PE-PE NCT05578664
ForOligometastatic RCC, ≤3 mets within 5y of prior nephrectomy, MDT-eligible
HR 0.47
95% CI 0.24-0.92, p=0.024; mRFS 25.4 vs 11.0 mo
TL;DRmRFS 25.4 vs 11.0 mo, HR 0.47 (0.24-0.92), p=0.024, adding 1y perioperative pembrolizumab to SABR/surgery MDT in omRCC.
SABR was the MDT in 73% vs 72% of pts, so this is mostly pembrolizumab plus SABR vs SABR alone. Dose/fractionation and the local relapse secondary aren't in the abstract, so it can't yet show whether P changes in-field control or only cuts distant events. The decision it moves is whether SABR to all sites goes ahead with or without IO.
In post-nephrectomy omRCC with ≤3 mets within 5y planned for SABR or metastasectomy, this supports discussing 1y of perioperative pembrolizumab with MDT; it does not extend to >3 mets or relapse beyond 5y of nephrectomy.
SABR was the MDT in 73% vs 72%, so this tests IO on top of SABR to all sites. Dose and the local relapse secondary are unreported, so whether P improves in-field control or only distant events is unknown, and SABR technique itself is unchanged.
One year of pembrolizumab, started before MDT, gave HR 0.47 for RFS beyond the M1NED-within-1y adjuvant setting. G3 AEs were 7% and hypothyroidism 23%, and with 2 deaths there is no OS read to weigh against deferring IO until relapse.
10 details
Randomized 2:1, open-label, multicenter phase II, N=85 (56 P+MDT, 29 MDT alone). Median follow-up 17.5 mo.
omRCC after prior nephrectomy, ≤3 metastases within 5 years of Nx or prior metastasectomy, candidates for MDT to all metastatic sites. Baseline characteristics reported as balanced.
Pembrolizumab 400 mg q6w × 9 cycles (1 year), first dose before MDT. Median cycles delivered 6.6.
MDT was SABR in 73% vs 72% and surgery in 23% vs 28%. Dose, fractionation and target definition not reported in source.
Primary: relapse-free survival. Secondary: distant RFS, local relapse, OS, safety; only safety secondaries are reported in the abstract.
Any AE 80% vs 21%, G3 7% vs 3%. P-related AEs 59%, led by hypothyroidism 23%; no G4/5.
KEYNOTE-564 established adjuvant pembrolizumab after Nx and for M1NED within 1 year of Nx. PE-PE moves that into unresected oligomets ablated by SABR or surgery, including relapse up to 5 years out.
No local vs distant relapse split, so the site of benefit is unknown. RFS assessor blinding not stated in an open-label trial, and 2 of the 9 planned cycles went undelivered at the median.
The RFS gain is large but rests on 35 events, with a CI upper bound of 0.92. It does not settle whether IO at MDT beats IO deferred to relapse, which is the comparison that decides treatment-free intervals.
| Endpoint | P+MDT (n=56) | MDT alone (n=29) | Effect |
|---|---|---|---|
| Median RFS | 25.4 mo | 11.0 mo | HR 0.47 (95% CI 0.24-0.92), p=0.024 |
| RFS events | 19 | 16 | 35 total |
| Any AE | 80% | 21% | n/a |
| G3 AE | 7% | 3% | n/a |
CONSORT flow
Randomized phase II met its RFS primary, but N=85 with a 29-pt control arm, 17.5-mo median f/u, open-label, and only 2 deaths; abstract-only.
- Local vs distant relapse split: where does pembrolizumab add benefit?
- OS benefit vs deferring IO until relapse after MDT
- SABR dose/fractionation interaction with perioperative IO
📚 Sources · 📄 1 paper
Abstract
The longer read
The question PE-PE answers is narrower than its headline. Adjuvant pembrolizumab is already standard after nephrectomy for high-risk disease and for M1NED within a year of nephrectomy, the population KEYNOTE-564 defined. PE-PE carries that principle to pts whose metastases are still present but limited (≤3) and ablatable, and whose relapse may come up to 5 years after nephrectomy, well outside the 1-year M1NED window. Read that way, the result extends accepted biology rather than departing from it: anti-PD-1 given around the time all visible disease is eradicated delays relapse. The point estimate, HR 0.47 with mRFS 25.4 vs 11.0 mo, is large, but the upper CI bound of 0.92 shows how little information 35 events carry, and the true effect could plausibly be modest.
For a radiation oncologist the trial is functionally pembrolizumab plus SABR vs SABR alone, since SABR was the MDT in 73% vs 72% of pts. The abstract gives no dose, fractionation or target definition, and it withholds the two secondaries that would tell us where the benefit sits: local relapse and distant RFS. That split decides how the RT reader should use the result. If the gain is entirely at new distant sites, P is treating occult disease SABR never saw, SABR technique is unchanged, and the question is purely a systemic add-on. If in-field failure also drops, there is an interaction worth testing directly, and dose de-escalation or lesion selection questions open up. Until those endpoints are published, neither reading is supported.
Several design features should temper confidence. The trial is open-label, and the source does not say whether relapse was assessed centrally or blinded; imaging-defined RFS in an unblinded trial is sensitive to assessment timing. The 2:1 allocation leaves only 29 control pts, so the MDT-alone median of 11.0 mo rests on 16 events. Exposure was also below the nominal year, with a median of 6.6 of 9 cycles. On the toxicity side, any-grade AEs of 80% vs 21% and hypothyroidism in 23% are a permanent cost for pts who might have stayed relapse-free after MDT alone.
What the trial cannot answer is the comparison that matters most at the bedside: pembrolizumab with MDT vs MDT with IO deferred to relapse. An RFS gain without OS data, here 2 deaths in total, does not show that early IO beats IO at progression, and the MDT-alone pts will presumably receive systemic therapy when they relapse. For now this is a reason to discuss perioperative pembrolizumab with MDT-eligible pts and to watch for the local relapse and distant RFS data, not a settled standard.