Resectable stage II-IIIB NSCLC (perioperative) — Merck Sharp & Dohme LLC. At ESMO 2026 (2519O), Dr. Marina Garassino presented an exploratory, post hoc biomarker analysis: the benefit of perioperative pembrolizumab over neoadjuvant chemotherapy was generally seen across T-cell–inflamed gene expression profile, tumor mutational burden and KRAS subgroups.
Discover KOL Sentiment on KEYNOTE-671 →Design - Phase 3 perioperative pembrolizumab + neoadjuvant chemo -> adjuvant pembrolizumab vs placebo, resectable stage II-IIIB NSCLC (NCT03425643).
EFS - 5-yr EFS 49.9% (95% CI 44.6-55.0) vs 26.5% (95% CI 21.7-31.5); HR 0.58 (95% CI 0.48-0.69). Median EFS 57.1 vs 18.4 mo. (Annals of Oncology, DCO 3 Jul 2025)
OS - 5-yr OS 64.6% (95% CI 59.5-69.2) vs 53.6% (95% CI 48.3-58.6); HR 0.74 (95% CI 0.59-0.92). Median OS not reached vs 70.7 mo. (Annals of Oncology, DCO 3 Jul 2025)
Earlier primary OS analysis, ESMO 2024: HR 0.72 (95% CI 0.56-0.93, P=0.0103) - a different data cut.
Safety - Grade >=3 AEs 44.9% vs 37.3%; AE discontinuation 13.0% vs 6.5%.
Regulatory - FDA approved October 2023.
Sponsor / drug - Merck (MSD); pembrolizumab (Keytruda).
Biomarkers - ESMO 2026 (2519O), exploratory post hoc analysis: benefit of perioperative pembrolizumab seen regardless of T-cell–inflamed GEP, TMB or KRAS status; details below. (2519O abstract, DCO 3 Jul 2025)
Compiled and reviewed by the KOL Pulse research team, led by Brian Shields, Founder, KOL Pulse. Last updated October 26, 2026.
Dr. Marina C. Garassino (University of Chicago) presented this analysis on October 26, 2026 in the ESMO 2026 Proffered Paper session on non-metastatic NSCLC and other thoracic malignancies (invited discussant: Jordi Remon). It is a post hoc, exploratory analysis of participants with an evaluable baseline tumor sample: 387 of 797 had RNA sequencing for the 18-gene T-cell–inflamed gene expression profile (GEP) and 459 had whole-exome sequencing for tumor mutational burden (TMB) and KRAS status. The presenter’s conclusions slide notes an ~50% biomarker-evaluable subset and a regional imbalance in the RNA-sequencing population, and says results should be interpreted with caution. (2519O abstract and slides)
Associations (nominal P values, α=0.05; 1-sided for the pembrolizumab arm, 2-sided for the placebo arm): T-cell–inflamed GEP was associated with event-free survival (P=0.021) and major pathological response (P=0.019) in the pembrolizumab arm, and with overall survival in the placebo arm (P=0.002). TMB was associated with event-free survival (P=0.000164), overall survival (P=0.005), major pathological response (P=0.001) and pathological complete response (P=0.000695) in the pembrolizumab arm, and with major pathological response (P=0.014) and pathological complete response (P=0.044) in the placebo arm. (2519O abstract)
| Biomarker subgroup (pembrolizumab vs placebo) | N | EFS HR (95% CI) | OS HR (95% CI) |
|---|---|---|---|
| T-cell–inflamed GEP low (<1st tertile) | 129 | 0.71 (0.46–1.11) | 0.59 (0.36–0.97) |
| T-cell–inflamed GEP high (≥1st tertile) | 258 | 0.53 (0.38–0.75) | 0.82 (0.55–1.22) |
| TMB low (<175 mutations/exome) | 266 | 0.63 (0.47–0.86) | 0.70 (0.49–1.01) |
| TMB high (≥175 mutations/exome) | 193 | 0.49 (0.32–0.73) | 0.67 (0.40–1.12) |
| KRAS mutation (nonsquamous, matched normal) | 86 | 0.73 (0.42–1.28) | 0.88 (0.45–1.75) |
