In BRCA-normal ovarian cancer, the gain in KEYLYNK-001 tracks olaparib, not pembrolizumab

In KEYLYNK-001, a phase 3 trial in women with advanced ovarian cancer without a BRCA mutation, adding pembrolizumab plus maintenance olaparib to first-line chemotherapy significantly lengthened progression-free survival — a 32% reduction in the risk of progression or death across the whole trial population (hazard ratio 0.68; 95% CI 0.58 to 0.81; p<0.0001) [s1]. But the trial's three-arm design exposes where the benefit comes from: pembrolizumab added to chemotherapy without olaparib did nothing (hazard ratio 0.95; 95% CI 0.77 to 1.19; p=0.33), so the gain tracks the PARP inhibitor, not the immunotherapy [s1].

That distinction is the whole point of reading this trial carefully. Olaparib, a PARP inhibitor, is already a first-line maintenance standard for the subset of ovarian cancers driven by a BRCA mutation, where the SOLO-1 trial cut the risk of progression or death by 70% (hazard ratio 0.30; 95% CI 0.23 to 0.41; P<0.001) [s2]. KEYLYNK-001 asked the harder question — whether an immunotherapy-plus-PARP strategy could help the larger group of patients whose tumours are BRCA non-mutated, and who are not served by that existing approval [s1][s2].

What the trial did

The trial enrolled patients with newly diagnosed stage III–IV epithelial ovarian, primary peritoneal or fallopian-tube cancer, all without a BRCA mutation, at 224 gynaecological-oncology centres in 22 countries [s1]. After one lead-in cycle of chemotherapy, 1,367 women were randomly assigned 1:1:1 into three groups: pembrolizumab plus chemotherapy then pembrolizumab-plus-olaparib maintenance (455 patients); pembrolizumab plus chemotherapy then pembrolizumab-plus-placebo maintenance (458); or placebo plus chemotherapy then placebo maintenance (454) [s1]. The median age was 61 years [s1]. Progression-free survival was the primary endpoint, tested first in patients whose tumours strongly expressed PD-L1 (a combined positive score of 10 or more) and then in the whole population, under a hierarchical rule that stops formal testing once a step fails [s1].

What it found

The pembrolizumab-plus-olaparib arm beat the control convincingly. At the first interim analysis, the hazard ratio for progression or death was 0.63 (95% CI 0.49 to 0.80; p<0.0001) in the high-PD-L1 group and 0.68 (95% CI 0.58 to 0.81; p<0.0001) across the whole population [s1]. At the final analysis, with median follow-up out to 49.6 months, the effect held: 0.66 (95% CI 0.53 to 0.83) in the high-PD-L1 group and 0.71 (95% CI 0.61 to 0.84) overall [s1].

The pembrolizumab-without-olaparib arm did not. Compared with control, its hazard ratio in the high-PD-L1 group was 0.95 (95% CI 0.77 to 1.19; p=0.33) — no benefit — and because that step of the hierarchy failed, progression-free survival was not formally tested for that arm in the whole population [s1]. Toxicity rose with the combination: grade 3 or higher treatment-related adverse events occurred in 66% of the pembrolizumab-plus-olaparib group, 56% of the pembrolizumab group and 51% of the control group, and serious treatment-related events in 24%, 21% and 9% respectively [s1]. Treatment-related deaths occurred in four patients (1%) in the combination arm and one (under 1%) in the pembrolizumab arm [s1].

How to read it

The result is real but its interpretation is bounded by what the trial could and could not separate. Because there was no olaparib-alone arm, KEYLYNK-001 cannot say how much of the combination's benefit is the PARP inhibitor doing work that olaparib might have done by itself, without the pembrolizumab bolted on [s1]. What it can say, unusually cleanly, is the negative: the immunotherapy alone added nothing, which is a useful and honest finding in a field where checkpoint inhibitors have repeatedly underdelivered in ovarian cancer [s1].

Two further limits matter. The reported endpoint is progression-free survival; the trial's own interpretation stops at "might have potential" for this population, language that signals the evidence is promising rather than definitive [s1]. And the trial was funded by Merck, which makes pembrolizumab — relevant when the headline regimen is the one built around that drug, even though the trial's structure is what reveals the drug pulling little weight [s1].

Why it matters

Most ovarian cancers are BRCA non-mutated, so a maintenance strategy that works in that group would reach far more patients than the BRCA-mutated approvals do [s1][s2]. KEYLYNK-001 offers a qualified yes — but one that points at the PARP inhibitor rather than the immunotherapy, and that raises an obvious follow-up question the trial was not built to answer: whether olaparib maintenance alone, without pembrolizumab and its added toxicity and cost, would deliver much of the same benefit.

The same PARP-inhibitor biology runs through other cancers defined by DNA-repair defects, including the BRCA-mutated prostate cancers where olaparib is also used — a reminder that these drugs are matched to a repair pathway, not to an organ.

What to watch

The decisive data are overall survival, not yet reported here, and any head-to-head that isolates olaparib's own contribution in BRCA non-mutated disease. Regulators and guideline committees will also weigh whether a combination that adds real toxicity earns its place when the component doing the work may be the cheaper, older half of it [s1].

This article describes research and regulatory developments and is not medical advice. Treatment decisions are for patients and their treating clinicians.

Sources

• Chemotherapy with or without pembrolizumab followed by maintenance pembrolizumab with or without olaparib in advanced BRCA non-mutated epithelial ovarian cancer (KEYLYNK-001) — The Lancet Oncology, online 7 September 2026
• Maintenance Olaparib in Patients with Newly Diagnosed Advanced Ovarian Cancer (SOLO-1) — New England Journal of Medicine, online 21 October 2018