Durvalumab with or without bevacizumab with transarterial chemoembolisation in hepatocellular carcinoma (EMERALD-1): a multiregional, randomised, double-blind, placebo-controlled, phase 3 study.
Sangro B, Kudo M, Erinjeri JP, Qin S, Ren Z, Chan SL, Arai Y, Heo J, Mai A, Escobar J, Lopez Chuken YA, Yoon JH, Tak WY, Breder VV, Suttichaimongkol T, Bouattour M, Lin SM, Peron JM, Nguyen QT, Yan L, Chiu CF, Santos FA, Veluvolu A, Thungappa SC, Matos M, Żotkiewicz M, Udoye SI, Kurland JF, Cohen GJ, Lencioni R, EMERALD-1 Investigators
- DOI
- 10.1016/S0140-6736(24)02551-0
- Record issued
- 2026-08-15
- Engine
- 7.39.0
- Exported
- 2026-09-20
Prepared by Alpha1. This document is confidential: it is intended for the recipient it was shared with and must not be redistributed. The live record at alpha1science.com/verify/eebee21a-9984-4d4c-a553-28f0596ebb4a is authoritative.
How this rating was calculated
- ClaimsEfficacy rests on an unvalidated surrogate endpoint−0.5★
- ClaimsTreatment effect not shown to be clinically meaningful−0.5★
- LinksDead data/code link−0.25★
- 01Efficacy rests on an unvalidated surrogate endpoint
The primary endpoint is progression-free survival (PFS), a surrogate for overall survival. The paper does not demonstrate target engagement at the tested dose (no PK/PD or dose-exposure data) and does not cite validated evidence linking PFS to overall survival in this setting. The authors acknowledge that overall survival data are not yet mature and that PFS may be confounded by competing risks.
“The primary endpoint was progression-free survival, by blinded independent central review (BICR), and per RECIST version 1.1, with durvalumab plus bevacizumab versus placebo alone in the intention-to-treat population”
- 02Treatment effect not shown to be clinically meaningful
The reported effect is a median PFS improvement from 8.2 to 15.0 months (HR 0.77). While statistically significant, the clinical meaningfulness is not anchored to a minimal clinically important difference or a validated threshold for PFS in hepatocellular carcinoma. The paper claims clinical relevance but does not provide an external anchor.
“Median progression-free survival was improved by 6·8 months versus the placebo plus TACE group, which had results that were consistent with previous reports.”
- 03Declared data/code link does not resolve
Dead link — nothing to verify.
“https://vivli.org/members/enquiries-about-studies-not-listed-on-the-vivli-platform/”
This Kaimen Rigor review uses Kaimen Rigor reviewers trained on a curated corpus of high-fidelity and retracted papers, with expert supervision and curation. It can still make mistakes; verify each finding against the source before relying on it.
This is a well-conducted and transparently reported phase 3 randomized controlled trial. The paper demonstrates strong scientific premise, rigorous design, and complete reporting across all eight rigor dimensions, with only minor copyedit issues and a single dead link in the data availability statement.
Both reviewers classified the study as interventional, and this was adopted. The evaluation covered all eight dimensions; several sub-criteria were marked not applicable due to the human clinical trial context (e.g., species/strain, housing, cell line authentication). The statistics verification checked only 2 tests (limited coverage), and the citation check found no retracted or unresolved references. The reproducibility check found 1 dead link among 4 checked.
Numerical inconsistencies
None foundValues that contradict each other or are impossible for the stated sample: recomputed p-values and test statistics, GRIM/GRIMMER checks on summary numbers, percentages against their own counts, totals against their parts, and estimates against their own confidence intervals.
Checked — nothing surfaced.
Recomputed 2 tests: 2 consistent, 0 inconsistent; 2 via agent-written checks.
- CONSISTENTreported p = .032 · recomputed p = .031Reviewers 1, 2Primary PFS HR for durvalumab+bevacizumab vs placebo
“Progression-free survival hazard ratio was 0·77 (95% CI 0·61–0·98; two-sided p=0·032) for durvalumab plus bevacizumab versus placebo.”
