to chemotherapy improved survival in head and neck cancer. If confirmed, this could make immunotherapy affordable for millions.
Use tumour DNA in the blood as the signal to pause and restart a lung cancer pill, keeping the tumour in check while slowing the rise of resistant cells.
Trials report side effects as a table of percentages that hides how long they lasted, how bad they felt and whether people stopped treatment. Tolerability should be measured with defined endpoints and a decision rule, like efficacy.
Patients report symptoms on their phone each week; a set of rules turns severe or worsening symptoms into a dose hold or reduction before the next clinic visit. This could keep people on treatment longer and feeling better.
Antibody-drug conjugates carry potent chemotherapy into tumours but also cause lung, eye and nerve damage that depends on total exposure. Trials comparing lower doses, longer intervals and payload caps could keep the benefit while cutting these harms.
Chemotherapy doses are calculated from height and weight, a formula from the 1950s. Doses based on actual muscle mass may cause fewer severe side-effects.
The dose for a frail 80-year-old is guessed from what fit 55-year-olds tolerated. Running dose-finding in older patients directly, using frailty assessment, would give doses they can actually take.
Rather than hitting the tumour as hard as possible, adaptive therapy uses just enough drug to keep it in check, letting drug-sensitive cells suppress resistant ones. A small prostate cancer pilot was promising; randomised trials are needed.
Instead of picking the dose by how much patients can tolerate, measure drug levels in blood and relate them to both benefit and harm across patients, then choose the dose that hits the sweet spot.
Immunotherapy antibodies stay active in the body for weeks, yet are often given every two or three weeks. After a few months, spacing doses out to every two or three months might work as well, with fewer hospital visits and much lower cost.
Companies lose money when they prove a shorter course works, so they never test it. Give them a modest reward, such as extra months of exclusivity, when they do.
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.
Combination trials usually keep one drug at full dose and push the other as high as patients can bear. Testing a grid of dose pairs would find combinations that work at lower, safer doses.
Cells that receive only a small amount of a drug survive and adapt. Measuring where inside a tumour the drug actually reaches would show where resistance is being bred.
Simulating thousands of virtual patients on a computer can suggest which dose and schedule to test, so fewer real patients receive doses that are too high or too low.
Dosing advice for people with weak kidneys or liver is often missing at approval and added years later, if ever. Requiring those studies before approval would protect a large group of real-world patients from day one.
The FDA now asks companies to find the right dose of a cancer drug before approval. If every regulator asked the same way, companies would do it once and doses would match worldwide.
Blood levels of oral cancer pills vary several-fold between people on the same dose, so some are under-treated and others poisoned. Checking levels and adjusting the dose, as is routine for some antibiotics, could fix both.
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.
People differ several-fold in how they absorb and clear cancer pills. Measure blood levels and adjust each person's dose, as is routine for some antibiotics and transplant drugs.
Health insurers and national health systems have every reason to find out whether half the dose or half the duration of a costly drug works as well. They would fund those trials directly and keep the savings.
Many approved cancer drugs probably work just as well at half the dose, which would halve their side effects and cost. Companies will not test this, so payers and public funders should.
Many cancer drugs work as well at lower doses or given less often, but companies have no reason to prove it. Public funders should run those trials.
Cancer drugs are usually tested at the highest dose patients can stand, and that dose sticks for life. Comparing two or more doses head-to-head before the big trial would find doses that work as well with fewer side effects.
Many cancer drugs are approved at the highest dose patients could stand, not the best dose. Run trials that test lower doses on approved drugs and measure how patients feel.
Track how much of each cancer drug patients actually receive and how often they need to reduce it, and use that to change the official dose when the label is too high.
The dose needed to shrink a tumour may be higher than the dose needed to keep it from growing back. Trials that lower the dose once a response is achieved could reduce long-term side effects without losing control.
Most trials use the same blood-test cut-offs for kidney and liver function regardless of how the drug is cleared from the body. Setting the cut-off from the drug's own pharmacology would let many more people join safely.
