KEYNOTE-866 Key Takeaways
Design
Phase III, double-blind, randomized 1:1: pembrolizumab 200 mg or placebo every 3 weeks with cisplatin 70 mg/m² and gemcitabine 1000 mg/m² for 4 neoadjuvant cycles, radical cystectomy with pelvic lymph node dissection, then 13 cycles of pembrolizumab or placebo (n=453 vs 454). Primary endpoint: EFS by investigator assessment; key secondary: OS and pCR by central pathology review. (ESMO 2026 4735O, Bedke; data cutoff 26-Jan-2026)
Efficacy
EFS HR 0.88 (95% CI 0.71–1.10), one-sided P=0.1366 (boundary 0.0225) — not statistically significant; median EFS 74.1 months vs not reached. OS HR 1.00 (95% CI 0.79–1.27). pCR 44.4% vs 35.8% (estimated difference 8.9%). OS and pCR were not formally tested. (ESMO 2026 4735O, Bedke; data cutoff 26-Jan-2026)
Safety
Immune-mediated AEs and infusion reactions: any grade 28.9% vs 9.4%; grade 3–5 8.9% vs 1.8%. The presenters reported a manageable safety profile with no new safety signals. (ESMO 2026 4735O, Bedke; data cutoff 26-Jan-2026)
Regulatory status
⚠️ Perioperative pembrolizumab plus neoadjuvant gemcitabine-cisplatin is not approved for MIBC; the presented conclusion states it is not supported. ✅ Perioperative enfortumab vedotin plus pembrolizumab is FDA-approved for MIBC: November 21, 2025 for cisplatin-ineligible patients (KEYNOTE-905) and July 10, 2026 for cystectomy candidates regardless of cisplatin eligibility (KEYNOTE-B15). (FDA)
KOL pulse
Physicians posting from the session called the trial negative, noted that the higher pCR rate did not translate into EFS or OS, and asked why NIAGARA was positive. Several noted the strong outcomes in the placebo arm.
Track oncology’s top KOLs in your specialty
Create a free account, pick the tumor types you cover, and go beyond the tweet — the intelligence pharma teams use to map influence and prepare for engagement:
- ✓Enhanced KOL profiles
- ✓Pharma influence & potential advisory-board patterns
- ✓2025 Open Payments financial analysis
- ✓Social-media sentiment & trend signals
Top KOLs Discussing KEYNOTE-866
KEYNOTE-866 Key Slides & Visuals
KEYNOTE-866 Top Tweets
𝕏It hard to explain why 1 trial with gem/cis +/-PD(L)1 (NIAGARA) is +ve for EFS/OS and 2 very similar trials (presented today) are negative in MIBC. Potentially NIAGARA has the highest risk population, not all respond to IO and the maintenance phase the most effect- but the OS of gem/cis pembro HR of 1 is concerning. This reflects what we saw in metastatic disease with contradictory results for the big combination trials. Overall EVP looks better and more consistent #ESMO26 @myESMO @Annals_Oncology @OncoAlert
𝕏🚨 New ESMO 2026 data challenge our understanding of perioperative therapy in MIBC! KEYNOTE-866 (phase III, n=707), ENERGIZE (phase III, n=653): Adding perioperative pembro or nivolumab to neoadjuvant gemcitabine/cisplatin did not significantly improve EFS and OS, despite numerically higher pCR! Yet NIAGARA demonstrated significant EFS (HR 0.68) and OS (HR 0.75) benefits with perioperative durvalumab. Why does one chemoimmunotherapy strategy succeed while others fails? Different patient populations, trial designs, or underlying biology? We urgently need predictive biomarkers—not just another checkpoint inhibitor. Meanwhile, new KEYNOTE-905 data in cisplatin-ineligible MIBC (ad hoc, n=255) show: • Pembrolizumab alone: pCR 25.3%; EFS HR 0.71 vs surgery alone. • EV+pembrolizumab: pCR 58.8%; EFS HR 0.40 vs surgery alone. The key question is no longer simply whether to add immunotherapy—but which patients need chemotherapy, ADC-based therapy, or both. Histology, NECTIN4/HER2 biology, and longitudinal ctDNA deserve prospective evaluation