Loading
Oncology learns more from what did not work than from what did, but failures vanish from pipelines and press releases. This room keeps them: drugs stopped, targets abandoned, approvals withdrawn, each with the lesson it left. Nothing here is a verdict on the idea; several of these targets later worked with a different weapon.
Why failures are shown: they are how the field learns. Every drug that did not work sharpened the next one: a target that survived a bad molecule, a dose that was wrong, a trial design that misled. Several targets in this room later worked with a different weapon, and the lesson attached to each exhibit is what the successful drugs were built on. Nothing here is a statement about any patient's outlook; it is the field's record of what it tried, kept so it does not have to be learned twice.
Encouraging early data, often single-arm response rates on a checkpoint-inhibitor backbone, that phase 3 could not reproduce.
Bempegaldesleukin was a re-engineered interleukin-2 meant to be a safer version of a famous old immunotherapy. It added nothing to nivolumab in three phase 3 trials.
Lesson: single-arm combination response rates in melanoma are unreliable because nivolumab alone already produces them; biological rationale (Treg sparing) did not translate.
An enzyme blocker meant to stop tumours starving T cells of tryptophan. Its 2018 phase 3 failure ended an entire class overnight.
Lesson: pharmacodynamic proof of target engagement in the tumour should precede phase 3; a large uncontrolled response rate on a pembrolizumab backbone in melanoma is not evidence of added benefit.
An immune-brake blocker that looked excellent in a phase 2 lung cancer trial and then failed every phase 3.
Lesson: a small randomised phase 2 with a surrogate endpoint in a selected population can mislead; redundancy between checkpoints means blocking a second one does not necessarily add to PD-1/PD-L1 blockade.
The target or concept survived; the specific drug did not do what was claimed, or carried the wrong payload.
ADU-S100 was the first STING agonist in the clinic. Injected directly into tumours, it produced almost no responses, alone or with checkpoint blockade.
Lesson: intratumoural delivery to one lesion rarely produces systemic immunity in humans as it does in mice; pharmacology (rapid clearance, dosing) matters as much as the target.
Billed as the first PARP inhibitor for triple-negative breast cancer, it failed its phase 3 in 2011. It turned out not to inhibit PARP at all.
Lesson: mechanism claims should be verified before a class label is attached; the true PARP inhibitors (olaparib, talazoparib) later succeeded in BRCA-mutant breast cancer.
Rovalpituzumab tesirine (Rova-T) was the first drug against DLL3 in small-cell lung cancer. AbbVie bought it for $5.8B and abandoned it after two failed phase 3 trials. The target later worked with a different weapon.
Lesson: target validation and modality are separable. DLL3 was the right address; a PBD payload with a narrow window in a frail population was the wrong weapon.
Tusamitamab ravtansine was Sanofi's ADC against the classic CEA tumour marker, stopped for futility in lung cancer in 2023.
Lesson: a tubulin-inhibitor payload at DAR ~4 with a non-permeable release mechanism could not match the TOP1-payload ADC bar set in lung cancer; CEACAM5 itself remains under study with T-cell engagers.
The mechanism looked sound in cells and mice but did not translate to patients.
Activity was real but toxicity, or a safer competitor, made it unusable.
The first 'don't eat me' signal blocker. Gilead paid $4.9B for it; it was stopped in 2024 after trials showed more deaths, not fewer.
Lesson: a ubiquitous target (CD47 is on every red cell) plus an immunosuppressed population is a narrow window; single-arm response rates in TP53-mutant disease were not predictive of survival.
A HER2 ADC with a DNA-alkylating payload that beat chemotherapy in a phase 3 trial yet never reached the market, because eye and lung toxicity and a stronger rival arrived first.
Lesson: statistical significance is not enough when a competitor redefines the standard; payload toxicity profile decides whether an ADC survives.
The chemotherapy partner or the companion diagnostic, not the drug, determined the result.
Accelerated approvals whose confirmatory trials disappointed or showed harm.
Negative or withdrawn items without a single dominant lesson.
A drug aimed at the tissue around the tumour delayed progression on scans by two and a half months, and the men who took it lived, if anything, slightly less long.
Abarelix was the first GnRH antagonist for prostate cancer, approved in the United States in 2003 for men who could not take agonists, but allergic reactions restricted it and it was withdrawn from the US market in 2005; it remained available in Germany.
When chemotherapy is given first and clears the lymph nodes, can a surgeon trust the sentinel node to prove it? Z1071 found the sentinel node missed residual cancer about one time in eight, which is too often, and that finding forced the technique to be redesigned.
