Line by Line

Most women with advanced ovarian, tubal or peritoneal cancer take drugs for years, one line of treatment after another. Platinum, PARP inhibitors, bevacizumab, antibody-drug conjugates: what each does, and what the trials show.

The Inner Lining described the surgery for cancers of the peritoneum: remove everything visible, sometimes wash the abdomen with heated chemotherapy. For the most common of these cancers, high-grade cancer of the ovary, fallopian tube or the peritoneum itself, surgery is only half of the treatment. The other half is drugs, and for many women it goes on for years.

Doctors count these drugs in lines. A line is one treatment plan, from the day it starts until the cancer grows again or the plan ends; several drugs given together are one line, and so are all their cycles. The first line combines surgery with chemotherapy. Some women have surgery first; others start with chemotherapy, have the operation between courses and finish the chemotherapy afterwards. Current American guidelines recommend surgery first when the patient is fit for it and complete removal of the visible tumour is likely, and chemotherapy first otherwise.[1] After a response, maintenance treatment may follow; it is part of the same line, not a new one. If the cancer comes back, as it often does after advanced disease, a second line follows, then a third. A 2019 review in the Lancet put the stakes plainly: recurrence “is usually incurable, and patient symptom control and quality of life are key considerations at this stage”.[2] Each line is chosen to control the disease for as long as possible, at a price the patient can accept.

This essay follows those lines: the platinum drugs that have anchored treatment for decades, the PARP inhibitors that turned a genetic weakness into a target, bevacizumab, the antibody-drug conjugates now arriving, and the immune drugs that failed here at first and have now, in one setting, worked. It describes what studies have found; it is not medical advice.

67%of women with a BRCA mutation alive seven years after starting olaparib maintenance, against 46.5% with placebo

6months: the usual line between platinum-sensitive and platinum-resistant cancer

+4months of median survival, about, added by the best recent treatments for platinum-resistant cancer

2drugs approved in the US in 2026 for platinum-resistant cancer: pembrolizumab with paclitaxel, and relacorilant

Fig. 1 — Sixty years of drugs for ovarian cancer, from the trials and laboratory papers cited in this essay. Black: older milestones; pink: since 2018. The 2026 entry is a US regulatory approval, not a publication.

The backbone

A metal that stops cells dividing

The story starts with an accident. In 1965 the biophysicist Barnett Rosenberg and colleagues reported that running an electric current through platinum electrodes stopped the bacterium Escherichia coli from dividing; the culprit was not the current but platinum compounds released from the electrodes.[3] One of them, cisplatin, became a cancer drug. Platinum drugs bind to DNA and tie its strands together, and a cell that cannot copy its DNA cleanly dies when it tries to divide.

The partner came next. In a trial published in 1996, 410 women with advanced ovarian cancer and tumour left after surgery received cisplatin with either the older drug cyclophosphamide or paclitaxel, a drug first extracted from the bark of the Pacific yew that freezes the scaffolding a cell needs to divide. With paclitaxel, 73% of women whose disease could be measured responded, against 60%; median time without progression rose from 13 to 18 months, and median survival from 24 to 38 months.[4] Seven years later a trial of 792 women showed that carboplatin, a gentler relative of cisplatin, worked as well, with less damage to the stomach and kidneys: median survival 57.4 months against 48.7.[5] Carboplatin and paclitaxel every three weeks, usually for six cycles, has been the first line since.

Attempts to improve the schedule show how hard progress is. A Japanese trial of 631 women gave paclitaxel weekly instead of every three weeks and found a striking gain: median time without progression of 28.0 months against 17.2.[6] An American trial of 692 women could not repeat it. Overall, weekly paclitaxel did not help (14.7 against 14.0 months); it helped only in the minority not also receiving bevacizumab.[7]

The price

Why healthy cells suffer, and why the cancer suffers more

Chemotherapy does not recognise cancer. Cisplatin and carboplatin bind to the DNA of any cell they enter, cross-linking its bases so that the cell cannot copy it cleanly; the damage triggers the cell’s own self-destruct programme.[8] Paclitaxel locks the internal scaffolding that a dividing cell needs to pull its chromosomes apart. Both do most harm where cells divide most, and in the body that means a few tissues that renew themselves constantly.

• The bone marrow, which makes blood. Fewer white cells means a risk of infection; fewer platelets, bleeding; fewer red cells, tiredness. Low blood counts are the main reason doses are delayed or reduced.
• The lining of the mouth and gut, renewed every few days: sores and diarrhoea. The damage is not only the loss of dividing cells; it sets off inflammation, which studies in rats have traced through the signal NF-κB and inflammatory cytokines.[9]
• Hair follicles, among the fastest-dividing cells in the body: hair loss, usually starting a few weeks into treatment and reversible after it ends.

That is why chemotherapy comes in cycles. Blood counts fall in the week or two after a dose and recover before the next; the gap lets healthy tissue regrow, while the hope is that the cancer, dose after dose, loses more than it can replace (Figure 2).

Fig. 2 — Why chemotherapy comes in cycles. A schematic, not data: healthy blood-forming cells dip after each dose and recover; cancer cells, ideally, fall step by step. When the cancer becomes resistant, its line stops falling.

Not every side effect comes from dividing cells. Several appear in cells that never divide, through other routes (Figure 3).

• Nerves. Paclitaxel often causes numbness, tingling or pain in the fingers and toes. Nerve cells do not divide, but their long fibres, up to a metre from the spine to the foot, rely on the same scaffolding as rails to carry supplies to their far ends; the leading explanation is that freezing those rails disrupts that transport, though the mechanism is not fully understood and there is still no effective way to prevent it.[10] It builds up with each cycle, which is why it often limits how much paclitaxel a woman can receive.
• Kidneys and inner ear. Cisplatin is taken up by kidney cells and by the sensory hair cells of the inner ear, through transport channels, and damages them: kidney injury and permanent hearing loss.[8][11] Carboplatin causes much less of both, which is one reason it replaced cisplatin; in the 2003 trial, stomach, kidney and metabolic side effects were all more frequent with cisplatin.[5]
• Nausea. Chemotherapy makes cells in the gut release serotonin, which signals the brain’s vomiting centre; that drives the nausea of the first day, which drugs blocking the serotonin receptor 5-HT3 largely prevent. The delayed nausea of the following days depends more on another signal, substance P, blocked by a second class of drugs.[12]
• Allergy. Reactions to paclitaxel usually happen at the first infusion and are blamed mainly on the castor-oil solvent the drug is dissolved in; medicines given beforehand and slower infusions made them rare enough for the drug to be used.[13] Carboplatin is the opposite: the allergy builds with exposure. In one series of 205 women, 12% developed it, after a median of eight courses, mostly during their second or third platinum treatment, and half of the reactions were at least moderately severe.[14] That is one practical reason platinum is sometimes “saved”.