Taken as given: The HR is 0.77 with 95% CI 0.61-0.98.; The CI is two-sided at 95%.; The p-value is two-sided from a log-rank test, approximated by the CI-based method.Method: Recomputed two-sided p from HR and 95% CI using normal approximation on log scale.How we recomputed it: pCI(0.77, 0.61, 0.98, 1) - CONSISTENTreported p = .640 · recomputed p = .599Reviewers 1, 2Secondary PFS HR for durvalumab+placebo vs placebo
“Progression-free survival hazard ratio was 0·94 (0·75–1·19; two-sided p=0·64) for durvalumab plus placebo versus placebo.”
Taken as given: The HR is 0.94 with 95% CI 0.75-1.19.; The CI is two-sided at 95%.; The p-value is two-sided from a log-rank test, approximated by the CI-based method.Method: Recomputed two-sided p from HR and 95% CI using normal approximation on log scale.How we recomputed it: pCI(0.94, 0.75, 1.19, 1)
Overstated conclusions
3 findings · worst highConclusions that reach past what the paper's own results support — including a significance claim that no longer holds when the statistic is recomputed, and efficacy resting on an unvalidated surrogate endpoint.
- Efficacy rests on an unvalidated surrogate endpointAssessed
- Treatment effect not shown to be clinically meaningfulAssessed
- Conclusions only partially backed by the presented evidenceAssessed
5 major claims checked against the paper's own evidence: 1 only partially supported (evidence backs part of the claim; gaps or caveats remain); the rest adequately supported.
- partialReviewers 1, 2The combination has the potential to set a new standard of care.The claim is based on a single positive PFS result, but overall survival data are not yet mature; thus, the claim is somewhat forward-looking.Evidence: Positive PFS result and manageable safety profile.
“Durvalumab plus bevacizumab plus TACE has the potential to set a new standard of care.”
AbstractFind in source - supportedReviewer 1Durvalumab plus bevacizumab plus TACE significantly improved progression-free survival versus placebo plus TACE.The primary endpoint was met with HR 0.77 (95% CI 0.61-0.98, p=0.032), which is statistically significant.Evidence: Reported HR and p-value in Results.
Progression-free survival hazard ratio was 0·77 (95% CI 0·61–0·98; two-sided p=0·032) for durvalumab plus bevacizumab versus placebo.
Abstractreviewer’s wording - supportedReviewers 1, 2Durvalumab plus placebo did not significantly improve progression-free survival versus placebo.The HR was 0.94 (95% CI 0.75-1.19, p=0.64), not statistically significant.Evidence: Reported HR and p-value in Results.
Progression-free survival hazard ratio was 0·94 (0·75–1·19; two-sided p=0·64) for durvalumab plus placebo versus placebo.
Abstractreviewer’s wording - supportedReviewers 1, 2The safety profile of durvalumab and bevacizumab with TACE was manageable and consistent with known profiles.Safety data show expected adverse events with no new signals; treatment-related deaths were rare.Evidence: Safety analysis set results in Table 2 and text.
“In combination with TACE, the safety profile of durvalumab and bevacizumab was manageable and consistent with the known safety profiles of durvalumab, bevacizumab, and TACE in unresectable hepatocellular carcinoma and the underlying liver disease.”
Discussion ¶5Find in source - supportedReviewer 2Durvalumab plus bevacizumab plus TACE significantly improved progression-free survival compared with TACE alone.The primary endpoint was met with a statistically significant improvement in PFS (HR 0.77, p=0.032).Evidence: Primary endpoint result: median PFS 15.0 vs 8.2 months, HR 0.77 (95% CI 0.61-0.98), p=0.032.
Durvalumab plus bevacizumab significantly improved progression-free survival by BICR per RECIST version 1.1 versus placebo (median 15·0 months [95% CI 11·1–18·9] vs 8·2 months [6·9–11·1]; HR 0·77 [95% CI 0·61–0·98]; two-sided log-rank p=0·032).