as treatment-selection tools. Exciting times for MIBC—but precision must catch up with therapeutic progress. @OncoAlert Faculty @DrChoueiri , @tompowles1 , @apolo_andrea , @neerajaiims , @montypal , @Silke_Gillessen , @AOmlin , @brian_rini , @PGrivasMDPhD , @TiansterZhang , @amerseburger , @DrDanielHeng , @HHammersMD , @scocmem , @AndreaNecchi , @AmandaNizamMD , @drenriquegrande , @nataliagandur , @AlbigesL , @cdanicas , @declangmurphy , @seanmmcbride , @shilpaonc , @DrRanaMcKay , @scserendipity1 , @DrYukselUrun , @ravikanesvaran , @BraunMDPhD , @EfstathiouEleni , @mzugman , @crisbergerot
𝕏More pCR didn’t translate into EFS or OS. KEYNOTE-866: an important negative phase 3 in cisplatin-eligible MIBC. #ESMO26 #BladderCancer https://t.co/Ctmh4fcbF5
𝕏Negative Ph.III KEYNOTE-866 by #DrJensBedke: perioperative pembrolizumab + neoadjuvant Gem/Cis vs placebo + Gem/Cis in cis-eligible MIBC. EFS HR 0.88 (95% CI 0.71-1.10), p=0.137. pCR 44.4% vs 35.8%; OS HR 1.00. Higher pCR did not translate into EFS benefit. #ESMO26 #BladderCancer @myESMO @OncoAlert
𝕏Full house for Bladder/RCC updates first with #KN866 #ESMO26 Gem/cis in MIBC + placebo or pembro (➖) study & less relevance in the era of EVP SOC Vs similar trial NIAGARA was ➕ pCR 44.4% vs 35.8% EFS & OS not significant We wait #VOLGA tomorrow for more in this space https://t.co/uSLgxYSoMO
𝕏🚨⭐ #ESMO26 |Abstract 4735O KEYNOTE-866 🟠 Perioperative pembrolizumab or placebo + neoadjuvant cisplatin and gemcitabine in muscle-invasive bladder cancer: Phase III KEYNOTE-866 🎤 Presented by Jens Bedke 🇩🇪 @OncoAlert 🔬 Can adding perioperative pembrolizumab to cisplatin-based neoadjuvant chemotherapy improve outcomes in cisplatin-eligible MIBC? 📊 STUDY DESIGN Phase III, randomized, double-blind, placebo-controlled trial. 👥 N=907 patients with cisplatin-eligible MIBC. 🔹 Pembrolizumab + gemcitabine/cisplatin → radical cystectomy → adjuvant pembrolizumab (n=453) 🔹 Placebo + gemcitabine/cisplatin → radical cystectomy → adjuvant placebo (n=454) Median follow-up: 60.9 months. 📊 KEY RESULTS ❌ Event-free survival (PRIMARY ENDPOINT) • Median EFS: 74.1 months vs NR • HR 0.88 (95% CI 0.71–1.10) • p=0.1366 ➡️ No statistically significant EFS improvement. 📈 Pathological complete response (pCR) ✅ 44.4% vs 35.8% • Estimated difference: +8.9 percentage points (95% CI 2.5–15.1). ⚠️ Higher pCR did NOT translate into a statistically significant improvement in the primary EFS endpoint. 📊 Overall survival • Median OS: NR in both arms • HR 1.00 (95% CI 0.79–1.27) ➡️ No observed OS advantage. pCR and OS were not formally tested for statistical significance under the hierarchical testing strategy. 🏥 Surgery • Underwent cystectomy: 91.6% vs 87.9%. ⚠️ SAFETY • Grade 3–5 treatment-emergent AEs: 80.2% vs 75.1%. • Grade 3–5 immune-mediated AEs/infusion reactions: 8.9% vs 1.8%. 🔬 CONCLUSION KEYNOTE-866 did not meet its primary EFS endpoint. Despite increased pCR rates, perioperative pembrolizumab added to neoadjuvant cisplatin/gemcitabine did not demonstrate a significant EFS benefit. ⭐ MY POINT OF VIEW | CLINICAL PEARL A higher pathological response rate is encouraging, but it is not enough to redefine a standard of care. In curative-intent MIBC, durable disease control and survival must drive treatment decisions. 📌 These findings do not support routinely adding perioperative pembrolizumab to neoadjuvant gemcitabine/cisplatin in cisplatin-eligible MIBC. 💡 More treatment does not necessarily mean better outcomes. The real challenge is identifying which patients benefit from intensification—and which do not. 🌍 GU Oncology Community @urologysummit @gbanna74 @ravikanesvaran @PGrivasMDPhD @tompowles1 @apolo_andrea @sonpavde @montypal @DrChoueiri @neerajaiims @HHammersMD @TiansterZhang @drenriquegrande @crisbergerot @DrDanielHeng @Silke_Gillessen @EfstathiouEleni @BraunMDPhD @brian_rini @AOmlin @DrYukselUrun @yekeduz_emre @cdanicas @ecancer @UrologyTimes @urotoday @wearemedsir @OpenMedicineHQ @UCSFCancer #BladderCancer #MIBC #GUOncology #ClinicalTrials #OncoAlertAF