Fig. 3 — Where the side effects of platinum and paclitaxel come from. Left: tissues that divide often. Right: effects in cells that need not divide, through other mechanisms. Below: why high-grade serous cancer is more vulnerable than healthy tissue. Sources: Dasari & Tchounwou 2014; Shin 2023; Wang et al. 2023; Rojas & Slusher 2012; Price & Castells 2002; Markman et al. 1999; the Cancer Genome Atlas 2011.

Why, then, does the cancer suffer more than the gut or the marrow? Not mainly because it divides faster: many tumours divide more slowly than the lining of the gut. The difference is in the brakes. A healthy cell that detects damaged DNA stops, repairs it and divides later. High-grade serous ovarian cancer has almost always lost the main brake: in the Cancer Genome Atlas, the gene TP53 was mutated in 96% of 316 tumours, and pathway analyses suggested that about half could not repair broken DNA properly by the route that depends on BRCA and related genes.[15] Such cells push on into division with damage they cannot fix, and die. This is the same weakness that PARP inhibitors exploit, and it shows in the clinic: in an Australian study of 1,001 women, those with an inherited BRCA mutation responded more often to platinum and to other drugs at relapse, even when the cancer had come back early.[16]

Resistance is the other side of the same coin. Each course kills the cells the drug can kill; those that survive, because they repair DNA better, pump the drug out or refuse to die, seed the next recurrence. Over successive lines the cancer is, in effect, bred for resistance.

The newer drugs have side effects of their own, from different mechanisms again: PARP inhibitors affect the bone marrow, rarely causing marrow cancers;[17] bevacizumab acts on blood vessels and raises blood pressure;[18] and antibody-drug conjugates can harm healthy tissues that carry a little of their target, or where their payload escapes, such as the eyes and the lungs.[19][20] They are described in the sections that follow.

The six-month line

Sensitive and resistant

When the cancer returns, the first question is how long it has been since the last platinum. If six months or more, the cancer is called platinum-sensitive and platinum usually works again. If less, it is platinum-resistant, and platinum is usually set aside. The cut-off is a convention, but it sorts patients into groups with very different prospects, and most trials are built around it. Figure 4 shows the usual sequence.

How a recurrence is found also matters less than many assume. The blood marker CA-125 often rises months before symptoms or scans show anything. In the MRC OV05 trial, 529 women whose CA-125 rose after a complete response were randomised to start treatment at once or to wait until symptoms appeared; there was no difference in survival (median 25.7 against 27.1 months).[21] A rising marker is a warning that prompts a closer look, not by itself an instruction to start the next line.

Fig. 4 — The usual sequence of treatment for advanced high-grade ovarian, tubal or primary peritoneal cancer. Details vary between patients and countries; the six-month cut-off is a convention, and doctors also weigh side effects, earlier drugs and the tumour’s genetics.

Because each course selects for resistant cells, as described above, doctors have sometimes tried to “save” platinum, giving another drug first so that the interval since the last platinum grows longer. A European trial tested this directly in women whose cancer had returned six to twelve months after platinum. Platinum first was better: median time without progression 9 months against 5, and tumours shrank in 52% against 19%.[22] There can still be good reasons to delay platinum, such as an allergy to carboplatin or nerve damage from paclitaxel, but lengthening the interval on purpose does not help.

How resistance works

Why platinum stops working

Platinum is the backbone of treatment for this cancer, and its failure is the central problem. A 2022 review put the paradox in two numbers: the first response to platinum-based chemotherapy is “as high as 80%”, yet “in most advanced patients, final relapse and death are caused by acquired drug resistance”.[23] Understanding how a cancer that once melted away learns to survive the same drug explains much of what happens in later lines, and where the new treatments aim.

Selection, not adaptation

A tumour is not one cell but billions, and they are not all alike. Each course of chemotherapy kills the cells that cannot defend themselves; those that can, survive and regrow. If the defenders are rare at first, the cancer returns still mostly sensitive, and platinum works again. Course after course, they become the majority. Figure 5 shows the logic with invented numbers: starting with one resistant cell in a thousand, the tumour is still mostly sensitive after the first line, mostly resistant after the second, and almost entirely resistant after the third. The real numbers differ from patient to patient, but the pattern, recurrences that respond for shorter and shorter times, is the one doctors see.

Fig. 5 — How treatment breeds resistance: an illustrative model, not data. At diagnosis 0.1% of the cells resist platinum; each of six cycles per line kills 60% of the sensitive cells and 15% of the resistant ones, and the tumour regrows between lines without changing its mix. The parameters are invented to show the logic; scripts/linebyline.py computes the bars.

What makes a cell a defender? Laboratory and patient studies point to five routes, which a single tumour can combine (Figure 6).[24]

Fig. 6 — Five ways a cancer cell escapes platinum. A schematic summarising the mechanisms reviewed by Havasi et al. (2023) and Yang et al. (2022), with the examples from ovarian cancer discussed in the text.

Less drug reaches the DNA

The first three routes keep platinum away from its target. A cell can take less of the drug in, pump more of it out, or neutralise it inside with sulphur-rich molecules such as glutathione before it reaches the DNA.[24][23] One such change was caught in patients. When Ann-Marie Patch, David Bowtell and colleagues sequenced the whole genomes of tumours from 92 women, comparing cancers that never responded, that responded and then failed, and that stayed sensitive, they found in resistant recurrences a recurring genetic fusion that switched on MDR1, a pump that expels drugs from the cell.[25] That pump mainly removes paclitaxel and doxorubicin rather than platinum, which is one reason resistance to one drug can travel to the next line.