Abstractreviewer’s wording
Premise concern: surrogate not validated for clinical benefit; effect size not shown to be clinically meaningful.
- INADEQUATESurrogate endpointThe primary endpoint is progression-free survival (PFS), a surrogate for overall survival. The paper does not demonstrate target engagement at the tested dose (no PK/PD or dose-exposure data) and does not cite validated evidence linking PFS to overall survival in this setting. The authors acknowledge that overall survival data are not yet mature and that PFS may be confounded by competing risks.
“The primary endpoint was progression-free survival, by blinded independent central review (BICR), and per RECIST version 1.1, with durvalumab plus bevacizumab versus placebo alone in the intention-to-treat population”
- INADEQUATEEffect sizeThe reported effect is a median PFS improvement from 8.2 to 15.0 months (HR 0.77). While statistically significant, the clinical meaningfulness is not anchored to a minimal clinically important difference or a validated threshold for PFS in hepatocellular carcinoma. The paper claims clinical relevance but does not provide an external anchor.
“Median progression-free survival was improved by 6·8 months versus the placebo plus TACE group, which had results that were consistent with previous reports.”
Data authenticity concerns
None foundAn adversarial read for patterns associated with data that may not be genuine: results that look too clean, implausibly large effects, duplicated data or images, and methods that do not match the results reported.
Checked — nothing surfaced.
Reporting gaps
None foundRequired detail the manuscript never states — study design, biological variables, ethics approval and consent, key resources, statistical reporting, data and code availability, and overall transparency.
Checked — nothing surfaced.
The introduction thoroughly reviews prior studies of TACE, sorafenib, brivanib, orantinib, bevacizumab, and immunotherapies, acknowledging both positive and negative results. The rationale for combining durvalumab and bevacizumab with TACE is logically derived from mechanistic evidence (VEGF and PD-L1 upregulation) and prior trial outcomes. The paper explicitly notes that no prior phase 3 study had shown significant improvement with systemic therapy added to TACE, and the study design addresses this gap.
“Randomised phase 2 studies with sorafenib, a multikinase inhibitor, have demonstrated the potential benefit of adding systemic therapies to TACE.”
“TACE increases vascular endothelial growth factor (VEGF) expression in hepatocellular carcinoma tumours, possibly allowing survival of residual tumour tissue after TACE.”
“However, to date, no phase 3 study has shown significant clinical improvements in any primary endpoint by adding systemic therapy to TACE.”
“However, to date, no phase 3 study has shown significant clinical improvements in any primary endpoint by adding systemic therapy to TACE.”
“Due to the aforementioned safety considerations, the study was designed to avoid concurrent bevacizumab and TACE (durvalumab was allowed with TACE, while bevacizumab was only started after completion of TACE).”
Randomization method (IxRS, blocked, stratified) and unit (participant) are reported. Blinding is described for participants, investigators, and outcome assessors. Power analysis is detailed with planned sample size and event-driven design. Inclusion/exclusion criteria are pre-specified. Outlier handling is addressed through pre-specified analysis populations (ITT, safety) and sensitivity analyses for missing data. Controls are inherent in the placebo arm. Independent replication is not applicable for a single pivotal trial.
“Participants were centrally assigned to their study treatment using an interactive voice or web response system (IxRS).”
“Participants, investigators, study centre staff, and those assessing outcomes were masked to treatment assignment until data analysis.”
“Participants were centrally assigned to their study treatment using an interactive voice or web response system (IxRS).”
“Participants, investigators, study centre staff, and those assessing outcomes were masked to treatment assignment until data analysis.”
Sex is reported for all participants (135 female, 481 male). Age is reported as median with IQR. Demographics include race and ethnicity. Health status is captured via ECOG performance status and Child-Pugh score. Species/strain and housing conditions are not applicable for a human trial.