ACT IV was a double-blind phase 3 test of rindopepimut, a vaccine against the EGFRvIII mutant protein, in 745 patients with newly diagnosed glioblastoma. It showed no benefit over a control vaccine, and because both arms beat historical expectations it taught the field how misleading historical-control comparisons can be.
ADIUVO asked whether people whose low-grade adrenal cancer had been fully removed should take the toxic drug mitotane for two years to prevent recurrence; recurrence and survival were no different from watching and waiting, so observation with regular scans is the standard for this group.
The second attempt with the same drug, this time after chemotherapy had already failed once. It also found nothing, including in the men expected to benefit most.
The phase 3 that was meant to bring the tumour-targeted doxorubicin aldoxorubicin to market: it did not meet its primary endpoint of progression-free survival.
ALEXANDRA asked whether a year of the immunotherapy atezolizumab added to chemotherapy after surgery stops triple-negative breast cancer coming back. It did not: relapses were no fewer, side effects were more, and the trial was stopped early for futility.
Algenpantucel-L was a whole-cell pancreatic cancer vaccine given after surgery; the phase 3 IMPRESS trial showed no survival benefit.
The North American version of the same question, with the same answer: cetuximab after surgery adds toxicity and no benefit.
An intensive infusion chemotherapy schedule, widely believed to be better for aggressive lymphoma, was compared head to head with the standard and was not better, only harder to tolerate.
Treating the blood test result before the cancer shows up on a scan did not work. It delayed relapse by about four months and caused severe blood toxicity in three-quarters of patients.
Complementary approaches used alongside treatment can help with symptoms. Choosing an alternative therapy instead of surgery, chemotherapy, radiotherapy or hormone therapy is a different decision: in a national US database, people who did so were two and a half times as likely to die within the study period, and more than five times as likely with breast cancer.
The randomised test of an off-the-shelf KRAS vaccine after pancreatic surgery. It missed its primary goal in June 2026.
Anetumab ravtansine was the first antibody-drug conjugate tested in a randomised trial in mesothelioma, aimed at the mesothelin protein that nearly all mesotheliomas carry; it was no better than vinorelbine chemotherapy as second-line treatment, and is now being tried with pembrolizumab.
The first randomised test of an antibody-drug conjugate in mesothelioma: no better than vinorelbine, which is why no ADC is approved for this cancer.
Antineoplastons are peptide fractions first isolated from urine and given at one private clinic in Texas for nearly fifty years, mostly to children with brain tumours. Despite dozens of registered trials, none has been published in full, no independent group has confirmed benefit, and the treatment is expensive and causes serious salt imbalance.
APACT tested whether adding nab-paclitaxel to gemcitabine after pancreatic surgery delays relapse; by the independent radiologists' reading it did not (19.4 against 18.8 months), although patients on the combination lived about four months longer at five years, so the regimen is not a formal adjuvant standard.
The only trial ever to pick men for prostate cancer treatment by a splice variant of the androgen receptor. It could not recruit, because the men who qualified were too ill to stay in it.
Three trials asked the same question and none was big enough alone. Pooling them, planned in advance so nobody could cherry-pick, showed that waiting for the PSA to rise is as good as treating everyone.
In healthy older people, daily low-dose aspirin did not extend disability-free life and, unexpectedly, was linked with more cancer deaths. It changed guidelines against routine aspirin in the elderly.
The other endothelin blocker. Adding it to chemotherapy for advanced prostate cancer changed nothing, and the trial was stopped for futility.
AVAglio tested adding the anti-blood-vessel drug bevacizumab to standard radiotherapy and chemotherapy for newly diagnosed glioblastoma; the tumours took longer to grow on scans, but people did not live any longer, so bevacizumab is not used up front for this cancer.
Adding the metabolism drug devimistat to FOLFIRINOX did not help people with metastatic pancreatic cancer live longer.
Black salve is a corrosive paste containing bloodroot and zinc chloride sold online to 'draw out' skin cancers. It burns whatever it touches, leaves disfiguring scars, does not reliably remove the cancer, and has let melanomas spread while people believed they were cured.
BRE12-158 sequenced the cancer left behind after pre-operative chemotherapy and matched a targeted drug to it. Matched therapy was no better than the doctor's usual choice, which increasingly meant capecitabine, and blood-borne tumour DNA after surgery predicted who relapsed.
The first big immunotherapy trial in prostate cancer. It missed by a hair, and the way it missed, harm early and benefit late, was the clue nobody managed to cash in.