𝕏Excellent presentation by Jens presenting results from KN866 trial that did not meet its primary endpoint in contrast with NIAGARA. Peri-op EV-pembro is the SOC in MIBC, awaiting VOLGA trial tomorrow! @myESMO #ESMO26 #ESMOAmbassadors @OncoAlert @Uromigos @shilpaonc @AndreaNecchi https://t.co/pDpZgq4etv https://t.co/dB1E6AlrWF
𝕏KEYNOTE-866 did not show EFS not OS. Great presentation 👏🏼by Prof Jens Bedke #ESMO26 @OncoAlert @Uromigos @urotoday @Uroweb @DGUrologie @high5md https://t.co/h2WiciZeJF
𝕏🚨⭐🟠 #ESMO26 GU Oncology Invited Discussion 🎤 Prof. Robert Jones 🇬🇧 @OncoAlert 🔬 Why did NIAGARA succeed while KEYNOTE-866 and ENERGIZE failed to meet their primary efficacy endpoints? 📊 EFS HR: ✅ NIAGARA: 0.68 ❌ KEYNOTE-866: 0.88 ❌ ENERGIZE: 0.94 🧩 Different populations, baseline risks, eligibility criteria and trial designs complicate comparisons. ⭐ MY POINT OF VIEW Different outcomes across trials do not establish superiority of PD-L1 over PD-1 blockade. 💡 In curative-intent MIBC, patient selection, endpoint definitions and absolute clinical benefit matter as much as the HR. Cross-trial comparisons generate hypotheses, not definitive conclusions. 🌍 GU Oncology Community @gbanna74 @ravikanesvaran @urologysummit @tompowles1 @PGrivasMDPhD @apolo_andrea @sonpavde @AndreaNecchi @drenriquegrande @DrYukselUrun @yekeduz_emre @cdanicas @ecancer @UrologyTimes @urotoday @wearemedsir @OpenMedicineHQ @UCSFCancer #BladderCancer #UrothelialCancer #GUOncology #ADC #ClinicalTrials
𝕏🚨 A negative phase III trial can still sharpen the field. @OncoAlert At #ESMO26, KEYNOTE-866 tested whether adding perioperative pembrolizumab to neoadjuvant gemcitabine/cisplatin and radical cystectomy could improve outcomes in cisplatin-eligible MIBC. 📊 n=907 ➡️ EFS: 74.1 mo vs NR HR 0.88 (0.71–1.10), p=.1366 ➡️ pCR: 44.4% vs 35.8% ➡️ OS: HR 1.00 (0.79–1.27) ⚠️ Grade 3–5 TEAEs: 80.2% vs 75.1% The key point: perioperative pembrolizumab did not significantly improve EFS. 🎯 Why this matters: not every perioperative IO strategy adds meaningful benefit to cisplatin-based chemotherapy. For me, this is a reminder that the perioperative MIBC space cannot be treated as one interchangeable class effect. We need to understand: 🔹 which combinations truly improve long-term outcomes 🔹 whether pCR gains translate into survival benefit 🔹 which patients need intensification 🔹 and how biomarkers may help us avoid adding toxicity without enough efficacy More therapy is not automatically better perioperative therapy. Better selection matters. Important data from @garysteinbergmd @DrChrisHoimes and the KEYNOTE-866 investigators. Pinging @DrChoueiri @tompowles1 @apolo_andrea @PGrivasMDPhD @AndreaNecchi @DrYukselUrun @drenriquegrande @crisbergerot @niklas_kluemper @nataliagandur #OncoAlert #ESMO26 #BladderCancer #MIBC #Immunotherapy #PerioperativeTherapy #GUOncology
About the KEYNOTE-866 Trial
KEYNOTE-866 asked whether perioperative pembrolizumab improves outcomes when added to neoadjuvant cisplatin-gemcitabine and radical cystectomy in cisplatin-eligible MIBC. It is described by the investigators as the first placebo-controlled trial of perioperative PD-1/PD-L1 immunotherapy with neoadjuvant cisplatin-gemcitabine. Most patients had cT3/cT4aN0 disease at baseline (84.5% vs 84.8%); 14.8% vs 14.5% had cT2N0 disease.