“375 (61%) were Asian, 176 (29%) were White, 22 (4%) were American Indian or Alaska Native, nine (1%) were Black or African American, one (<1%) was native Hawaiian or other Pacific Islander, and 33 (5%) were other races.”
“375 (61%) were Asian, 176 (29%) were White, 22 (4%) were American Indian or Alaska Native, nine (1%) were Black or African American, one (<1%) was native Hawaiian or other Pacific Islander, and 33 (5%) were other races.”
The paper states that the trial protocol was approved by an Institutional Review Board or Independent Ethics Committee at each site, and written informed consent was obtained. Compliance with Good Clinical Practice, Declaration of Helsinki, and CIOMS guidelines is stated.
“Written informed consent was obtained from participants (or their legal representatives if the participant lacked capacity to consent) before participation.”
“Good Clinical Practice guidelines of the International Council on Harmonisation, and ethical considerations of the Declaration of Helsinki and Council for International Organisations of Medical Sciences, were followed.”
“Written informed consent was obtained from participants (or their legal representatives if the participant lacked capacity to consent) before participation.”
“Good Clinical Practice guidelines of the International Council on Harmonisation, and ethical considerations of the Declaration of Helsinki and Council for International Organisations of Medical Sciences, were followed.”
Durvalumab and bevacizumab are named with manufacturer (AstraZeneca implied) and dosing details. The statistical software SAS is identified. Antibodies, cell lines, mycoplasma, and organisms are not applicable as this is a clinical trial without wet-lab assays.
“Participants in the durvalumab plus bevacizumab group received TACE plus durvalumab (1500 mg intravenously, once every 4 weeks) followed by concurrent durvalumab (1120 mg once every 3 weeks) plus bevacizumab (15 mg/kg intravenously, once every 3 weeks).”
“SAS (version 9.1 or higher) was used for all analyses.”
“Participants in the durvalumab plus bevacizumab group received TACE plus durvalumab (1500 mg intravenously, once every 4 weeks) followed by concurrent durvalumab (1120 mg once every 3 weeks) plus bevacizumab (15 mg/kg intravenously, once every 3 weeks).”
“SAS (version 9.1 or higher) was used for all analyses.”
The primary analysis uses stratified log-rank test and Cox proportional hazards model, with HRs and 95% CIs. Exact p-values are given (e.g., p=0.032). Assumptions are handled by design (stratified analysis, sensitivity analyses). Data presentation includes Kaplan-Meier curves and forest plots. Mathematical plausibility is not applicable for large-N continuous outcomes.
“Primary and secondary progression-free survival endpoints were analysed using a stratified log-rank test adjusting for TACE modality (drug-eluting bead vs conventional), region (Japan vs the rest of Asia excluding Japan vs other regions), and portal vein invasion (Vp1 or Vp2 [±Vp1] vs none).”
“Progression-free survival hazard ratio was 0·77 (95% CI 0·61–0·98; two-sided p=0·032) for durvalumab plus bevacizumab versus placebo, and 0·94 (0·75–1·19; two-sided p=0·64) for durvalumab plus placebo versus placebo.”
“Progression-free survival hazard ratio was 0·77 (95% CI 0·61–0·98; two-sided p=0·032)”
The data sharing statement names Vivli as a managed-access platform and provides URLs for requesting data. This is adequate for patient-level data. Repository deposit and accession numbers are not applicable due to privacy. Code sharing is not applicable as no bespoke code is mentioned.
Trial registration number is provided (NCT03778957). Methods are detailed enough for replication. Limitations are explicitly discussed (e.g., regional differences, competing risks). Conclusions are appropriately cautious, noting ongoing overall survival follow-up. Funding and conflicts of interest are disclosed. Reporting guideline adherence is implied by the structured abstract and CONSORT-like flow diagram, though not explicitly named.
“This study is registered with ClinicalTrials.gov (http://ClinicalTrials.gov) , NCT03778957 (https://clinicaltrials.gov/ct2/show/NCT03778957) , and is closed to accrual.”