Adding the anti-angiogenic antibody bevacizumab to chemotherapy delayed progression and shrank more tumours, but men did not live longer and four times as many died of the treatment.
CALLA tested adding the PD-L1 antibody durvalumab to chemoradiotherapy for locally advanced cervical cancer and found no significant benefit, in contrast to the later positive KEYNOTE-A18 trial of pembrolizumab in a higher-risk population.
CanStem111P is the largest first-line pancreatic cancer trial ever run, 1,134 patients, and it found that the stemness inhibitor napabucasin added nothing to nab-paclitaxel and gemcitabine; the trial was stopped for futility and its control arm is now the best description of what that chemotherapy achieves.
Carotuximab was an antibody against endoglin, a protein on growing blood vessels and on angiosarcoma cells. Added to pazopanib in the phase 3 TAPPAS trial it did not help, and the trial was stopped in 2019.
CD47 is the 'don't eat me' signal: it binds SIRP-alpha on macrophages to stop them engulfing the cell, and over 90% of AML blasts and large B-cell lymphoma cells display it. Blocking it should let macrophages eat tumour cells, but red cells carry CD47 too, so anaemia is built in, and the lead antibody magrolimab was dropped after failed trials.
Biotech that ran one of the early GPNMB ADCs in triple-negative breast cancer, which failed, and has since pivoted to mast-cell (KIT) antibodies for allergic disease.
The trial that showed a PD-1 antibody can fail in the first line if the PD-L1 threshold is set too low, and the reason the 50 percent cut-off exists.
CheckMate 143 was the first large randomised test of an immune checkpoint drug in glioblastoma. When the tumour came back, nivolumab kept patients alive no longer than bevacizumab.
CheckMate 498 asked whether nivolumab could replace temozolomide in the glioblastoma patients who gain least from it. Patients given nivolumab lived a shorter time.
CheckMate 548 added nivolumab to standard radiotherapy and temozolomide in newly diagnosed glioblastoma and found no benefit. With CheckMate 143 and 498 it makes three large negative trials: the immunotherapy that transformed melanoma and lung cancer did nothing in glioblastoma.
CheckMate 915 asked whether adding a low dose of ipilimumab to nivolumab after melanoma surgery would prevent more recurrences; it did not, and added toxicity, so single-agent PD-1 blockade remains the adjuvant standard.
Giving the heated abdominal wash as insurance, before any peritoneal disease has appeared, did not prevent it.
A drug that made bone scans look dramatically better did not make men live longer. It is the best illustration in prostate cancer that a bone scan is not the disease.
CONKO-005 was the first adjuvant trial to add a targeted drug to chemotherapy after pancreatic surgery; six months of erlotinib with gemcitabine made no difference to the time before relapse (11.4 months in both arms) or to survival.
Copanlisib is an intravenous PI3K inhibitor for relapsed follicular lymphoma, approved in 2017 and withdrawn in 2023 when its confirmatory trial failed.
Several drugs have been put on the scalp to stop chemotherapy hair loss before it starts: minoxidil lotion, vitamin D analogues, and others. The trials were done and they did not work. Scalp cooling remains the only method cleared by a regulator to prevent it.
CROSSFIRE is the only randomised comparison of the two local treatments offered after chemotherapy for pancreatic cancer that cannot be removed, MR-guided stereotactic radiotherapy and irreversible electroporation; survival was 16.1 against 12.5 months with no significant difference, the trial stopped early for futility, and neither has been shown to beat chemotherapy alone.
A drug designed to switch off a protein that helps cancer cells survive chemotherapy. Two large trials found it added nothing.
De-ESCALaTE, the European counterpart of RTOG 1016, found that cetuximab gave no less toxicity than cisplatin and worse tumour control in HPV-positive throat cancer.
Depatuxizumab mafodotin carried a cell-killing payload to glioblastomas with extra copies of the EGFR gene. It caused serious eye problems and, in the phase 3 INTELLANCE-1 trial, did not help patients live longer, so development stopped in 2019.
Devimistat tried to starve pancreatic cancer cells of their energy supply alongside chemotherapy; the phase 3 AVENGER 500 trial found no survival gain.
ECLIPSE was the randomised test of the Johns Hopkins pancreatic cancer vaccine strategy, a whole-cell GVAX vaccine followed by a Listeria vector carrying mesothelin; after promising early results it did not beat chemotherapy, with survival of 3.7 months against 4.6, and the programme ended.
EA1131 asked whether platinum chemotherapy after surgery beats capecitabine for triple-negative breast cancer that survived pre-operative chemotherapy. It was stopped early because platinum was not going to prove better and was more toxic, so capecitabine stayed the standard.