Jens Bedke presented the final EFS analysis at ESMO 2026 (Abstract 4735O), with a concurrent paper in Annals of Oncology. The presented conclusions state that efficacy in the placebo arm surpassed conventional benchmarks for neoadjuvant cisplatin-gemcitabine and that perioperative enfortumab vedotin plus pembrolizumab has become a standard of care in MIBC.
Trial Methodology & Results
Study Design
Phase III, randomized 1:1, double-blind, placebo-controlled. (Study design slide; ClinicalTrials.gov NCT03924856)
Population
Cisplatin-eligible MIBC; ITT 453 vs 454. Baseline cT2N0 14.8% vs 14.5%, cT3/cT4aN0 84.5% vs 84.8%, cT1–4aN1 0.7% vs 0.7%; PD-L1 CPS ≥10 56.7% vs 56.2%. (Baseline slide)
Interventions
Pembrolizumab 200 mg or placebo Q3W + cisplatin 70 mg/m² + gemcitabine 1000 mg/m² x4 → RC + PLND → pembrolizumab or placebo Q3W x13.
Endpoints
Primary: EFS by investigator assessment. Key secondary: OS; pCR (pT0N0) by central pathology review. Other: safety.
Presenter
Jens Bedke — ESMO 2026 Proffered paper session, Genitourinary tumours, renal and urothelial, October 23, 2026, Abstract 4735O. Concurrent publication: Annals of Oncology (DOI 10.1016/j.annonc.2026.10.106).
Sponsor
Merck Sharp & Dohme LLC (ClinicalTrials.gov).
Survival and response, KEYNOTE-866 (ESMO 2026 4735O, Bedke; data cutoff 26-Jan-2026)
| Measure | Pembro + cis/gem (n=453) | Placebo + cis/gem (n=454) |
|---|---|---|
| EFS events | 157 (34.7%) | 168 (37.0%) |
| Median EFS | 74.1 mo (72.4–NR) | Not reached (67.7–NR) |
| 24-month EFS | 76.7% | 70.0% |
| 36-month EFS | 69.2% | 66.3% |
| 48-month EFS | 66.0% | 62.3% |
| 60-month EFS | 63.5% | 61.9% |
| OS events | 135 (29.8%) | 135 (29.7%) |
| Median OS | Not reached | Not reached |
| 24-month OS | 82.8% | 81.7% |
| 36-month OS | 76.8% | 76.6% |
| 48-month OS | 72.9% | 73.4% |
| 60-month OS | 70.8% | 70.9% |
| pCR (central review) | 44.4% (200/450) | 35.8% (162/453) |
EFS HR 0.88 (95% CI 0.71–1.10), one-sided P=0.1366, not significant (one-sided boundary 0.0225). OS HR 1.00 (95% CI 0.79–1.27), not formally tested. pCR estimated difference 8.9% (95% CI 2.5–15.1), not formally tested. Landmark values as labeled on the presented Kaplan-Meier slides.