The damage is repaired

The fourth route is repair, and in ovarian cancer it has a remarkable twist. Many of these cancers respond to platinum in the first place because they cannot repair DNA properly, often because a BRCA gene is broken. Under the pressure of treatment, a cell can acquire a second mutation in the broken gene that restores its reading frame, and with it the ability to repair. In 2008 Wataru Sakai, Toshiyasu Taniguchi and colleagues showed this for BRCA2, including in a recurrent ovarian cancer that had become resistant to cisplatin; cells with these “reversion” mutations were resistant to both cisplatin and a PARP inhibitor.[26] The same group then found restoring mutations in BRCA1 in four of six recurrent platinum-resistant ovarian cancers, and in none of three that were still sensitive.[27]

Later studies showed how common and how many-sided this is. Patch and colleagues found “multiple independent reversions” of BRCA1 or BRCA2 in individual patients, different parts of the same cancer finding the same escape separately, and, in tumours where BRCA1 had been switched off chemically rather than mutated, loss of that switch.[25] In 30 women with inherited BRCA mutations, reversions were found only in recurrent disease, in five of 16, and every patient with one was resistant to platinum or PARP inhibitor treatment; they could sometimes be detected in tumour DNA circulating in the blood.[28] In a trial of the PARP inhibitor rucaparib, reversions in blood samples taken before treatment were found in 18% of platinum-refractory and 13% of platinum-resistant cancers, against 2% of platinum-sensitive ones, and women with them gained far less from the drug: a median of 1.8 months without progression, against 9.0.[29]

Repair can also improve without BRCA. The cell’s main tool for cutting platinum lesions out of DNA is nucleotide excision repair, in which the enzyme ERCC1 plays a key part. In laboratory models of ovarian cancer, resistant cells removed platinum damage about three times faster and made more ERCC1;[30] whether ERCC1 levels predict resistance in patients has been studied for decades without a clear answer.[31]

The cell refuses to die

The fifth route is to survive the damage. Platinum kills by triggering the cell’s self-destruct programme, and that programme can be turned down. High-grade serous cancers have already lost its main trigger, p53, in almost every case,[15] and survival signals can raise the threshold further. One of them is cortisol: when it reaches its receptor inside the tumour cell, it “has been shown to suppress the apoptotic pathways used by cytotoxic agents”, the rationale for adding relacorilant to chemotherapy.[32]

Resistant from the start

Some cancers never respond well. In the genome study, amplification of CCNE1, a gene that drives cells through division, was common in cancers that resisted or failed to respond to their first platinum.[25] It tends to occur in tumours whose BRCA pathway is intact, so they lack the very weakness platinum exploits; laboratory work suggests that these tumours depend on BRCA1 and on the cell’s protein-disposal system instead, a possible target for new drugs.[33]

What follows for treatment

• Platinum and PARP resistance often travel together. Both drugs exploit the same broken repair, so a reversion that restores BRCA defeats both.[26][29] That is part of why PARP inhibitors work best after a first, platinum-sensitive response.
• The six-month rule is a proxy. It measures time, not mechanism; women with inherited BRCA mutations sometimes respond to platinum again even after an early relapse,[16] which is why doctors sometimes retry it.
• Mechanisms point to new treatments. Antibody-drug conjugates deliver a different poison by a different route; relacorilant targets a survival signal; and research is under way on tests for reversions in blood and on drugs for CCNE1-amplified tumours.[28][33]

A broken repair kit

PARP inhibitors

Many high-grade ovarian cancers have a fault in the machinery that repairs broken DNA, often in the genes BRCA1 or BRCA2, sometimes elsewhere; doctors test for it as BRCA status and homologous recombination deficiency (HRD). In one large trial, 51% of tumours were HRD.[34] In 2005 two laboratories, publishing side by side in Nature, showed how to turn that fault into a target. Cells lacking BRCA depend on a second repair enzyme, PARP; block PARP, and they die, while normal cells, which still have BRCA, mostly survive.[35][36] Figure 7 shows the logic, which biologists call synthetic lethality.

Fig. 7 — Why PARP inhibitors work best in BRCA-deficient tumours. A schematic after Farmer et al. and Bryant et al. (2005): a cell survives with either repair route, and dies when both are lost.

PARP inhibitors are tablets, given as maintenance: after chemotherapy has shrunk the cancer, to keep it from growing back. The first big test came in platinum-sensitive relapse. With olaparib, median time without progression after chemotherapy was 8.4 months against 4.8 with placebo.[37] In women with a BRCA mutation, a later trial found a median survival of 51.7 months with olaparib against 38.8, although the difference narrowly missed statistical significance, partly because many women on placebo later received a PARP inhibitor.[17]

The largest effect came when PARP inhibitors moved to the first line. In women with a BRCA mutation, after their first chemotherapy, 60% of those on olaparib were free of progression at three years, against 27% on placebo.[38] Seven years on, 67% were alive against 46.5%, and 45.3% were alive without having needed any further treatment, against 20.6%.[39] The survival difference did not formally meet the strict statistical threshold set in advance (a p-value of 0.0004 against the 0.0001 required), but for a disease in which recurrence was almost the rule, the trend is a real change.[39] Niraparib, in a broader group, doubled the median time without progression in HRD tumours, from 10.4 to 21.9 months;[34] but the final analysis, after a median of more than six years, found no difference in overall survival, either in the whole trial or in HRD tumours, partly because many women on placebo later received a PARP inhibitor.[40] Olaparib added to bevacizumab did best in HRD-positive tumours: 37.2 months without progression against 17.7.[41] In its final analysis, overall survival was not improved in the trial as a whole (56.5 against 51.6 months), but in HRD-positive tumours 65.5% of women were alive at five years against 48.4%.[42] The trial had no arm with olaparib alone, so it cannot show that the combination beats olaparib on its own.

Time without progression and overall survival answer different questions, and the PARP trials show why both are needed. A drug can delay the cancer’s return without a proven effect on how long women live, especially when, as here, many in the control group later receive the same kind of drug.

The costs are real too. Anaemia is common, and a small number of patients develop bone-marrow cancers: in the relapse trial, six deaths on olaparib were judged treatment-related, from myelodysplastic syndrome or acute myeloid leukaemia.[17]

The blood supply

Bevacizumab

Bevacizumab is a biological medicine of the kind described in One Signal at a Time: an antibody made to order. It mops up VEGF, a signal tumours use to grow new blood vessels. In the first line, given with chemotherapy and then continued, it lengthened median time without progression from 10.3 to 14.1 months in a trial of 1,873 women, without a significant difference in overall survival.[43] A European trial of 1,528 women found a smaller average gain, larger in women at high risk of progression.[44]

Its main role now is in recurrent disease, including resistant disease. In the AURELIA trial, adding it to a single chemotherapy drug doubled the median time without progression, from 3.4 to 6.7 months, and tumours shrank in 27% against 12%; overall survival, 16.6 against 13.3 months, was not significantly different, partly because women on chemotherapy alone could switch to bevacizumab later. It raises blood pressure, and the bowel perforated in 2.2% of those who received it.[18] That risk is why women with a history of bowel obstruction were excluded from the trial, and why doctors use it cautiously in women with bowel problems.