Immunotherapy is given as one flat dose regardless of body size, which means smaller patients get more than they need. Dosing by weight would save a fifth of the drug at no cost to patients.
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.
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.
A cheap gene test before starting common chemotherapy identifies people at high risk of severe or fatal side effects so their dose can be lowered. Europe recommends it; many places still do not do it.
Many antibody drugs moved from doses based on body weight to one fixed dose for everyone, which is convenient but means lighter patients get relatively more drug. Trials comparing the two, plus rounding doses to standard vial sizes, could keep effectiveness while cutting cost and waste.
Giving a targeted drug in pulses rather than continuously might slow the emergence of resistant cells and reduce side effects. Early results are mixed, so this needs careful trials with clear rules for when to try it.
Frail older patients are often given full doses and then have them cut after bad side effects, or are given nothing. Trials should test starting at a lower dose from the beginning.
Some cancer pills are absorbed several times better with food, but the label says take them fasting at a high dose. Taking a quarter of the dose with breakfast can give the same drug levels at a quarter of the price.
Low doses of anti-blood-vessel drugs briefly make tumour vessels work better, which helps immune cells and other drugs get in. Scans can find that window for each patient.
Patients with melanoma that has spread to palpable lymph nodes should now be offered immunotherapy before rather than only after surgery: two cycles of low-dose ipilimumab with nivolumab, then surgery, with the pathology result deciding whether any more treatment is needed. Most patients respond well and are spared a year of adjuvant therapy. Serious side effects are more common than with nivolumab alone, mostly endocrine, and the approach requires close coordination between oncologists, surgeons and pathologists.
Ribociclib is a second adjuvant CDK4/6 option, and the only one with data in node-negative stage II disease. Roughly 3 in 100 patients avoid a relapse or death at three years, so the decision depends heavily on individual risk, tolerance of a three-year oral drug, and cost. Whether the benefit persists after treatment ends, as it did with abemaciclib, needs longer follow-up.
Patients with KRAS G12C lung cancer that has progressed after chemo-immunotherapy can take an oral KRAS inhibitor instead of docetaxel and gain a somewhat longer time to progression with fewer severe side effects, but should understand that most tumours become resistant within a year and that survival is not improved. KRAS G12C testing is worthwhile, but first-generation inhibitors are a step rather than a cure; combinations and next-generation inhibitors are the active research fronts.
Patients with metastatic colorectal cancer carrying a KRAS G12C mutation (about 3-4% of cases) who have exhausted standard chemotherapy now have a targeted option that works far better than trifluridine-tipiracil or regorafenib. The higher sotorasib dose is clearly superior, and the EGFR antibody is essential because KRAS inhibition alone barely works in bowel cancer. Responses are still modest and short-lived compared with EGFR or ALK inhibitors in lung cancer.
Patients with small-cell lung cancer that has relapsed after chemotherapy now have a drug that works far better than topotecan or lurbinectedin, and it is the first T-cell engager approved for a solid tumour. Treatment requires inpatient monitoring for the first doses because of cytokine release syndrome, which most centres now manage on a short-stay basis. It does not yet apply to first-line treatment, where trials are ongoing.
Elranatamab confirmed that BCMA bispecifics are a class, not a one-off, and its protocol-built dose reduction after response set a precedent for lowering the immunosuppressive burden of T-cell engagers. Patients now have two approved BCMA bispecifics and one against GPRC5D (talquetamab). Choosing between them, and sequencing them with CAR-T, remains guided by availability and toxicity profile rather than head-to-head data.
QuANTUM-First gave FLT3-ITD AML patients a second front-line targeted option and showed that continuing a FLT3 inhibitor as long-term maintenance, including after transplant, pays off. Quizartinib was approved for this indication in 2023. Head-to-head data against midostaurin are lacking, and the design leaves open how much of the benefit came from maintenance.
Teclistamab showed that an off-the-shelf bispecific can approach CAR-T-like response rates in late myeloma, giving patients who cannot wait for or access cell therapy a real option. It also exposed the price: prolonged T-cell engagement causes profound immunosuppression, so infection prophylaxis and immunoglobulin replacement are now routine. Less frequent dosing after response is being adopted to reduce this burden.