“Limitations of this study include that eligibility criteria and treatment practices for TACE differ by region and treatment site, and EMERALD-1 did not stratify randomisation by site or country.”
“This study was funded by AstraZeneca.”
“Limitations of this study include that eligibility criteria and treatment practices for TACE differ by region and treatment site, and EMERALD-1 did not stratify randomisation by site or country.”
“In conclusion, the EMERALD-1 study was positive: durvalumab plus bevacizumab plus TACE significantly improved progression-free survival compared with TACE alone, with no new adverse events identified in a heterogeneous population of participants with unresectable hepatocellular carcinoma amenable to embolisation.”
Registered (1 ID: ClinicalTrials.gov). No reporting guideline cited.
Broken references and links
1 finding · worst mediumReferences checked against Crossref, OpenAlex and Retraction Watch for retractions and resolvability, plus declared data and code links probed for whether they resolve to content matching the paper.
- Dead data/code linksRecomputed
Checked 30 references by DOI: 27 verified — 3 no DOI (shown, not verified).
- NO DOINivolumab (NIVO) plus ipilimumab (IPI) vs lenvatinib (LEN) or sorafenib (SOR) as first-line treatment for unresectable hepatocellular carcinoma (uHCC): first results from CheckMate 9DWNo DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
- NO DOIOutcomes of patients (pts) with hepatocellular carcinoma (HCC) treated with transarterial chemoembolization (TACE): global OPTIMIS final analysisNo DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
- NO DOISafety and efficacy of durvalumab plus bevacizumab in unresectable hepatocellular carcinoma: results from the phase 2 study 22No DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
4 data/code links checked; 3 live, 1 dead.
- datahttps://astrazenecagrouptrials.pharmacm.com/ST/Submission/DisclosureLIVEHTTP 200Resolves, but the content could not be matched to the paper.
- datahttp://www.vivli.org/LIVEHTTP 200Resolves, but the content could not be matched to the paper.
- datahttps://vivli.org/members/enquiries-about-studies-not-listed-on-the-vivli-platform/DEADHTTP 404Dead link — nothing to verify.
- datahttps://vivli.org/ourmember/astrazeneca/LIVEHTTP 200Resolves, but the content could not be matched to the paper.
Copyediting
5 minorWording, consistency and formatting errors that need correcting before submission.
No major wording or formatting errors. 5 minor suggestions below.
5 copyedit issues flagged: mostly typo, consistency, clarity.
- MINORtypoAbstract, Findings“With 1 additional follow-up of the EMERALD-1 study”→ Change to 'With additional follow-up of the EMERALD-1 study'The '1' appears to be a typographical error.
- MINORconsistencyResults, Safety“The most common maximum grade 3–4 adverse events were hypertension in participants who received durvalumab and bevacizumab (nine [6%] of 154 participants), anaemia in participants who received durvalumab and placebo (ten [4%] of 232 participants), and post-embolisation syndrome in participants who received placebo alone (eight [4%] of 200 participants).”→ Ensure consistent use of 'participants' vs 'patients' throughout.Minor stylistic inconsistency.
- MINORtypoAbstract, Findings“durvalumab plus placebo (n=207), or placebo alone (n=205; ITT population).”→ Consider rephrasing for clarity: 'durvalumab plus placebo (n=207), or placebo alone (n=205), comprising the ITT population.'Minor punctuation issue.
- MINORconsistencyResults, paragraph 3“durvalumab plus placebo alone did not significantly improve progression-free survival versus placebo (median 10·0 months [95% CI 9·0–12·7] vs 8·2 months [6·9–11·1]; HR 0·94 [95% CI 0·75–1·19]; two-sided log-rank p=0·64; ).”→ Remove the extra semicolon before the period.Punctuation error.