After platinum

When platinum stops working

Platinum-resistant cancer is where treatment is hardest. The standard is one chemotherapy drug at a time: weekly paclitaxel, liposomal doxorubicin or topotecan, often with bevacizumab. On those drugs alone, in AURELIA, tumours shrank in about one woman in eight, and the median time without progression was three to four months.[18] The aim is to control symptoms and buy time with a treatment the patient can live with, and it is here that most new drugs are being tested.

One recent success came from an unexpected direction. Relacorilant blocks the receptor for cortisol, the stress hormone; the trial tested whether blocking that signal makes chemotherapy work better. In 381 women with platinum-resistant cancer, all previously treated with bevacizumab, 61% with a PARP inhibitor and 44% after three earlier lines, adding relacorilant to nab-paclitaxel raised median survival from 11.9 to 16.0 months; at 18 months, 46% were alive against 27%.[45][46] The combination arm received a lower dose of nab-paclitaxel (80 against 100 mg/m²), so the trial compared two regimens rather than adding one drug to an identical one.[45] The US Food and Drug Administration approved the combination in March 2026, for women who have had one to three earlier lines including bevacizumab.[47] What stress hormones do in the body, and how strong the evidence is that stress affects disease, cancer included, is the subject of Under Strain.

Guided chemotherapy

Antibody-drug conjugates

The newest family joins the two ideas above: an antibody that recognises a protein on cancer cells, carrying a chemotherapy drug on a chemical linker. The antibody finds the cell; the cell swallows it; the linker is cut inside; the drug is released where it is needed. Some drugs also leak into neighbouring cells, which helps when not every cancer cell carries the target. Figure 8 shows the steps.

Fig. 8 — How an antibody-drug conjugate works. A schematic; the bystander effect has been shown for drugs whose payload crosses cell membranes, such as the DXd payload of raludotatug deruxtecan.

The first to prove its worth in ovarian cancer was mirvetuximab soravtansine, which targets folate receptor alpha, a protein many ovarian cancers carry in large amounts. In the MIRASOL trial, 453 women with platinum-resistant cancer whose tumours had high levels of the receptor, after one to three lines, received it or chemotherapy. Tumours shrank in 42.3% against 15.9%, median survival was 16.46 months against 12.75, and severe side effects were less frequent: 41.7% against 54.1%.[48] It is approved for this group of patients; antibody-drug conjugates as a class can affect the eyes, lungs and nerves, which need watching.[19]

Others are in trials. Trastuzumab deruxtecan, already used in breast cancer, targets HER2; in 40 women with HER2-expressing ovarian cancer, tumours shrank in 45%, but inflammation of the lungs occurred in 10.8% of 120 women with gynaecological cancers, and one died of it.[20] Raludotatug deruxtecan targets CDH6, a protein expressed by some ovarian and kidney cancers; in laboratory models it shrank tumours, entered the cancer cells and affected their neighbours.[49] In the first, dose-finding part of a phase 2/3 trial, presented in October 2025, tumours shrank in 50.5% of 107 women with platinum-resistant cancer; that is an early result reported by the sponsors, not a comparison with chemotherapy.[50] The comparative part, planned for 860 women worldwide, is under way, with results expected from 2028.[51] Sacituzumab tirumotecan (target TROP2) is being tested as maintenance in platinum-sensitive relapse,[52] and ifinatamab deruxtecan (target B7-H3) in early-phase trials across several cancers.[53] The pattern is clear: find a protein the cancer carries more of than normal tissue, and use it as an address.

A late success

The brakes, released at last

The checkpoint drugs described in The Brakes Released, which transformed melanoma and lung cancer, at first failed in ovarian cancer. In platinum-resistant disease, avelumab alone did worse than chemotherapy (median time without progression 1.9 months against 3.5), and added to chemotherapy it did not significantly improve survival (15.7 against 13.1 months).[54] In the first line, atezolizumab added to chemotherapy and bevacizumab in 1,301 women made no significant difference (19.5 against 18.4 months without progression).[55]

Then a trial worked. In KEYNOTE-B96, women with platinum-resistant cancer received weekly paclitaxel, with or without bevacizumab, plus pembrolizumab or a placebo. Among the 466 whose tumours carried the protein PD-L1 (a combined positive score of at least 1), median time without progression was 8.3 months against 7.2, and median survival 18.2 months against 14.0 (hazard ratio for death 0.76). On 10 February 2026 the FDA approved the combination for these women, after one or two earlier regimens, together with the PD-L1 test that selects them.[56] It does not undo the earlier failures, and why this trial succeeded where they did not is not settled: the chemotherapy partner, the timing and the selection by PD-L1 all differ. But the blanket statement that immunotherapy does not work in ovarian cancer no longer holds.

The shrinking gain

What each line adds

Put the trials side by side and a pattern appears (Figure 9). In the first line, maintenance can add a year or more to the time without progression. In platinum-sensitive relapse, months. In platinum-resistant disease, the best recent treatments add one to three months without progression and about four months of median survival: 12.75 to 16.46 months with mirvetuximab, 11.9 to 16.0 with relacorilant, 14.0 to 18.2 with pembrolizumab in tumours with PD-L1.[48][46][56] Those months matter, and some women gain far more than the median; but further along, the average gain tends to be smaller, although the trials differ too much for this to be a law.

Fig. 9 — Median months without progression, control (black) against the new treatment (pink), in the trials discussed. The trials enrolled different patients and measured from different starting points (Study 19, for instance, from the end of chemotherapy), so the chart shows the pattern across settings, not a comparison between drugs. Sources: Ray-Coquard et al. 2019; González-Martín et al. 2019; Ledermann et al. 2012; Pujade-Lauraine et al. 2014 and 2021; Moore et al. 2023; Olawaiye et al. 2025; FDA 2026 (KEYNOTE-B96, tumours with PD-L1 CPS of 1 or more).