Patients with newly diagnosed advanced clear-cell kidney cancer should receive an immunotherapy-based combination; lenvatinib plus pembrolizumab gives the highest response rate and longest PFS of the available options, at the cost of more side effects requiring dose adjustment. Sunitinib alone is no longer an appropriate standard. Choosing among the combinations depends on risk group, symptoms, comorbidity and the value placed on treatment-free survival, which favours nivolumab plus ipilimumab in intermediate and poor risk.
The dose on the label is often not the best dose for patients; it is the highest one that was tolerable for a few weeks. Project Optimus means new cancer drugs should arrive with evidence on dose, and it gives clinicians licence to consider dose reduction for toxicity. For older drugs, the evidence gap persists.
Older patients are more likely to be harmed by standard-dose chemotherapy, and a structured assessment of function, cognition, nutrition and social support lets oncologists adjust treatment safely. Starting lower does not appear to shorten life. Most older patients still do not get such an assessment.
Pages like this
not linked directly; found by shared links- BottleneckToxicity and quality of life are undervalued
Shares Define tolerability endpoints as rigorously as efficacy endpoints, Dose adjustments driven by patients' own symptom reports, tested against clinician judgement, Real-world dose intensity and toxicity monitoring to revise labelled doses, Analyse and dose by sex: women get more toxicity from many cancer drugs at the same dose.
- BottleneckOlder and multimorbid patients are excluded and undertreated
Shares A mandatory over-70s cohort with geriatric assessment in every pivotal trial, Kidney and liver impairment dosing studies completed before approval, not years after, Shorter venetoclax courses in unfit AML, Dose-finding in older and frail patients, not extrapolation from fit ones.
- BottleneckPrices and value
Shares Publicly funded trials of lower and less frequent doses of expensive cancer drugs, Restore weight-based dosing and vial sharing for immunotherapy in the label, Use food effects to cut the dose and cost of oral drugs that absorb better with meals, Test flat versus weight-based dosing of antibodies, and use dose banding to cut waste.
- BottleneckAcquired resistance to every therapy
Shares Evolution-guided 'adaptive therapy' dosing tested in randomised phase 2 trials, Find the parts of a tumour the drug never reaches, ctDNA-guided dose holidays for lung cancer targeted therapy, Test intermittent dosing of targeted drugs to delay resistance, with honest priors.
- TermDrug resistance (primary and acquired)
Shares Evolution-guided 'adaptive therapy' dosing tested in randomised phase 2 trials, Find the parts of a tumour the drug never reaches, Test intermittent dosing of targeted drugs to delay resistance, with honest priors, Autonomous closed-loop adaptive therapy driven by blood tests and evolutionary models.
- ProductOsimertinib
Shares Publicly funded trials of lower and less frequent doses of expensive cancer drugs, Extra exclusivity for sponsors who run treatment-duration and de-escalation trials, ctDNA-guided dose holidays for lung cancer targeted therapy, Autonomous closed-loop adaptive therapy driven by blood tests and evolutionary models.
- ProductImatinib
Shares Measure blood levels of oral targeted drugs and adjust doses to a target range, Model-informed individual dosing with drug-level monitoring for every oral cancer drug, Publicly funded dose-reduction trials of expensive approved drugs, Reduce the dose once the cancer responds: response-adapted de-escalation trials.
- TermObjective response rate (ORR)
Shares CodeBreaK 300: sotorasib plus panitumumab in chemotherapy-refractory KRAS G12C colorectal cancer, CLEAR: lenvatinib plus pembrolizumab versus sunitinib as first treatment for advanced kidney cancer, CodeBreaK 200: sotorasib versus docetaxel in KRAS G12C-mutated lung cancer, a modest win for the first KRAS drug, MagnetisMM-3: elranatamab, a second BCMA bispecific, with a switch to fortnightly dosing after response.