| KRAS wild-type (nonsquamous, matched normal) | 169 | 0.75 (0.50–1.12) | 0.85 (0.49–1.47) |
Hazard-ratio point estimates were below 1 in every subgroup, with overlapping confidence intervals, and 8 of the 12 confidence intervals include 1; these subgroup analyses were not powered to test for interaction, and overall survival data in the subgroups are immature. The authors concluded that the benefit of perioperative pembrolizumab was observed regardless of these biomarkers, supporting its use in early-stage NSCLC. (2519O abstract table; slides: “HRs for subgroups are exploratory/underpowered”)
Exploratory: benefit seen regardless of T-cell–inflamed GEP, TMB or KRAS status
| Endpoint | Perioperative pembrolizumab (n=397) | Placebo + chemo (n=400) |
|---|---|---|
| Median EFS | 57.1 months (38.0–NR) | 18.4 months (14.8–22.1) |
| 12-month EFS | 73.8% (69.2–77.9) | 61.1% (56.1–65.8) |
| 24-month EFS | 61.2% (56.2–65.9) | 42.0% (37.0–46.9) |
| 36-month EFS | 56.1% (51.0–60.9) | 36.1% (31.3–40.9) |
| 48-month EFS | 52.1% (46.9–57.0) | 29.6% (25.0–34.4) |
| 60-month EFS | 49.9% (44.6–55.0) | 26.5% (21.7–31.5) |
| Events / participants | 191/397 | 273/400 |
| Hazard ratio (95% CI) | 0.58 (0.48–0.69) | — |
| Endpoint | Perioperative pembrolizumab (n=397) | Placebo + chemo (n=400) |
|---|---|---|
| Median OS | Not reached | 70.7 months (53.7–NR) |
| 12-month OS | 87.6% (84.0–90.5) | 87.7% (84.0–90.5) |
| 24-month OS | 78.8% (74.4–82.5) | 74.8% (70.2–78.8) |
| 36-month OS | 72.1% (67.4–76.3) | 65.6% (60.7–70.0) |
| 48-month OS | 67.4% (62.5–71.8) | 57.8% (52.8–62.6) |
| 60-month OS | 64.6% (59.5–69.2) | 53.6% (48.3–58.6) |
| Deaths / participants | 142/397 | 182/400 |
| Hazard ratio (95% CI) | 0.74 (0.59–0.92) | — |
Top tweets by impressions — click to view on X
Newest first; each post shown in full and verbatim.
Another presentation Dr. @marinagarassino from @UCCancerCenter @UChicagoMed with biomarker analysis of KEYNOTE-671 for resectable early stage NSCLC. https://t.co/7urBmqr9iR
🆙#ESMO26 Proffered Paper: #LCSM Non-Mets 🔥Association of Biomarkers With Efficacy in the Phase 3 KEYNOTE-671 Study of Perioperative Pembrolizumab (Pembro) Plus Neoadjuvant Chemotherapy (Chemo) in Early-Stage Non–Small-Cell Lung Cancer (NSCLC) 🎙️ @marinagarassino 🔢2519O 🎯Higher TMB Associated With Better EFS (p=0.000164) and OS (p=0.005) With Pembro 🎯Pembro Benefit Over Chemo Consistent Regardless of T-cell-inflamed GEP, TMB, or KRAS Status ☑️NCT03425643 🔗 https://t.co/grEfaUchKN @OncoAlert @Larvol @myESMO @IASLC
#ESMO26 The brilliant Dr. @marinagarassino with biomarker analysis from the phase III KEYNOTE 671 study of perioperative pembrolizumab for resectable NSCLC showing association of T-cell inflamed gene expression profile and TMB with efficacy. Benefit across subgroups; inflamed profile associated with better prognosis. #ESMOAmbassadors
Updated KEYNOTE-671 results: 5-year EFS 49.9% vs 26.5%; OS 64.6% vs 53.6% with periop pembro
Yet key questions remain: Who needs neoadjuvant-only vs periop therapy? Is the adjuvant phase needed—and for how long? Can PD-L1, pCR or ctDNA/MRD guide selection?
https://t.co/mb00TcgnAM
5y outcomes from phase III KEYNOTE-671: perioperative pembrolizumab vs neoadjuvant chemo for resectable NSCLC @Annals_Oncology.
