Reviews/Perioperative Immunotherapy Solid Tumours/Adjuvant pembrolizumab disease-free and overall survival in renal cell carcinoma
RCC · PERIO-RCC-01 · v2
Borderline: within 4 points of the Established/Likely boundary, so a re-grade could land either side.
One large, well-conducted trial in which patients were randomly given either pembrolizumab or a dummy treatment after kidney-cancer surgery, with neither patients nor doctors knowing which, found that the drug lengthened the time patients stayed free of returning cancer and, at a later look, lengthened how long they lived [11],[10],[13]. The same trial's findings held up as patients were followed for longer and pointed the same way in its East Asian participants [21],[5]. The main limitation is that every efficacy number comes from this single trial; the corpus contains no second, independent trial testing the same question. A separate modelling study raised a doubt about whether the survival advantage is solid, because many patients in the comparison group did not go on to receive today's standard follow-up treatments [16]. The survival advantage also only reached statistical significance at a late analysis, not at earlier ones [2]. The panel treated the benefit as supported but not settled, because of the reliance on one trial and the live question over the survival signal's durability.
The corpus carries both limbs of the claim through KEYNOTE-564, a double-blind placebo-controlled phase 3 trial of 994 randomized patients [11],[10]. Disease-free survival improved at the primary analysis (hazard ratio 0.68, 95% CI 0.53 to 0.87, P=0.002) [11] and remained consistent at later follow-ups (HR 0.63 at 30 months [21]; HR 0.72 at ~57 months [10],[13]). Overall survival reached statistical significance at the third prespecified interim analysis (HR 0.62, with 48-month OS 91.2% vs 86.0%) [10],[13]. The East Asian subgroup of the same trial was directionally concordant for both endpoints (DFS HR 0.70, OS HR 0.47) [5]. All of this efficacy derives from one trial [2],[5],[10],[11],[13],[21]. The contested point: The division is over the durability of the overall-survival benefit. A simulation reanalysis found that 52% of progressing control-arm patients never received first-line checkpoint therapy, and after correcting for this undertreatment the re-estimated base-case OS hazard ratio was 1.41 (0.93 to 2.14), with the authors stating the benefit 'may not be statistically durable' [16]. The OS result also crossed the significance boundary only at the third interim analysis, not at earlier reports where it did not cross the testing boundary [2].
All 27 papers behind this claim with their DOIs and what each does to it, the panel’s arbitration, the evidence gap map and the open research questions. Free, and sent to your inbox.
We’ll also send you other claims from Perioperative Immunotherapy Solid Tumours, one at a time, and tell you when a release or a pivotal trial changes a grade. Unsubscribe by replying.
One reviewer's only job is to attack the finding. The challenges below were weighed before the grade was set.
All efficacy estimates derive from one trial, KEYNOTE-564 ([2],[5],[10],[11],[13],[21] are the same NCT03142334, not independent studies). The corpus contains no independent adjuvant-pembrolizumab RCT, so evidential independence is near zero.
The OS benefit's durability is challenged: [16] shows 52% of progressing control-arm patients did not receive first-line checkpoint therapy, and a simulation correcting for this undertreatment yields a re-estimated base-case OS HR of 1.41 (0.93-2.14), stating the benefit 'may not be statistically durable.'
Where it remains uncertain. Every efficacy estimate comes from one trial [2],[5],[10],[11],[13],[21], and a simulation correcting for control-arm undertreatment re-estimated an OS HR of 1.41 (0.93 to 2.14) [16].
Aleatory
OS event counts are modest (55 vs 86 events [13]) and subgroup and reanalysis intervals are wide (e.g. [16] base-case 0.93 to 2.14), so random variation still shapes the survival estimate.
Epistemic
All efficacy derives from one trial, KEYNOTE-564 [2],[5],[10],[11],[13],[21]; the corpus contains no second adjuvant-pembrolizumab RCT to confirm the effect.
Inferential
The trial enrolled 'increased risk of recurrence' spanning intermediate-high, high, and M1 NED disease [2],[10], broader than the claim's 'high-risk', so applying the result exactly to 'high-risk' requires narrowing the tested population.
The load-bearing papers from a corpus of 27, each classified by what it does to the claim. Every DOI is taken from the corpus the claim was graded on.
Establishes the OS (hard-outcome) benefit at significance.
“A significant improvement in overall survival was observed with pembrolizumab as compared with placebo (hazard ratio for death, 0.62; 95% CI, 0.44 to 0.87; P = 0.005). The estimated overall survival at 48 months was 91.2% in the pembrolizumab group, as compared with 86.0% in the placebo group”
Panel auditInternal validity: high · scope of inference: Adjuvant pembrolizumab vs placebo in clear-cell RCC at increased risk after surgery.Key limitations: higher serious/grade 3-4 adverse events with pembrolizumab
Primary endpoint establishing the DFS benefit.
“Pembrolizumab therapy was associated with significantly longer disease-free survival than placebo (disease-free survival at 24 months, 77.3% vs. 68.1%; hazard ratio for recurrence or death, 0.68; 95% confidence interval [CI], 0.53 to 0.87; P = 0.002 [two-sided]).”
Panel auditInternal validity: high · scope of inference: Adjuvant pembrolizumab vs placebo in high-risk clear-cell RCC after nephrectomy.Key limitations: OS immature at this interim · higher grade 3+ adverse events with pembrolizumab
Confirms the OS benefit and notes it is the first phase 3 adjuvant RCC survival signal.
“Statistically significant improvement in OS was observed with pembrolizumab vs placebo (medians not reached, HR 0.62, 95% CI 0.44−0.87; P=.0024).”
Panel auditInternal validity: high · scope of inference: Adjuvant pembrolizumab vs placebo in clear-cell RCC at increased recurrence risk.Key limitations: conference-abstract presentation of the interim data
The simulation argues the reported adjuvant OS benefit may not be statistically durable once control-arm undertreatment is corrected.
Panel auditInternal validity: low · scope of inference: Sensitivity of the KEYNOTE-564 OS benefit to subsequent-therapy assumptions in the control arm.Key limitations: modeled not observed data · results vary across mortality assumptions · re-estimated HR CIs cross 1
4 of 27 papers shown. The full review lists all of them, with the panel’s arbitration →
Claims are graded at each quarterly release, and re-graded between releases only when a pivotal trial lands. Every change is a new version of the claim.
Both endpoints are statistically significant in a high-quality double-blind RCT, but reliance on one trial and a live OS-durability challenge warrant a Likely rather than Established tier.
Release 2026-q4 re-graded every claim from its corpus under protocol v0.19. An earlier release's grade is shown for the record; it came from an earlier protocol and is not a movement.
Assumptions that carry weight but are not themselves established.
The panel records the gap that would most move the claim, and the study that would close it.
No independent replication of the adjuvant-pembrolizumab effect. Every efficacy estimate comes from a single trial, so a confirmatory result cannot be distinguished from a one-trial artifact. The panel rates this the highest-importance open gap for the claim, and names the study design that would close it: An independent, adequately powered double-blind placebo-controlled adjuvant trial of pembrolizumab in high-risk clear-cell RCC reporting DFS and OS.
Open question: Does adjuvant pembrolizumab improve overall survival when control-arm patients receive current standard-of-care subsequent therapy at progression? Why it is still open: The pivotal trial's OS benefit was estimated against a control arm in which 52% of progressing patients never received first-line checkpoint therapy, and a simulation correcting for this re-estimated an OS HR of 1.41 (0.93 to 2.14) [16].
Priors is a sceptical first reader, not peer review, and not medical advice.
Release 2026-q4 · graded 21 Oct 2026 · protocol v0.19