A cleaner version of abiraterone held the cancer back on scans for five extra months and did not make anyone live longer. It is the clearest warning in prostate cancer that scans are not survival.
A drug blocking the endothelin signal that prostate cancer uses in bone did not extend life, and it caused swelling, headaches and heart failure.
EORTC 22033 asked whether temozolomide could replace radiotherapy as the first treatment for high-risk low-grade glioma and found no difference in progression-free survival, with the molecular subtype mattering more than the choice of treatment.
Tried to show that six months of hormone therapy with radiotherapy is enough for locally advanced prostate cancer. It is not: three years works better.
Combining a bone-seeking radioactive drug with abiraterone did not help, and it nearly tripled the rate of broken bones. The trial was unblinded early and the combination is now contraindicated.
The only randomised trial of a ketogenic diet in brain tumours found no benefit. Patients could follow the diet and their ketone levels rose, but progression came just as quickly.
Eryaspase packed a leukaemia enzyme into red blood cells to starve pancreatic cancer of an amino acid; the phase 3 TRYbeCA-1 trial missed its survival goal.
ESSA Pharma developed masofaniten, an AR N-terminal-domain inhibitor whose phase 2 was stopped for futility in 2024-25.
Essiac tea, Hoxsey tonic and similar herbal mixtures have been sold as cancer cures for a century. Laboratory tests and reviews by the NCI and Canadian regulators found no anticancer effect, and the clinics that sell them have never produced a controlled trial.
Evofosfamide was the most advanced hypoxia-activated drug, designed to kill the oxygen-starved tumour cells radiotherapy and chemotherapy miss, but it failed both of its phase 3 trials in 2015.
Dog dewormers and anti-parasite drugs are promoted on social media as hidden cancer cures on the strength of cell-culture experiments and anecdotes. No clinical trial shows benefit in people, liver damage has been reported, and drug repurposing is real but works through trials, not forums.
Fianlimab is Regeneron's LAG-3 blocking antibody, paired with the PD-1 blocker cemiplimab. A 60% phase 1 response rate in untreated melanoma prompted phase 3 trials, but the metastatic trial against pembrolizumab missed its progression endpoint in 2026.
FIGHT-302 set out to test whether the FGFR2 inhibitor pemigatinib should be used before chemotherapy in bile duct cancers driven by an FGFR2 fusion; the trial closed early after the standard of care changed, and its 2026 report showed pemigatinib delayed progression compared with chemotherapy, with similar overall survival.
Cabazitaxel is no better than docetaxel as the first chemotherapy for castration-resistant prostate cancer. It is a different set of side effects, not a better drug.
An experimental prostate cancer tablet meant to work in men whose cancer makes a broken form of the androgen receptor. The trial designed to prove it collapsed.
GATSBY tested whether the antibody-drug conjugate trastuzumab emtansine, which works in HER2-positive breast cancer, would also beat chemotherapy in HER2-positive stomach cancer after first treatment; it did not, and the drug was never approved for this cancer.
This Spanish and Latin American trial gave eight extra cycles of capecitabine after standard chemotherapy for early triple-negative breast cancer and did not significantly reduce relapses. Only tumours without the basal-like pattern seemed to benefit, a finding that still needs confirmation.
GeparDouze, the largest neoadjuvant immunotherapy trial in triple-negative breast cancer, found that adding atezolizumab to chemotherapy before and after surgery did not significantly reduce relapses (a 3.3-point gain at four years that missed statistical significance), unlike pembrolizumab in KEYNOTE-522.
The Gerson regimen prescribes hourly juices, a strict low-salt vegetarian diet, supplements and several coffee enemas a day, and is sold at clinics in Mexico. It has never shown benefit in any controlled study, and coffee enemas have caused fatal electrolyte disturbances and infections.
The first trial to test chemotherapy at the start of hormone therapy for metastatic prostate cancer found no benefit. It was too small, and CHAARTED and STAMPEDE later showed it had missed a real effect.
Trying to spare people radiotherapy by using a clear scan after two rounds of chemotherapy cost them about seven in a hundred in disease control, so the radiotherapy stayed.
GI-6301 was a vaccine made from yeast carrying brachyury, the protein that chordomas depend on. Given with radiotherapy in a randomised trial it did not improve tumour responses.
A newer antibody against the same target as rituximab was tested in first-line treatment for the commonest aggressive lymphoma and did not work better, while causing more side effects.
Neoantigen-vaccine pioneer (GRANITE, SLATE) that ran out of money and filed for bankruptcy in October 2024; a cautionary tale for the personalised-vaccine field.