5y EFS 49.9% vs 26.5% (HR 0.58) and 5y OS 64.6% vs 53.6% (HR 0.74) with no detriment to QoL with pembro vs placebo.
https://t.co/73LjhFfiqd
🚨 KEYNOTE-671: 5-year data confirm durable benefit of perioperative pembrolizumab in resectable NSCLC
Stage II–IIIB(N2) NSCLC
Neoadjuvant pembrolizumab + platinum-doublet chemo → surgery → adjuvant pembrolizumab
vs placebo + chemo → surgery → placebo.
📊 5-year outcomes
• EFS: 49.9% vs 26.5%
HR 0.58
• OS: 64.6% vs 53.6%
HR 0.74
Median EFS: 57.1 vs 18.4 months
Importantly, benefit persisted even in patients without pCR:
5-year EFS 50.0% vs 29.7%
HR 0.63
⚠️ Grade 3–5 treatment-related AEs: 45.2% vs 37.8%
🩺 Clinical verdict:
✅ PRACTICE CONFIRMING
Five-year follow-up firmly establishes the durability of perioperative pembrolizumab + chemotherapy for resectable early-stage NSCLC.
@ASCO @oncoalert
#NSCLC #LungCancer #KEYNOTE671 #Immunotherapy
🔥🚨 @OncoAlert Hot off the press.
Just published @Annals_Oncology
⭐️Five-Year Outcomes of #Keynote671
trial of #Perioperative #Pembrolizumab + chemo vs chemo alone in resectable stage II-IIIB non-small cell #LungCancer (#NSCLC).
✅⬆️5-year #EFS
49.9% (pembro arm) vs 26.5% (HR, 0.58)
✅⬆️5-year #OS
64.6% (pembro arm) vs 53.6% (HR, 0.74)
👇🏻
https://t.co/DPEblT4Rry
🔥KEYNOTE-671: 5-year outcomes of perioperative pembrolizumab for early-stage NSCLC
🆙 @Annals_Oncology @myESMO
🎯5-y EFS 49.9 vs 26.5% (HR 0.58, 95% CI, 0.48‒0.69)
🎯5-y OS 64.6 vs 53.6% (HR 0.74, 95% CI, 0.59‒0.92)
🎙 @HwakeleeMD
#LCSM @OncoAlert @Larvol
https://t.co/eaZb5WnfDz
Is it applicable for ANY patients?
But we will need to revise all of these thoughts in the wake of recent data presentations on AEGEAN, Neotorch, & upcoming #ASCO23 presentation of KEYNOTE-671.…
☑️#WCLC25 #LCSM Mini Oral Abstract🆙
🔥KEYNOTE-671 4-year update
🔥Perioperative Pembrolizumab in Non-Small Cell Cancer (NSCLC): 4-Year Outcomes by Nodal Status in the KEYNOTE-671 Study
🎙️…
🚦 pCR + MRD in periop NSCLC:
🔴 MRD+ = worst prognosis
▫️Will adj IO be enough?
🟡 MRD– / no pCR (~80%) = gray zone
▫️Who actually benefits (and is IO ‘right’)?