Reading these numbers needs three cautions. A median describes a group, not the months any one woman gains. The clock starts at different points: maintenance trials count from the end of a successful chemotherapy, trials in recurrence from the start of the new treatment, and neither from diagnosis. And the trials select their patients, by BRCA, HRD, folate receptor alpha or PD-L1, and by fitness: they enrolled women fit enough to take part, usually without bowel obstruction and with a limited number of earlier lines: AURELIA excluded women with more than two earlier regimens or a history of bowel obstruction,[18] and MIRASOL took women after one to three lines.[48] After four or five lines the evidence is much thinner, which is one reason clinical trials are often offered at that point.

The balance

What is known and what is not

Known. Carboplatin and paclitaxel remain the backbone of the first line, three decades on. Testing for BRCA and HRD matters, because PARP inhibitor maintenance after the first chemotherapy has a large, lasting benefit in BRCA-mutated cancer. The six-month platinum-free interval sorts patients and treatments, and deliberately lengthening it does not help. In platinum-resistant disease, mirvetuximab (for tumours with high folate receptor alpha), relacorilant with nab-paclitaxel and pembrolizumab with paclitaxel (for tumours with PD-L1) have each extended median survival by about four months in randomised trials. Checkpoint immunotherapy failed in its first large trials and has now worked in one selected group. Surgery with heated chemotherapy after initial chemotherapy extended survival in its long-term follow-up.

Not known. The best order of treatments after the first relapse, now that there are more options. Whether antibody-drug conjugates work better earlier, or in combination. Which tumours will respond to which conjugate: target levels help, but not perfectly. How well these results apply to women after many lines, or with bowel problems, who are mostly excluded from trials. And the trials measure time; fewer measure how that time is lived.

Next

Where this could go

If resistance is the central problem, the next treatments can be sorted by how they try to get past it. There are three broad ways (Figure 10): bring more platinum to the cancer, make the damage it causes lethal again, or switch to a different weapon that the platinum defences do not stop. The first is the oldest idea and the most limited; the third is where most of the new drugs are.

Fig. 10 — Three ways past platinum resistance, with examples and how far each has got. Pink marks treatments tested in randomised trials; grey, early studies or laboratory work. Sources in the text.

More platinum where the cancer is

Raising the dose everywhere does not work well: the bone marrow and kidneys set the ceiling, and giving the drug more often helped in one trial but not in a larger one. Delivering it locally does better. In the OVHIPEC-1 trial, 245 women with stage III ovarian cancer who had surgery after their first chemotherapy were randomised to heated cisplatin washed through the abdomen at the end of the operation, or not. Median time without recurrence rose from 10.7 to 14.2 months.[57] In the final analysis, after a median of more than ten years, median survival was 44.9 months against 33.3 (hazard ratio for death 0.70).[58] The method, and its limits in other cancers, are described in The Inner Lining. But it belongs to the first line, after surgery has removed the visible tumour; drug washed over a surface reaches only the outer layers of a deposit, so it is not a remedy for bulky resistant disease.

Getting more platinum into the cell is still experimental. Platinum enters partly through a copper transporter, and lowering the body’s copper with the drug trientine makes cells build more of it. In an exploratory study of 55 patients with advanced cancers, only eight of them with ovarian cancer and 45 who had already failed platinum, carboplatin with trientine caused no dose-limiting toxicity; nine patients had their disease held for six months or more or shrinking, including one partial response in platinum-resistant ovarian cancer, and these patients had lower copper levels than those whose disease progressed.[59] Nanoparticles that carry platinum, sometimes together with a PARP inhibitor, past the cell’s defences are being developed in the laboratory.[60] Neither approach has been tested in a randomised trial.

Make the damage lethal again

The second way accepts that the cell is resistant and attacks its defences. PARP inhibitors are the established example. The newer drugs block the pause a cell takes to repair DNA before dividing. High-grade serous cells, which almost all lack p53,[15] lean heavily on one such pause, controlled by the enzyme WEE1. In a randomised phase 2 trial in 99 women with platinum-resistant or platinum-refractory high-grade serous cancer, adding the WEE1 inhibitor adavosertib to gemcitabine lengthened median time without progression from 3.0 to 4.6 months, at the cost of more low blood counts: severe neutropenia in 62% against 30%.[61] Relacorilant, which removes cortisol’s survival signal, belongs in the same family, and is the only one of these newer drugs so far shown in a phase 3 trial to lengthen life.[46] For tumours with CCNE1 amplification, which are resistant from the start, laboratory work points to their own dependencies, such as the cell’s protein-disposal system.[33]

Switch weapons

The third way is the one most new drugs take, and it is about addresses. Each antibody-drug conjugate needs a protein that the cancer carries more of than healthy tissue, and the list for ovarian cancer is growing: folate receptor alpha, HER2, CDH6, TROP2, B7-H3 and others in trials.[19] They get past platinum resistance because they enter through their target, not the copper transporter, and carry a different poison: a drug that disrupts the cell’s scaffolding in mirvetuximab, one that blocks the untangling of DNA in the deruxtecan family. Restored BRCA and extra glutathione give little protection against them. But they have resistance of their own. A cancer can lose its target from the surface, and pumps such as MDR1, switched on in some resistant ovarian tumours,[25] can expel some of their payloads. Their side effects on the lungs and eyes will decide how far and how early they spread.[19][20]

Choosing by test

All three ways depend on knowing what the tumour has. Tumours are already tested for BRCA and HRD to choose PARP inhibitors, for folate receptor alpha before mirvetuximab, which in MIRASOL was given only to women whose tumours had high levels,[48] and, since 2026, for PD-L1 before pembrolizumab.[56] Blood tests may add more. Reversion mutations, which restore BRCA and defeat both platinum and PARP inhibitors, can sometimes be found in tumour DNA circulating in the blood,[28] and in the rucaparib trial their presence before treatment marked women who gained little from the drug.[29] The likely direction is that each woman’s next line will be chosen by testing her cancer, rather than taken from a fixed list.

Approval is not the same as access. The American and European decisions mentioned here say what regulators accepted; whether a drug is available and paid for in a given country, and for whom, is decided separately. For patients, two practical things follow. Ask which markers the tumour has been tested for, because they open or close options. And ask about clinical trials early, while the lines are still few: eligibility often depends on how many there have been, and on being well enough to take part.