HALO-301 was the phase 3 test of breaking down the dense stroma of pancreatic tumours with a hyaluronidase enzyme so chemotherapy could reach them; response rates went up but survival did not move, and the stroma-busting idea in this form was abandoned.
HERBY asked whether adding the anti-blood-vessel antibody bevacizumab to radiotherapy and temozolomide would help children with high-grade brain tumours; it did not delay the tumours returning, so the drug is not part of standard treatment for these children.
Homeopathic remedies are diluted until no molecules of the starting substance remain, so any effect would need new physics. A Cochrane review of trials for chemotherapy and radiotherapy side effects found no convincing evidence that they work; they are harmless as long as they do not replace real treatment.
HORRAD found no overall survival benefit from irradiating the prostate in men presenting with bone metastases, though a possible gain in those with few metastases fed into STAMPEDE's later positive finding.
Hydrazine sulfate was promoted in the 1970s and 80s to reverse cancer weight loss and prolong life. Three large randomised trials sponsored by the National Cancer Institute found no benefit and more side effects, and the compound is a suspected carcinogen.
Adding a checkpoint inhibitor to enzalutamide did not help men with advanced prostate cancer live longer, and the trial explained why: prostate tumours carry very little of what these drugs need.
The trial that tried to make immunotherapy work in ordinary bowel cancer by adding a MEK inhibitor. It failed outright.
The largest vaccine trial after pancreatic cancer surgery found that algenpantucel-L added nothing to standard chemotherapy.
Infigratinib is an FGFR inhibitor pill given accelerated approval in the United States in 2021 for bile duct cancer with an FGFR2 fusion, then withdrawn in 2024 when its confirmatory trial could not enrol; it is now being developed for achondroplasia instead.
INTELLANCE-1 tested an antibody-drug conjugate that targets the EGFR amplification found in about half of glioblastomas, added to standard treatment; it did not help people live longer, and the drug was abandoned.
IPATunity130 tried to confirm a promising early signal for the AKT-blocking tablet ipatasertib with paclitaxel in triple-negative breast cancers carrying PI3K-pathway mutations. It failed: progression and survival were the same with or without the drug, and the ipatasertib programme in breast cancer ended.
This Brazilian trial tested ivermectin against COVID-19 in people with cancer, not against the cancer itself. It was stopped early because an interim analysis found no difference between the drug and placebo.
KEYNOTE-119 tested pembrolizumab on its own against chemotherapy as second- or third-line treatment for metastatic triple-negative breast cancer. It did not lengthen life, even in tumours with high PD-L1, which is why immunotherapy in this disease is given with chemotherapy or an antibody-drug conjugate and only in the first line.
KEYNOTE-412 tested adding pembrolizumab to standard chemoradiotherapy for locally advanced head and neck cancer; the improvement in keeping patients free of disease fell short of statistical significance, so chemoradiotherapy alone remains the standard.
Adding pembrolizumab to enzalutamide did not help. The trial was stopped for futility, and three times as many men had a serious side effect.
Adding pembrolizumab to chemotherapy for advanced prostate cancer did not work, and it added side effects and lung inflammation.
The last attempt to make immunotherapy work in prostate cancer, this time as early as possible. It was stopped for futility and made rashes and serious side effects much commoner.
Laetrile, sold as vitamin B17 or apricot kernel extract, was tested in a large National Cancer Institute study in the 1980s and did nothing against cancer. It releases cyanide in the gut, and people have been poisoned by it.
LAP07 tested whether adding radiotherapy after four months of chemotherapy helps people whose pancreatic cancer has not spread but cannot be removed; it did not lengthen life, although it did keep the tumour in check locally for longer, and adding erlotinib to gemcitabine did not help either.
The anti-stroma antibody pamrevlumab did not help people with locally advanced pancreatic cancer live longer.
The trial that ended twenty years of argument about radiotherapy to the mediastinum after lung cancer surgery: it does not help, and it damages the heart.
The first phase 3 to run inside Lung-MAP asked whether adding a second immunotherapy drug, ipilimumab, to nivolumab helps people with squamous lung cancer. It did not: survival was no better and side effects were worse.
The British trial that tested a cheap, staged alternative to scanning everyone, and found it did not work: more than half the cancers never showed up in the sputum at all.
MAJIC-ET found that ruxolitinib was no better than the usual second-line drugs at controlling platelets in ET, although it eased itching and other symptoms.
A drug designed to block the part of the androgen receptor that resistant variants keep; its phase 2 was stopped for futility and development ended.