🟢 pCR + MRD– = best…
More news in early stage #lungcancer: @Merck announces Ph3 KEYNOTE-671 study results: Neoadj pembro/chemo ➡️ adju pembro ⬆️ EFS + ⬆️ MPR vs. neoadj chemo alone in stage II, IIIA, IIIB. #LCSM…
🆙 #ELCC26 @myESMO 🇩🇰
🔥Mini Oral session 2
☑️KEYNOTE-671: 5 years F/U in pCR subgroup
🎯EFS HR 0.37, 5-y rate 80.8% vs. 55.7%
🎙️Dr. Margarita Majem
🎙️Chair: Dr. Kersti Oselin…
Today was the first time I observed @SalmaJabbour1 in a debate. It wouldn’t have mattered who was on the other side as she was absolutely spectacular using data and logic to clarify CRT is often the…
🆙 #ELCC26 @myESMO 🇩🇰
🔥Mini Oral session 2
☑️KEYNOTE-671: 5 years F/U in non-pCR subgroup
🎯EFS HR 0.69, 5-y rate 42.9% vs. 25.2%, OS immature
🎙️ @MartinReck2
🎙️Chair: Dr. Kersti Oselin…
🔥 Does pCR translate into long-term benefit? #ELCC26
5-year outcomes from KEYNOTE-671 in early-stage NSCLC answer this 👇
🧬 Study population
• Resectable stage II–IIIB (N2) NSCLC
• Perioperative…
@SalmaJabbour1 and I are going all in for our debate on stage III lung cancer today at #TexasLung26 - it will be a ton of fun and sure to entertain!
The only winner will be our patients who have two…
#ELCC26 @MartinReck2 on KN671 exploratory analysis of patients without pCR
5-year EFS 43% vs 25% , HR =0,69, 95%CI 0.57-0.83) https://t.co/mFvqfzt6Wu
KEYNOTE-671 demonstrated both EFS and OS benefits with perioperative pembrolizumab + chemotherapy, leading to FDA approval October 2023. Alongside AEGEAN (durvalumab), CheckMate-77T (nivolumab), and CheckMate-816 (neoadjuvant-only nivolumab), establishes perioperative immunotherapy as a core stage II-IIIB NSCLC strategy. Choice among regimens individualized by clinical and biomarker factors.
5-year analysis (Annals of Oncology, data cutoff 3 Jul 2025; median follow-up 60.4 months, range 42.6-85.8). Median EFS 57.1 months (95% CI 38.0-NR) with perioperative pembrolizumab vs. 18.4 months (95% CI 14.8-22.1) with placebo. Five-year EFS 49.9% (95% CI 44.6-55.0) vs. 26.5% (95% CI 21.7-31.5); HR 0.58 (95% CI 0.48-0.69), with 191/397 vs. 273/400 events. Benefit was consistent across prespecified subgroups including stage, nodal status, histology and PD-L1 TPS.
Primary analysis (NEJM 2023). Median EFS not reached vs. 17.0 months; HR 0.58 (95% CI 0.46-0.72), P<0.001. pCR 18.1% vs. 4.0%; mPR 30.2% vs. 11.0%.
Pathologic-response subgroups (ELCC 2026 update). pCR subgroup 5-yr EFS 80.8% vs. 55.7% (HR 0.37); non-pCR subgroup 5-yr EFS 42.9% vs. 25.2% (HR 0.69) - durable benefit even without a pathologic complete response.
5-year analysis (Annals of Oncology, data cutoff 3 Jul 2025). Median OS not reached with perioperative pembrolizumab vs. 70.7 months (95% CI 53.7-NR) with placebo. Five-year OS 64.6% (95% CI 59.5-69.2) vs. 53.6% (95% CI 48.3-58.6); HR 0.74 (95% CI 0.59-0.92), with 142/397 vs. 182/400 deaths. No detriment to health-related quality of life was identified with longer follow-up, and safety was consistent with the known profile of each treatment.
Primary OS analysis (ESMO 2024) - a different data cut. HR 0.72 (95% CI 0.56-0.93), P=0.0103, a 28% reduction in the risk of death. The 0.72 and 0.74 hazard ratios come from different cutoffs and should not be quoted interchangeably.