On method and tools

This essay was researched and written with Claude Opus 5.5 (Anthropic): human specification, editorial direction and critical review; machine research, drafting and figure generation. The subject was Luis’s choice. Every reference was checked against its PubMed record, and the figures and quotations come from the abstracts; the trials still under way are described from their records on ClinicalTrials.gov. The 2026 US approvals of pembrolizumab with paclitaxel and of relacorilant, and the KEYNOTE-B96 figures, come from the FDA’s announcements, and the early results for raludotatug deruxtecan from the sponsors’ press release; KEYNOTE-B96 had not been read in a peer-reviewed paper. Several corrections and additions in this version (the KEYNOTE-B96 result, the final survival analyses of OVHIPEC-1, PRIMA and PAOLA-1, the CA-125 trial, the nab-paclitaxel dose difference in ROSELLA and the PFS and survival cautions) were suggested by a review of the draft with ChatGPT (OpenAI) and then checked against the sources by Claude. The general statements about how paclitaxel acts, which tissues renew fastest, the timing of falls in blood counts and of hair loss, and the reasons for giving chemotherapy in cycles are textbook background, not taken from the papers cited; the cycles figure is a schematic with invented shapes. The figure on how treatment breeds resistance comes from a small illustrative model in the same script, with invented parameters (0.1% resistant cells at diagnosis; per cycle, 60% of sensitive and 15% of resistant cells killed; six cycles per line); it shows the logic, not real proportions, which are listed in docs/linebyline-results.json. That the MDR1 pump mainly expels paclitaxel and doxorubicin, and that glutathione neutralises platinum, are textbook pharmacology. So are the statements that drug washed over a surface penetrates only its outer layers, that high-grade serous cells depend on the WEE1-controlled pause, and how the payloads of antibody-drug conjugates act. The figures are drawn by scripts/linebyline.py from the published numbers named in their captions; the medians are listed in docs/linebyline-results.json. There is no model in this essay. The cover is a drawing, not data. This essay describes what studies have found; it is not medical advice.

Authored by: Luis Matos Ferreira — Physicist, Developer, Writer

Related essays on this blog

1. The Inner Lining — cancers of the peritoneum, and what surgery with heated chemotherapy can do.
2. The Brakes Released — cancer immunotherapy: checkpoint drugs, CAR-T cells and personalised vaccines.
3. One Signal at a Time — the biological medicines, antibodies made to order, used in autoimmune disease.