Mifepristone, best known as an abortion pill and a treatment for Cushing's syndrome, blocks progesterone receptors that most meningiomas carry. A large phase 3 trial found it did not slow inoperable meningiomas.
Mobocertinib was the first oral drug for EGFR exon 20 insertion lung cancer, approved in 2021 and withdrawn in 2023-24 after its confirmatory trial failed.
The trial that tested whether giving GPs a diagnostic gadget would make their skin cancer referrals better, found that it made them more numerous instead, and is the reason the NHS asks for training and dermoscopy rather than devices in primary care.
An immunotoxin for hairy cell leukaemia that produced lasting remissions in relapsed patients but was withdrawn from sale in 2023 for commercial reasons.
The UK trial that checked whether adding an antibody to first-line chemotherapy for advanced bowel cancer helps people live longer, and found it does not.
Napabucasin, an oral drug marketed as a cancer stemness inhibitor, was tested in 1,134 people with untreated metastatic pancreatic cancer on top of nab-paclitaxel and gemcitabine; survival was identical and the trial was stopped for futility.
NEOLAP asked whether switching to FOLFIRINOX after two months of gemcitabine and nab-paclitaxel converts more inoperable pancreatic cancers to operable ones; about a third to 44 percent were removed with either sequence and survival was similar, so both induction regimens are acceptable.
NeoTRIP added the immunotherapy atezolizumab to platinum chemotherapy before surgery without anthracyclines and did not significantly increase the share of women whose tumour disappeared (49 versus 44 percent); its main endpoint, whether fewer women relapse, is still awaited.
Adding the EGFR antibody around liver surgery shortened life by more than two years. It is the clearest warning in bowel cancer that a drug that helps in advanced disease can harm in the curative setting.
CAR-NK company that shelved its oncology programmes after falling response rates and pivoted to autoimmune disease.
NORPACT-1 asked whether giving FOLFIRINOX chemotherapy before surgery helps people whose pancreatic cancer can be removed straight away; it did not, and patients who went straight to surgery were if anything more likely to be alive at 18 months, so upfront surgery remains the standard for clearly resectable disease.
The largest partial-breast trial found recurrence rates within one percentage point of whole-breast treatment but could not formally prove equivalence, so accelerated partial-breast irradiation is offered to selected lower-risk women rather than everyone.
If chemotherapy given before surgery clears the lymph nodes completely, irradiating those nodes afterwards adds nothing. Fewer than one woman in ten had the cancer come back either way, and the difference between the two groups was within chance.
Olaratumab was approved in 2016 with doxorubicin for soft tissue sarcoma after a small trial suggested it added almost a year of life, but the large confirmatory trial found no benefit and it was withdrawn in 2019, a warning case for accelerated approvals.
Olmutinib was a Korean-developed EGFR pill approved in South Korea in 2016 for lung cancers that had developed the T790M resistance mutation, the same niche as osimertinib; severe skin reactions and osimertinib's success ended its development.
Oprelvekin was the first drug approved, in 1997, to prevent the severe platelet falls that chemotherapy causes, but fluid retention and heart rhythm problems limited it and it was withdrawn in 2011; thrombopoietin agonists took over the problem.
A more selective version of abiraterone that was meant to avoid the need for steroids. It delayed the cancer on scans but did not help men live longer, and was abandoned.
PACIFIC-2 asked whether starting durvalumab during chemoradiotherapy, rather than after it, would help people with stage III lung cancer; it did not, so the sequence established by the original PACIFIC trial stands.
Pamrevlumab targeted the scar-like tissue around pancreatic tumours; its phase 3 LAPIS trial in locally advanced disease did not lengthen survival.
An HDAC inhibitor for relapsed myeloma approved in 2015 and withdrawn in 2021 after its confirmatory trial was never completed.
Pegilodecakin, a long-acting interleukin-10 meant to expand tumour-killing T cells, added nothing to FOLFOX in the phase 3 SEQUOIA trial of second-line pancreatic cancer, and Eli Lilly dropped it.
PEGPH20 was designed to break down the hyaluronan that stiffens pancreatic tumours and let chemotherapy in; in the phase 3 HALO-301 trial it raised response rates but not survival, and Halozyme stopped developing it in 2019.
PENELOPE-B tested whether a year of the CDK4/6 inhibitor palbociclib after chemotherapy and surgery would prevent recurrence in women at high risk; it did not, one of two trials to show that palbociclib does not work in early breast cancer even though later drugs in the class do.