Grade ≥3 adverse events: 44.9% (pembro) vs. 37.3% (placebo). Discontinuation due to AEs: 13.0% (pembro) vs. 6.5% (placebo). Key AEs: neutropenia / chemo-related cytopenia, immune-related AEs (thyroid, skin). Treatment-related deaths 0.8% (pembro) vs. 0.3% (placebo). AE profile consistent with chemotherapy + checkpoint inhibitor combinations.
✅ Perioperative pembrolizumab established as SOC for resectable stage II-IIIB NSCLC. KEYNOTE-671 demonstrated both EFS and OS benefits with perioperative pembrolizumab + chemotherapy, leading to FDA approval October 2023. Alongside AEGEAN (durvalumab), CheckMate-77T (nivolumab), and CheckMate-816 (neoadjuvant-only nivolumab), establishes perioperative immunotherapy as a core stage II-IIIB NSCLC strategy. Choice among regimens individualized by clinical and biomarker factors.
In an exploratory, post hoc analysis presented at ESMO 2026 (abstract 2519O; data cutoff July 3, 2025), higher T-cell-inflamed gene expression profile and tumor mutational burden were associated with certain efficacy outcomes in both treatment arms (including the placebo arm), and event-free and overall survival hazard-ratio point estimates favored perioperative pembrolizumab over neoadjuvant chemotherapy regardless of gene expression profile, tumor mutational burden or KRAS status. Only about half of participants had biomarker data, and the subgroup hazard ratios were not powered to test for interaction. See At a Glance.
KEYNOTE-671 is a Phase 3 randomized trial (NCT03425643) of perioperative pembrolizumab (Keytruda) - neoadjuvant pembrolizumab plus platinum-based chemotherapy followed by adjuvant pembrolizumab - versus neoadjuvant chemotherapy plus placebo in patients with resectable stage II-IIIB non-small cell lung cancer. Event-free survival and overall survival were the dual primary endpoints.
At the 5-year analysis published in Annals of Oncology on August 21, 2026 (data cutoff July 3, 2025; median follow-up 60.4 months), 5-year event-free survival was 49.9% (95% CI 44.6-55.0) with perioperative pembrolizumab versus 26.5% (95% CI 21.7-31.5) with placebo (HR 0.58; 95% CI 0.48-0.69), and 5-year overall survival was 64.6% (95% CI 59.5-69.2) versus 53.6% (95% CI 48.3-58.6) (HR 0.74; 95% CI 0.59-0.92). Median EFS was 57.1 versus 18.4 months and median OS was not reached versus 70.7 months. Separately, at the earlier ESMO 2024 primary overall-survival analysis the OS hazard ratio was 0.72 (95% CI 0.56-0.93, P=0.0103). These are different data cuts and should not be quoted interchangeably.
No. In the 5-year analysis published in Annals of Oncology (August 21, 2026), no detriment to health-related quality of life was identified in the pembrolizumab arm versus the placebo arm with longer follow-up, and safety was consistent with the known profiles of each treatment.
Yes. In October 2023 the FDA approved pembrolizumab (Keytruda) with neoadjuvant platinum-containing chemotherapy and then continued as a single agent as adjuvant treatment for resectable (tumors >=4 cm or node-positive) NSCLC, based on KEYNOTE-671.
Grade >=3 adverse events occurred in 44.9% of pembrolizumab patients versus 37.3% with placebo, and treatment discontinuation due to adverse events was 13.0% versus 6.5%. Treatment-related deaths were uncommon (0.8% vs 0.3%). The profile is consistent with a chemotherapy-plus-checkpoint-inhibitor combination, with immune-related events (thyroid, skin) among the notable toxicities.
KEYNOTE-671 helped establish perioperative immunotherapy as a core approach for resectable stage II-IIIB NSCLC, alongside AEGEAN (durvalumab), CheckMate-77T (nivolumab), and the neoadjuvant-only CheckMate-816 (nivolumab). It is distinguished by demonstrating both an EFS and a statistically significant OS benefit.