KEYLYNK-001: the three-arm design is what makes this result readable

Design

Randomised, double-blind, placebo-controlled phase 3 trial with three arms at 224 centres in 22 countries

Population

1,367 patients with stage III–IV BRCA1/2 non-mutated epithelial ovarian, primary peritoneal or fallopian tube cancer, median age 61

Primary outcome

Progression-free survival, tested hierarchically in the PD-L1 CPS 10 or higher population then the intention-to-treat population

Effect

Pembrolizumab-olaparib HR 0.66 (95% CI 0.53 to 0.83) at CPS 10 or higher and 0.71 (0.61 to 0.84) in ITT; pembrolizumab alone HR 0.95 (0.77 to 1.19, p = 0.33)

Across 224 centres in 22 countries, 1,367 patients with stage III–IV epithelial ovarian, primary peritoneal or fallopian tube cancer without a BRCA mutation were randomly assigned in a 1:1:1 ratio, double-blind, after one lead-in cycle of chemotherapy. One arm had pembrolizumab with chemotherapy then pembrolizumab-olaparib maintenance; one had pembrolizumab with chemotherapy then pembrolizumab alone; one had placebo throughout. Bevacizumab was at investigator discretion. At the final analysis, median follow-up 49.6 months, pembrolizumab-olaparib improved progression-free survival against control — HR 0.66 (95% CI 0.53 to 0.83) where the PD-L1 combined positive score was 10 or higher, and 0.71 (0.61 to 0.84) in the intention-to-treat population. Pembrolizumab alone gave HR 0.95 (0.77 to 1.19, p = 0.33).

That middle arm is the reason this trial answers something. Adding a checkpoint inhibitor to chemotherapy in this population did nothing, so the progression-free survival gain in the combination arm is attributable to olaparib maintenance rather than to immunotherapy. Ovarian cancer has resisted checkpoint blockade repeatedly, and this is a well-powered confirmation rather than another equivocal signal.

The cost side has to be stated with the benefit. Grade 3 or higher treatment-related adverse events occurred in 66% of the pembrolizumab-olaparib group against 51% of controls, serious treatment-related events in 24% against 9%, and four patients in that arm died of treatment-related causes — cerebral haemorrhage, sclerosing cholangitis, haemophagocytic lymphohistiocytosis and colitis. Overall survival is not reported in this analysis. So the honest framing for a patient without a BRCA mutation is a delay in progression, of unproven effect on survival, bought with a substantially heavier toxicity burden and two years of maintenance therapy.

• Do not add pembrolizumab to first-line chemotherapy for BRCA non-mutated ovarian cancer on this evidence.
• Attribute the maintenance benefit to olaparib — the pembrolizumab-only arm was flat.
• Quote the toxicity when consenting: serious treatment-related events rose from 9% to 24%.
• Say clearly that no overall survival benefit has been shown in this analysis.
• PD-L1 combined positive score of 10 or more did not identify a group who benefited from pembrolizumab alone.

The statistics, in plain English

The two hazard ratios that matter here are 0.71 for the combination and 0.95 for pembrolizumab alone. The first excludes 1.0 with a reasonably tight interval, the second sits on it — and because the arms shared the same chemotherapy backbone and the same trial population, the difference between them isolates the contribution of olaparib far better than a cross-trial comparison could. Note also what a progression-free survival hazard ratio does not tell you: it measures time to radiological or clinical progression, not how long the patient lives, and in ovarian cancer the two have often diverged. With overall survival unreported, the size of the toxicity difference is the more certain number in this trial.

Read the rest in the app

You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.