Irinotecan works in advanced bowel cancer and does nothing after surgery. The trial's tumour bank turned out to be worth more than its result.
Adding the EGFR antibody cetuximab to chemotherapy after surgery did not help, even in patients whose tumours had no KRAS mutation.
Pevonedistat was a first-in-class drug that jammed part of the cell's protein-disposal system. Promising with azacitidine in a mid-stage trial for higher-risk myelodysplastic syndromes, it failed to beat azacitidine alone in the phase 3 PANTHER trial in 2021.
For the thicker, lump-like basal cell carcinoma, light treatment clears about as many at three months as surgery does, but over five years three times as many come back. What it buys is a better-looking result, and that is the whole of the trade.
Pixantrone is an anthracycline-like drug engineered to spare the heart. The European Union authorised it in 2012 as a single agent for adults with aggressive B-cell lymphoma that had relapsed several times; the authorisation expired in June 2024 when the company did not renew it.
The American screening trial that found no benefit, and the reason is the trial itself: half the men in the group that was supposed to go unscreened were getting PSA tests from their own doctors.
Poziotinib was a tablet designed to fit the awkward shape of HER2 and EGFR exon 20 mutations in lung cancer. It shrank tumours in some patients but caused severe rash and diarrhoea, and the FDA declined to approve it in 2022.
PREOPANC-2 asked whether the strongest chemotherapy given entirely before surgery beats the Dutch standard of gemcitabine chemoradiotherapy before and gemcitabine after; survival was the same, about 22 months in both arms, so both remain reasonable ways to treat first, and the question of surgery first against treatment first moves to Alliance A021806 and PREOPANC-3.
Washing the abdomen with heated oxaliplatin after complete tumour removal added nothing to survival and caused more late complications. The surgery, not the wash, is what works.
Reopening the abdomen to look for and wash out invisible disease did not help, and one patient in four had a serious complication.
The prostate cancer vaccine that looked like it added eight and a half months of life in a small trial added nothing in a big one.
A vaccine made from two viruses carrying the PSA gene, meant to teach the immune system to attack prostate cancer. The big trial found it did nothing at all.
QUARTZ found that whole-brain radiotherapy added nothing measurable, in survival or quality of life, for lung cancer patients whose brain metastases could not be treated with surgery or radiosurgery.
Waiting to see whether the PSA rises after prostate surgery, and only then giving radiotherapy, works as well as treating everybody straight away, and spares two men in three the radiotherapy altogether.
Blocking Src, the kinase that lets prostate cancer cells talk to bone-dissolving cells, did nothing when added to chemotherapy.
RESOLVE tested the leukaemia drug ibrutinib as a way to reprogramme the immune cells around pancreatic tumours; added to nab-paclitaxel and gemcitabine it shortened the time to progression and cut the chemotherapy patients could tolerate, with no survival gain.
Rindopepimut is a peptide vaccine against EGFRvIII, a mutant protein found only on some glioblastomas. After a phase 2 that beat historical controls, the 745-patient double-blind phase 3 ACT IV found no benefit in 2016, and tumours in both arms had lost EGFRvIII at recurrence, a lesson in antigen escape.
RTOG 0617 is the trial that stopped dose escalation in lung cancer: patients given 74 Gy lived shorter than those given the standard 60 Gy, and cetuximab added nothing.
RTOG 1016 tried to spare HPV-positive throat cancer patients the toxicity of cisplatin by using cetuximab instead, and found that survival was worse: cisplatin stays the standard.
A phase 2 trial of Acasunlimab, Pembrolizumab in non-small-cell lung cancer and small-cell lung cancer, run by Genmab, stopped early.
Scraping the lump off first and then using the cream was tested against simply cutting it out, for the thicker kind of basal cell carcinoma. It did not come close: about one in five of those treated without surgery had the tumour back within five years, against one in fifty after the operation.
The idea that sharks do not get cancer (they do) launched a supplement industry. Two randomised trials, including a large phase 3 in lung cancer, found that shark cartilage extract does nothing for survival, and the trade contributed to shark population declines.
The largest trial of a cream against an operation for the commonest cancer found the cream is worse, and stayed worse at five years. About one person in six treated with the cream still needed something else done, against one in fifty after surgery. It is still a reasonable choice for some people, but it is not equivalent.
The first randomised trial in retroperitoneal sarcoma found that radiotherapy before surgery did not reduce recurrence in the abdomen, ending a decades-long debate for most patients.
Biotin, marine-protein and multi-ingredient capsules are advertised directly to people whose hair has thinned after treatment. No randomised trial of any of them has been run in chemotherapy or endocrine-therapy hair loss. High-dose biotin also distorts hospital blood tests, including the one used to diagnose a heart attack.
