Hepatocellular Carcinoma (HCC)
KEYNOTE-937
About the KEYNOTE-937 Trial
Table of Contents
Major Presentations and Milestones
KEYNOTE-937 Trial design, results, and conclusions
KEYNOTE-937 Sentiments and Criticisms
KEYNOTE-937 Temporal Sentiment Arc
Professional Resources : Interactive Tweet History, Influence Diagram, Sentiment Table, AI Chatbot
KEYNOTE-937 Trial: Major Presentations and Milestones
Primary speakers driving the story
At ASCO GI 2026 (#GI26), multiple GI and liver cancer KOLs amplified the phase III KEYNOTE-937 readout evaluating adjuvant pembrolizumab versus placebo after curative-intent therapy for hepatocellular carcinoma (HCC). Jun Gong, MD highlighted the on-site presentation by Stephen L. Chan (Chinese University of Hong Kong) and emphasized that the study did not meet its primary endpoint, aligning with the broader pattern of negative adjuvant IO trials in HCC.
Dr. Chan @CUHKofficial rand PhIII KN-937 trial of adj #pembrolizumab X1 year in #HCC after surgery or local ablation ➡️ no sig improvement in #RFS (primary endpoint) vs placebo (HR 1.06), no diff in OS or met-free survival. Follows negative IMbrave050 trial @OncoAlert #GI26 https://t.co/fTm1RpkNFE
— Jun Gong (@jgong15) January 9, 2026
In parallel, European liver cancer voices underscored the clinical disappointment and the need to redirect research efforts toward earlier (neoadjuvant) strategies. Arndt Vogel, MD, emphasized the lack of separation in recurrence-free survival despite long follow-up.
Keynote-937 phs-3: pembrolizumab for HCC after surgical resection or local ablation #ASCOGI26 👉mRFS: 46 vs 45 mo 👈4-yr RFS rate: 50 vs 50% 🤔Very disappointing, but inline with IMBRAVE-050, now we need to invest into neoadj. strategies @myesmo @ASCO @EASLedu @ILCAnews https://t.co/EoOx5jOgB2
— Arndt Vogel (@ArndtVogel) January 9, 2026
By the end of the meeting cycle, the discussion consolidated around a pragmatic conclusion: in resected/ablated HCC with complete radiographic response, KEYNOTE-937 did not support adoption of adjuvant pembrolizumab in routine practice.
KEYNOTE-937 Trial Design, Results, and Conclusions
Trial Design:
KEYNOTE-937 is a randomized phase III trial in HCC evaluating adjuvant pembrolizumab for 1 year versus placebo in patients who achieved a complete radiologic response after curative-intent therapy (surgical resection or local ablation). KOL commentary consistently framed the study as a test of whether checkpoint inhibition could reduce recurrence after “curative” therapy in a population frequently characterized by underlying liver disease.
Primary Results:
Across independent KOL summaries, the primary endpoint (recurrence-free survival, RFS) was not improved with pembrolizumab:
RFS: 46.7 months vs 45.5 months; HR 1.06; P=0.72 (as reported by Nieves Martinez Lago, MD, PhD).
#GI26 KEYNOTE-937 (phase III) 🔀 Adj HCC (PEM vs PBO) after complete radiologic response (resection/ablation) 📉 RFS: 46.7 vs 45.5 mo; (HR 1.06; P=0.72) 📉 OS: 48-mo OS 79% vs 81% 🛡️ Higher grade ≥3 AEs with pembrolizumab ➡️ Adj PEM does not improve outcomes in this setting https://t.co/h6Lp1ZL19S
— Nieves Martinez Lago MD PhD (@DraMartinezLago) January 9, 2026
Overall survival was also not improved in the reported summaries. Cathy Eng, MD noted: “No difference in OS (HR 1.06)” and that the study “Did not proceed to final analysis as the RFS hypothesis was not met.”
Safety:
Safety signals were not described in granular detail in the available tweet set, but Martinez Lago reported higher grade ≥3 adverse events with pembrolizumab versus placebo—an important consideration given the competing risks in an HCC population often affected by cirrhosis and limited hepatic reserve.
Key Conclusions:
Based on the #GI26 KOL discussion captured here, KEYNOTE-937 was interpreted as a negative adjuvant immunotherapy trial in HCC, with no RFS or OS advantage and higher high-grade toxicity. Multiple commentators explicitly concluded there is no current role for adjuvant pembrolizumab in this post-resection/ablation setting, and several called for alternative approaches (including neoadjuvant strategies and novel trial designs tailored to HCC biology and cirrhosis).
KEYNOTE-937 Sentiments and Criticisms
Consensus clinical takeaway (practice impact):
Cathy Eng, MD: “No role for adjuvant IO” https://x.com/CathyEngMD/status/2009674543032402014
Comparative framing (pattern across adjuvant HCC IO trials):
Daneng Li, MD: “keynote 937 negative study for adjuvant #HCC in line with prior results from IMbrave 050.” https://x.com/DanengLi/status/2009676073798381717
Arndt Vogel, MD: “Very disappointing, but inline with IMBRAVE-050, now we need to invest into neoadj. strategies” https://x.com/ArndtVogel/status/2009674004349759938
KEYNOTE-937 Temporal Sentiment Arc
2021 (Anticipation of an adjuvant HCC IO signal)
Primary/KOL tweets:
- https://x.com/pashtoonkasi/status/1401625170947739656
- Tone: Forward-looking interest in whether adjuvant immunotherapy successes in other tumor types could translate to GI malignancies, specifically HCC.
- Shift: The conversation was framed as “what’s next” for adjuvant IO expansion into HCC.
2026 (ASCO GI 2026 readout: consolidation around a negative result)
Primary/KOL tweets:
- https://x.com/DraMartinezLago/status/2009673655937032502
- https://x.com/jgong15/status/2009674274819453235
- https://x.com/CathyEngMD/status/2009674543032402014
- https://x.com/DanengLi/status/2009676073798381717
- Tone: Predominantly disappointed but clinically decisive—multiple KOLs described no RFS/OS benefit and highlighted that negative trials still meaningfully shape practice.
- Shift: The discussion moved from “will adjuvant IO work in HCC?” to “why are adjuvant IO trials in HCC failing, and what trial strategies should come next (e.g., neoadjuvant approaches, better risk selection,?”
Across the full arc, KEYNOTE-937 evolved from a “high-interest” adjuvant IO question (2021) to a practice-informing negative readout (2026), reinforcing a broader community reassessment of how to design perioperative immunotherapy strategies in HCC.
KEYNOTE-937 Professional Resources