- MINORclarityMethods, Statistical analysis“At the final progression-free survival analysis, the threshold for significance of the primary objective was 0·044, based on the α spent at the progression-free survival interim analysis (2·27%) and the actual number of events at the final progression-free survival analysis”→ Add a period at the end of the sentence.Missing period.
The published work is robust and well-reported; an informed reader should weigh the minor reporting gaps (explicit CONSORT statement, dead link in data availability) and the fact that only a subset of statistics were machine-verified. No erratum is warranted for the core findings, but the dead link and minor copyedit issues could be corrected in a future revision or erratum.
- 1.HIGHdata codeFix the dead link in the data availability statement (reproducibility check found 1 dead link among 4 checked).A broken link undermines the stated data access route and could prevent readers from obtaining the data.
- 2.MEDIUMreportingExplicitly state adherence to CONSORT reporting guidelines in the Methods or a dedicated section.Both reviewers noted that reporting guideline adherence is implied but not explicitly stated, which is a minor transparency gap.
- 3.MEDIUMreportingAdd a note in the data sharing statement specifying the timeframe and conditions for data availability.The current statement lacks specific conditions or timelines, which would improve transparency for potential data requesters.
- 4.MEDIUMreportingClarify in the Discussion that the progression-free survival benefit is based on a single trial and that overall survival data are pending.This avoids overgeneralization and ensures conclusions are appropriately cautious.
- 5.MEDIUMreportingConsider including a CONSORT flow diagram in the main text (currently referenced as Figure 2) to improve clarity of participant disposition.A flow diagram in the main text enhances transparency and reader comprehension.
- 6.LOWcopyeditFix the typo in the Abstract, Findings: change 'With 1 additional follow-up' to 'With additional follow-up'.The '1' appears to be a typographical error that could confuse readers.
- 7.LOWcopyeditRemove the extra semicolon before the period in Results, paragraph 3: '...two-sided log-rank p=0·64; ).'Punctuation error that should be corrected for professional presentation.
- 8.LOWcopyeditAdd a period at the end of the sentence in Methods, Statistical analysis: '...at the final progression-free survival analysis'.Missing period at the end of the sentence.
- 9.LOWcopyeditRephrase the Abstract, Findings sentence for clarity: 'durvalumab plus placebo (n=207), or placebo alone (n=205; ITT population).'Minor punctuation issue that can be improved for readability.
- 10.LOWcopyeditEnsure consistent use of 'participants' vs 'patients' throughout the manuscript.Minor stylistic inconsistency noted in the Results, Safety section.
The star rating is the report’s one-glance summary. Every paper starts at 5★ and loses stars for the concrete problems the review finds — so a rating is never a vague average, it’s a running total you can read line by line under “How this rating was calculated.”
- Reporting — 8 dimensionseach dimension that fully fails−½★
- each dimension partially met−¼★
- Statistics · Integrity · Claimseach serious problem−1★
- each medium problem−½★
- Citationseach retracted or unverifiable reference−¼★
- Copyeditonly when the manuscript needs a full edit−½★
The rating never drops below 1★, and a demonstrable critical failure (an impossible statistic, a proven ethics violation) caps it at 1★ on its own — so the stars can never look healthy when the verdict is CRITICAL.
The rating draws on a panel of agents. Three independent Kaimen Rigor reviewers grade the eight dimensions below across several independent passes (the shown verdict is their majority vote — steadier than any single run), isolate the paper’s major claims and check its own evidence backs them, and flag integrity concerns. Alongside them, a citation agent resolves every reference against Crossref, OpenAlex, and Retraction Watch; a statistics agent recomputes reported tests; and rule-based checks verify that declared data/code links actually resolve. Full text is required — an abstract-only submission is not analyzed.
Graded against NIH, MDAR, ARRIVE 2.0, CONSORT, EQUATOR, and RRID guidelines. A dimension that doesn’t apply to the study type is skipped, never penalized.
This Kaimen Rigor review is model-assisted and is not a substitute for formal expert review. It complements human evaluation by surfacing potential methodological concerns — verify each finding against the source.