For women with one to three involved nodes or a larger node-negative tumour, radiotherapy to the chest wall after mastectomy did not help them live longer. It roughly halved the small chance of the cancer returning on the chest wall, from one in forty to one in ninety.
Adding an endothelin blocker to chemotherapy changed nothing, not survival, not progression, not pain.
Adding a third chemotherapy drug, nab-paclitaxel, to gemcitabine and cisplatin did not help people with advanced bile duct or gallbladder cancer live longer in this 452-patient US trial, although the gallbladder subgroup showed a hint of slower progression that needs its own test.
Blocking the survival protein clusterin did not make chemotherapy work better for advanced prostate cancer.
When patients were checked with brain MRI first, prophylactic brain irradiation no longer lengthened survival in extensive-stage small-cell lung cancer, so many centres now offer MRI surveillance instead.
TAPPAS tested whether adding an antibody against endoglin, a protein on the abnormal blood vessel cells that make up angiosarcoma, would improve on the standard tablet pazopanib; it did not, and the trial was stopped for futility.
An oral drug aimed at the tissue around the tumour rather than the cancer cells. It held the cancer back on scans and, if anything, shortened life.
Tazemetostat was the first EZH2 inhibitor, approved for epithelioid sarcoma and follicular lymphoma in 2020 and withdrawn worldwide in 2026 after secondary blood cancers.
The attempt to move the EZH2 inhibitor tazemetostat into first-line epithelioid sarcoma alongside doxorubicin; the safety run-in enrolled 25 patients and the study was stopped, leaving the accelerated approval without its confirmatory trial.
TIGIT is an inhibitory receptor on T and natural killer cells that binds PVR (CD155) on tumour cells, so blocking it was expected to amplify PD-1 and PD-L1 inhibitors. Tiragolumab, domvanalimab and others then failed to add benefit in phase 3 lung cancer trials despite encouraging phase 2 signals, and the lack of a TIGIT-specific biomarker remains a weakness.
TreeTopp tested whether the HER-family blocker varlitinib added to capecitabine helps bile duct and gallbladder cancer after first-line chemotherapy. It did not: response, progression and survival were the same, and the planned phase 3 part was abandoned. A small gallbladder subgroup hinted at slower progression but the finding was not conclusive.
Adding chemotherapy to the radiotherapy given after surgery for the most dangerous skin squamous cancers did not help. What the trial did show is how well surgery and radiotherapy already work: more than four in five of these high-risk cancers were still controlled five years later.
Eryaspase, an enzyme carried inside red blood cells, did not lengthen survival when added to second-line chemotherapy for pancreatic cancer.
The largest screening trial ever run for ovarian cancer found earlier detection but no reduction in deaths, so there is still no recommended screening.
A PI3K inhibitor for marginal zone and follicular lymphoma approved in 2021 and withdrawn in 2022 after the UNITY-CLL trial suggested more deaths.
A pill that blocked the HER family of growth receptors, tested with capecitabine as second-line treatment for bile duct and gallbladder cancer. It did not beat capecitabine alone in the TreeTopp trial and development stopped.
Veliparib is a PARP inhibitor that AbbVie tested with chemotherapy in breast cancer. It added nothing to carboplatin before surgery (BrighTNess) and lengthened progression-free but not overall survival with carboplatin and paclitaxel in inherited BRCA advanced disease (BROCADE3), so it was never licensed.
VERONA tested whether adding venetoclax to azacitidine, the combination that transformed treatment of acute myeloid leukaemia in older patients, would also help people with higher-risk myelodysplastic syndromes live longer; it did not meet that goal.
The definitive test of whether vitamin D or fish-oil pills prevent cancer or heart disease in healthy adults. They did not.
Three thousand breast cancer survivors were coached for years to eat far more vegetables, fruit and fibre. They did, and it made no difference to recurrence or survival.
WINTHER was the first trial to pick cancer drugs using RNA, comparing what genes a tumour switched on against a biopsy of the patient's normal tissue, for patients whose DNA showed no obvious target. It missed its formal goal, but showed the approach was workable and that better-matched treatment went with better outcomes.
Xevinapant was a tablet meant to make head and neck cancer cells easier to kill with chemoradiation by removing the proteins that stop them dying. A promising mid-stage trial was followed by a phase 3 that was stopped in 2024 because it was not working.
A drug blocking a signal that prostate cancer uses to grow in bone. Three trials in prostate cancer found no survival benefit and it was dropped.