Sources

1. S. Gaillard, C. Lacchetti, D. K. Armstrong et al., “Neoadjuvant Chemotherapy for Newly Diagnosed, Advanced Ovarian Cancer: ASCO Guideline Update”, Journal of Clinical Oncology 43, 868–891 (2025), doi.org.
2. S. Lheureux, C. Gourley, I. Vergote et al., “Epithelial ovarian cancer”, Lancet (London, England) 393, 1240–1253 (2019), doi.org.
3. B. Rosenberg, L. VanCamp & T. Krigas, “Inhibition of cell division in Escherichia coli by electrolysis products from a platinum electrode”, Nature 205, 698–699 (1965), doi.org.
4. W. P. McGuire, W. J. Hoskins, M. F. Brady et al., “Cyclophosphamide and cisplatin compared with paclitaxel and cisplatin in patients with stage III and stage IV ovarian cancer”, The New England Journal of Medicine 334, 1–6 (1996), doi.org.
5. R. F. Ozols, B. N. Bundy, B. E. Greer et al., “Phase III trial of carboplatin and paclitaxel compared with cisplatin and paclitaxel in patients with optimally resected stage III ovarian cancer: a Gynecologic Oncology Group study”, Journal of Clinical Oncology 21, 3194–3200 (2003), doi.org.
6. N. Katsumata, M. Yasuda, F. Takahashi et al., “Dose-dense paclitaxel once a week in combination with carboplatin every 3 weeks for advanced ovarian cancer: a phase 3, open-label, randomised controlled trial”, Lancet (London, England) 374, 1331–1338 (2009), doi.org.
7. J. K. Chan, M. F. Brady, R. T. Penson et al., “Weekly vs. Every-3-Week Paclitaxel and Carboplatin for Ovarian Cancer”, The New England Journal of Medicine 374, 738–748 (2016), doi.org.
8. S. Dasari & P. B. Tchounwou, “Cisplatin in cancer therapy: molecular mechanisms of action”, European Journal of Pharmacology 740, 364–378 (2014), doi.org.
9. R. M. Logan, A. M. Stringer, J. M. Bowen et al., “Is the pathobiology of chemotherapy-induced alimentary tract mucositis influenced by the type of mucotoxic drug administered?”, Cancer Chemotherapy and Pharmacology 63, 239–251 (2009), doi.org.
10. G. J. Shin, “Towards a mechanistic understanding of axon transport and endocytic changes underlying paclitaxel-induced peripheral neuropathy”, Experimental Neurology 359, 114258 (2023), doi.org.
11. X. Wang, Y. Zhou, D. Wang et al., “Cisplatin-induced ototoxicity: From signaling network to therapeutic targets”, Biomedicine & Pharmacotherapy = Biomedecine & Pharmacotherapie 157, 114045 (2023), doi.org.
12. C. Rojas & B. S. Slusher, “Pharmacological mechanisms of 5-HT₃ and tachykinin NK₁ receptor antagonism to prevent chemotherapy-induced nausea and vomiting”, European Journal of Pharmacology 684, 1–7 (2012), doi.org.
13. K. S. Price & M. C. Castells, “Taxol reactions”, Allergy and Asthma Proceedings 23, 205–208 (2002), pubmed.ncbi.nlm.nih.gov.
14. M. Markman, A. Kennedy, K. Webster et al., “Clinical features of hypersensitivity reactions to carboplatin”, Journal of Clinical Oncology 17, 1141 (1999), doi.org.
15. Cancer Genome Atlas Research Network, “Integrated genomic analyses of ovarian carcinoma”, Nature 474, 609–615 (2011), doi.org.
16. K. Alsop, S. Fereday, C. Meldrum et al., “BRCA mutation frequency and patterns of treatment response in BRCA mutation-positive women with ovarian cancer: a report from the Australian Ovarian Cancer Study Group”, Journal of Clinical Oncology 30, 2654–2663 (2012), doi.org.
17. A. Poveda, A. Floquet, J. A. Ledermann et al., “Olaparib tablets as maintenance therapy in patients with platinum-sensitive relapsed ovarian cancer and a BRCA1/2 mutation (SOLO2/ENGOT-Ov21): a final analysis of a double-blind, randomised, placebo-controlled, phase 3 trial”, The Lancet Oncology 22, 620–631 (2021), doi.org.
18. E. Pujade-Lauraine, F. Hilpert, B. Weber et al., “Bevacizumab combined with chemotherapy for platinum-resistant recurrent ovarian cancer: The AURELIA open-label randomized phase III trial”, Journal of Clinical Oncology 32, 1302–1308 (2014), doi.org.
19. J. He, M. Chen, L. Jia et al., “Antibody-drug conjugates in gynecologic cancer: current landscape, clinical data, and emerging targets”, International Journal of Gynecological Cancer 35, 101978 (2025), doi.org.
20. J. Y. Lee, A. Oaknin, L. Manso et al., “Trastuzumab deruxtecan in HER2-expressing gynecologic cancers from DESTINY-PanTumor02: antitumor activity, safety, and exploratory biomarker analyses”, Journal of Gynecologic Oncology 37, e65 (2026), doi.org.
21. G. J. Rustin, M. E. van der Burg, C. L. Griffin et al., “Early versus delayed treatment of relapsed ovarian cancer (MRC OV05/EORTC 55955): a randomised trial”, Lancet (London, England) 376, 1155–1163 (2010), doi.org.
22. M. C. Piccirillo, G. Scambia, A. Bologna et al., “Quality-of-life analysis of the MITO-8, MaNGO, BGOG-Ov1, AGO-Ovar2.16, ENGOT-Ov1, GCIG study comparing platinum-based versus non-platinum-based chemotherapy in patients with partially platinum-sensitive recurrent ovarian cancer”, Annals of Oncology 29, 1189–1194 (2018), doi.org.
23. L. Yang, H. J. Xie, Y. Y. Li et al., “Molecular mechanisms of platinum‑based chemotherapy resistance in ovarian cancer (Review)”, Oncology Reports 47 (2022), doi.org.
24. A. Havasi, S. S. Cainap, A. T. Havasi et al., “Ovarian Cancer-Insights into Platinum Resistance and Overcoming It”, Medicina (Kaunas, Lithuania) 59 (2023), doi.org.
25. A. M. Patch, E. L. Christie, D. Etemadmoghadam et al., “Whole-genome characterization of chemoresistant ovarian cancer”, Nature 521, 489–494 (2015), doi.org.
26. W. Sakai, E. M. Swisher, B. Y. Karlan et al., “Secondary mutations as a mechanism of cisplatin resistance in BRCA2-mutated cancers”, Nature 451, 1116–1120 (2008), doi.org.
27. E. M. Swisher, W. Sakai, B. Y. Karlan et al., “Secondary BRCA1 mutations in BRCA1-mutated ovarian carcinomas with platinum resistance”, Cancer Research 68, 2581–2586 (2008), doi.org.
28. E. L. Christie, S. Fereday, K. Doig et al., “Reversion of BRCA1/2 Germline Mutations Detected in Circulating Tumor DNA From Patients With High-Grade Serous Ovarian Cancer”, Journal of Clinical Oncology 35, 1274–1280 (2017), doi.org.
29. K. K. Lin, M. I. Harrell, A. M. Oza et al., “BRCA Reversion Mutations in Circulating Tumor DNA Predict Primary and Acquired Resistance to the PARP Inhibitor Rucaparib in High-Grade Ovarian Carcinoma”, Cancer Discovery 9, 210–219 (2019), doi.org.
30. K. V. Ferry, T. C. Hamilton & S. W. Johnson, “Increased nucleotide excision repair in cisplatin-resistant ovarian cancer cells: role of ERCC1-XPF”, Biochemical Pharmacology 60, 1305–1313 (2000), doi.org.
31. E. Reed, “Platinum-DNA adduct, nucleotide excision repair and platinum based anti-cancer chemotherapy”, Cancer Treatment Reviews 24, 331–344 (1998), doi.org.
32. A. B. Olawaiye, J. W. Kim, A. Bagameri et al., “Clinical Trial Protocol for ROSELLA: a phase 3 study of relacorilant in combination with nab-paclitaxel versus nab-paclitaxel monotherapy in advanced platinum-resistant ovarian cancer”, Journal of Gynecologic Oncology 35, e111 (2024), doi.org.