Event-free survival (primary)
EFS events 157 (34.7%) vs 168 (37.0%). Median EFS 74.1 months (95% CI 72.4–NR) with pembrolizumab vs not reached (67.7–NR) with placebo. HR 0.88 (95% CI 0.71–1.10), one-sided P=0.1366; not statistically significant against a one-sided boundary of 0.0225. 24-month EFS 76.7% vs 70.0%; 60-month EFS 63.5% vs 61.9%. (ESMO 2026 4735O, Bedke; data cutoff 26-Jan-2026)
EFS HR 0.88 · P=0.1366 · not metSource: Bedke et al., Annals of Oncology (DOI 10.1016/j.annonc.2026.10.106) ↗
Overall survival and pCR (key secondary)
OS events 135 (29.8%) vs 135 (29.7%); median not reached in either arm; HR 1.00 (95% CI 0.79–1.27); 60-month OS 70.8% vs 70.9%. pCR by central pathology review 44.4% (200/450; 95% CI 39.8–49.2) vs 35.8% (162/453; 95% CI 31.3–40.4); estimated difference 8.9% (95% CI 2.5–15.1). Neither OS nor pCR was formally tested for statistical significance under the statistical design. (ESMO 2026 4735O, Bedke; data cutoff 26-Jan-2026)
OS HR 1.00 · pCR 44.4% vs 35.8%Source: ClinicalTrials.gov NCT03924856 · presented slides (KOL capture) ↗
Safety
Immune-mediated AEs and infusion reactions occurred in 130 (28.9%) vs 42 (9.4%) patients; grade 3–5 in 40 (8.9%) vs 8 (1.8%); serious in 31 (6.9%) vs 8 (1.8%); none led to death with pembrolizumab vs 1 (0.2%) with placebo. Hypothyroidism (10.4%) was the most frequent immune-mediated AE with pembrolizumab. Any subsequent therapy 18.8% vs 23.8%. (ESMO 2026 4735O, Bedke; data cutoff 26-Jan-2026)
Grade 3–5 immune-mediated AEs 8.9% vs 1.8%Source: presented safety slide (posted by @nataliagandur) ↗
KEYNOTE-866 in the News
KEYNOTE-866 FAQ
What is the KEYNOTE-866 trial?
KEYNOTE-866 (NCT03924856) is a Merck double-blind phase III trial in cisplatin-eligible muscle-invasive bladder cancer. Patients received pembrolizumab or placebo with 4 cycles of neoadjuvant cisplatin-gemcitabine, radical cystectomy with pelvic lymph node dissection, then 13 cycles of pembrolizumab or placebo (n=453 vs 454).
What did KEYNOTE-866 show at ESMO 2026?
The primary endpoint was not met: EFS HR 0.88 (95% CI 0.71-1.10), one-sided P=0.1366, with median EFS 74.1 months vs not reached (data cutoff January 26, 2026). OS HR was 1.00 (95% CI 0.79-1.27). pCR was 44.4% vs 35.8%. OS and pCR were not formally tested.
Is pembrolizumab with gemcitabine-cisplatin approved for MIBC?
No. Perioperative pembrolizumab with neoadjuvant gemcitabine-cisplatin is not approved for MIBC, and the presented conclusions state it is not supported. Perioperative enfortumab vedotin plus pembrolizumab is FDA-approved for MIBC, based on KEYNOTE-905 (cisplatin-ineligible, November 21, 2025) and KEYNOTE-B15 (regardless of cisplatin eligibility, July 10, 2026).
Why was NIAGARA positive when KEYNOTE-866 was negative?
The reason is not established. Physicians and the invited discussant pointed to differences in disease stage at entry, control-arm outcomes, endpoint definitions and trial design. NIAGARA (durvalumab) improved EFS (HR 0.68) and OS (HR 0.75); these are cross-trial observations, not a head-to-head comparison, and do not show that one antibody is superior.
What were the safety findings?
Immune-mediated adverse events and infusion reactions occurred in 28.9% of patients with pembrolizumab vs 9.4% with placebo, including grade 3-5 events in 8.9% vs 1.8%. The investigators reported a manageable safety profile with no new safety signals.
Key KOL Sentiments - KEYNOTE-866
Compiled and reviewed by the KOL Pulse research team, led by Brian Shields, Founder, KOL Pulse. Last updated October 23, 2026. Statistics carry their source and data-cut label in place; every quote is verbatim from the physician’s own post.