33. D. Etemadmoghadam, B. A. Weir, G. Au-Yeung et al., “Synthetic lethality between CCNE1 amplification and loss of BRCA1”, Proceedings of the National Academy of Sciences of the United States of America 110, 19489–19494 (2013), doi.org.
34. A. González-Martín, B. Pothuri, I. Vergote et al., “Niraparib in Patients with Newly Diagnosed Advanced Ovarian Cancer”, The New England Journal of Medicine 381, 2391–2402 (2019), doi.org.
35. H. Farmer, N. McCabe, C. J. Lord et al., “Targeting the DNA repair defect in BRCA mutant cells as a therapeutic strategy”, Nature 434, 917–921 (2005), doi.org.
36. H. E. Bryant, N. Schultz, H. D. Thomas et al., “Specific killing of BRCA2-deficient tumours with inhibitors of poly(ADP-ribose) polymerase”, Nature 434, 913–917 (2005), doi.org.
37. J. Ledermann, P. Harter, C. Gourley et al., “Olaparib maintenance therapy in platinum-sensitive relapsed ovarian cancer”, The New England Journal of Medicine 366, 1382–1392 (2012), doi.org.
38. K. Moore, N. Colombo, G. Scambia et al., “Maintenance Olaparib in Patients with Newly Diagnosed Advanced Ovarian Cancer”, The New England Journal of Medicine 379, 2495–2505 (2018), doi.org.
39. P. DiSilvestro, S. Banerjee, N. Colombo et al., “Overall Survival With Maintenance Olaparib at a 7-Year Follow-Up in Patients With Newly Diagnosed Advanced Ovarian Cancer and a BRCA Mutation: The SOLO1/GOG 3004 Trial”, Journal of Clinical Oncology 41, 609–617 (2023), doi.org.
40. B. J. Monk, M. P. Barretina-Ginesta, B. Pothuri et al., “Niraparib first-line maintenance therapy in patients with newly diagnosed advanced ovarian cancer: final overall survival results from the PRIMA/ENGOT-OV26/GOG-3012 trial”, Annals of Oncology 35, 981–992 (2024), doi.org.
41. I. Ray-Coquard, P. Pautier, S. Pignata et al., “Olaparib plus Bevacizumab as First-Line Maintenance in Ovarian Cancer”, The New England Journal of Medicine 381, 2416–2428 (2019), doi.org.
42. I. Ray-Coquard, A. Leary, S. Pignata et al., “Olaparib plus bevacizumab first-line maintenance in ovarian cancer: final overall survival results from the PAOLA-1/ENGOT-ov25 trial”, Annals of Oncology 34, 681–692 (2023), doi.org.
43. R. A. Burger, M. F. Brady, M. A. Bookman et al., “Incorporation of bevacizumab in the primary treatment of ovarian cancer”, The New England Journal of Medicine 365, 2473–2483 (2011), doi.org.
44. T. J. Perren, A. M. Swart, J. Pfisterer et al., “A phase 3 trial of bevacizumab in ovarian cancer”, The New England Journal of Medicine 365, 2484–2496 (2011), doi.org.
45. A. B. Olawaiye, L. Gladieff, D. M. O'Malley et al., “Relacorilant and nab-paclitaxel in patients with platinum-resistant ovarian cancer (ROSELLA): an open-label, randomised, controlled, phase 3 trial”, Lancet (London, England) 405, 2205–2216 (2025), doi.org.
46. D. Lorusso, L. Gladieff, D. M. O'Malley et al., “Overall survival with relacorilant and nab-paclitaxel in patients with platinum-resistant ovarian cancer (ROSELLA): a phase 3 randomised controlled trial”, Lancet (London, England) 407, 1513–1524 (2026), doi.org.
47. US Food and Drug Administration, “FDA approves relacorilant with nab-paclitaxel for platinum-resistant epithelial ovarian, fallopian tube, or primary peritoneal cancer”, 25 March 2026 (page consulted 3 November 2026), fda.gov.
48. K. N. Moore, A. Angelergues, G. E. Konecny et al., “Mirvetuximab Soravtansine in FRα-Positive, Platinum-Resistant Ovarian Cancer”, The New England Journal of Medicine 389, 2162–2174 (2023), doi.org.
49. H. Suzuki, S. Nagase, C. Saito et al., “Raludotatug Deruxtecan, a CDH6-Targeting Antibody-Drug Conjugate with a DNA Topoisomerase I Inhibitor DXd, Is Efficacious in Human Ovarian and Kidney Cancer Models”, Molecular Cancer Therapeutics 23, 257–271 (2024), doi.org.
50. Merck and Daiichi Sankyo, press release on the phase 2 part of REJOICE-Ovarian01 presented at ESMO 2025 (abstract LBA42), 19 October 2025; a sponsor report, not a comparison with chemotherapy, merck.com.
51. ClinicalTrials.gov, NCT06161025, “A Phase 2/3, Multicenter, Randomized Study of Raludotatug Deruxtecan (R-DXd), a CDH6-directed Antibody-drug Conjugate, in Subjects With Platinum-resistant, High-grade Ovarian, Primary Peritoneal, or Fallopian Tube Cancer” (record consulted 2 November 2026), clinicaltrials.gov.
52. ClinicalTrials.gov, NCT06824467, TroFuse-022: sacituzumab tirumotecan maintenance versus standard of care in platinum-sensitive recurrent ovarian cancer (record consulted 2 November 2026), clinicaltrials.gov.
53. ClinicalTrials.gov, NCT06330064, IDeate-PanTumor02: ifinatamab deruxtecan in recurrent or metastatic solid tumours (record consulted 2 November 2026), clinicaltrials.gov.
54. E. Pujade-Lauraine, K. Fujiwara, J. A. Ledermann et al., “Avelumab alone or in combination with chemotherapy versus chemotherapy alone in platinum-resistant or platinum-refractory ovarian cancer (JAVELIN Ovarian 200): an open-label, three-arm, randomised, phase 3 study”, The Lancet Oncology 22, 1034–1046 (2021), doi.org.
55. K. N. Moore, M. Bookman, J. Sehouli et al., “Atezolizumab, Bevacizumab, and Chemotherapy for Newly Diagnosed Stage III or IV Ovarian Cancer: Placebo-Controlled Randomized Phase III Trial (IMagyn050/GOG 3015/ENGOT-OV39)”, Journal of Clinical Oncology 39, 1842–1855 (2021), doi.org.
56. US Food and Drug Administration, “FDA approves pembrolizumab with paclitaxel for platinum-resistant epithelial ovarian, fallopian tube, or primary peritoneal carcinoma”, 10 February 2026, with the KEYNOTE-B96 results in the PD-L1 CPS ≥1 population (page consulted 3 November 2026), fda.gov.
57. W. J. van Driel, S. N. Koole, K. Sikorska et al., “Hyperthermic Intraperitoneal Chemotherapy in Ovarian Cancer”, The New England Journal of Medicine 378, 230–240 (2018), doi.org.
58. S. L. Aronson, M. Lopez-Yurda, S. N. Koole et al., “Cytoreductive surgery with or without hyperthermic intraperitoneal chemotherapy in patients with advanced ovarian cancer (OVHIPEC-1): final survival analysis of a randomised, controlled, phase 3 trial”, The Lancet Oncology 24, 1109–1118 (2023), doi.org.
59. S. Fu, M. M. Hou, J. Wheler et al., “Exploratory study of carboplatin plus the copper-lowering agent trientine in patients with advanced malignancies”, Investigational New Drugs 32, 465–472 (2014), doi.org.
60. L. B. Mensah, S. W. Morton, J. Li et al., “Layer-by-layer nanoparticles for novel delivery of cisplatin and PARP inhibitors for platinum-based drug resistance therapy in ovarian cancer”, Bioengineering & Translational Medicine 4, e10131 (2019), doi.org.
61. S. Lheureux, M. C. Cristea, J. P. Bruce et al., “Adavosertib plus gemcitabine for platinum-resistant or platinum-refractory recurrent ovarian cancer: a double-blind, randomised, placebo-controlled, phase 2 trial”, Lancet (London, England) 397, 281–292 (2021), doi.org.