Trastuzumab deruxtecan versus trastuzumab emtansine in HER2-positive metastatic breast cancer: long-term survival analysis of the DESTINY-Breast03 trial.
Cortés J, Hurvitz SA, Im SA, Iwata H, Curigliano G, Kim SB, Chiu JWY, Pedrini JL, Li W, Yonemori K, Bianchini G, Loi S, Borges GS, Wang X, Bachelot T, Nakatani S, Ashfaque S, Liang Z, Egorov A, Hamilton E
- DOI
- 10.1038/s41591-024-03021-7
- Record issued
- 2026-08-16
- 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/f8dcff95-c9fa-483b-8586-0d4cd5d73254 is authoritative.
How this rating was calculated
- CitationsUnresolved reference−0.25★
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 rigorously conducted and transparently reported phase 3 randomized controlled trial with strong scientific premise, detailed methods, and appropriate statistical analysis. Minor reporting gaps exist (explicit power analysis, outlier handling, exact p-values for exploratory analyses, and statistical assumption verification), but none undermine the validity of the findings. The paper is well-suited for publication and the post-publication robustness is high.
Both reviewers independently scored all eight dimensions and agreed on the status of each, with only minor sub-criterion differences in study design and statistical analysis. The study type is interventional (phase 3 RCT). Non-applicable sub-criteria (e.g., animal housing, cell line authentication) were excluded. The statistics verification recomputed 5 reported tests/effect estimates, all consistent; coverage is limited to tests with test statistics/df or effect estimates with CIs, so other statistics remain unverified. The citation check found 1 reference not found in registry (potential fabrication signal), which is addressed in the action items.
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 5 tests: 5 consistent, 0 inconsistent; 5 via agent-written checks.
- CONSISTENTreported p = .014 · recomputed p = .017Reviewer 1Check HR for OS from reported 95% CI
“median OS was 52.6 versus 42.7 months (HR, 0.73; 95% CI, 0.56–0.94)”
Taken as given: The HR is 0.73 with 95% CI 0.56-0.94.; The CI is two-sided at 95%.; The HR is on a log scale.Method: Compute p-value from HR and 95% CI using the pCI function with log=1.How we recomputed it: pCI(0.73, 0.56, 0.94, 1) - CONSISTENTreported p < .001 · recomputed p = <.001Reviewer 1Check HR for PFS from reported 95% CI
“median PFS by investigator assessment was 29.0 versus 7.2 months (hazard ratio (HR), 0.30; 95% confidence interval (CI), 0.24–0.38)”
Taken as given: The HR is 0.30 with 95% CI 0.24-0.38.; The CI is two-sided at 95%.; The HR is on a log scale.Method: Compute p-value from HR and 95% CI using the pCI function with log=1.How we recomputed it: pCI(0.30, 0.24, 0.38, 1) - CONSISTENTreported p = .014 · recomputed p = .017Reviewer 2OS HR from abstract
“median OS was 52.6 versus 42.7 months (HR, 0.73; 95% CI, 0.56–0.94) with T-DXd versus T-DM1, respectively.”
Taken as given: The HR is 0.73; The 95% CI is (0.56, 0.94); The CI is two-sided at 95%; The HR is on the log scale (ratio)Method: p-value from HR and 95% CI using the formula for a log-normal distribution of the log-HR.How we recomputed it: pCI(0.73, 0.56, 0.94, 1) - CONSISTENTreported p < .001 · recomputed p = <.001Reviewer 2PFS HR from abstract
“median PFS by investigator assessment was 29.0 versus 7.2 months (hazard ratio (HR), 0.30; 95% confidence interval (CI), 0.24–0.38)”
Taken as given: The HR is 0.30; The 95% CI is (0.24, 0.38); The CI is two-sided at 95%; The HR is on the log scale (ratio)Method: p-value from HR and 95% CI using the formula for a log-normal distribution of the log-HR.How we recomputed it: pCI(0.30, 0.24, 0.38, 1) - CONSISTENTreported p < .001 · recomputed p = <.001Reviewer 2PFS2 HR from results
“Median PFS2 ... was 45.2 months ... with T-DXd and 23.1 months ... with T-DM1 (HR, 0.53; 95% CI, 0.41–0.68)”
Taken as given: The HR is 0.53; The 95% CI is (0.41, 0.68); The CI is two-sided at 95%; The HR is on the log scale (ratio)Method: p-value from HR and 95% CI using the formula for a log-normal distribution of the log-HR.How we recomputed it: pCI(0.53, 0.41, 0.68, 1)
Overstated conclusions
1 finding · worst lowConclusions 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.
- 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, 2Median OS with T-DXd is the longest reported in this disease setting.The claim is partially supported by cross-trial comparisons (e.g., CLEOPATRA, EMILIA), but the paper acknowledges that cross-trial comparisons should be interpreted cautiously due to differences in study design and post-trial therapies.Evidence: Discussion: 'To our knowledge, the median OS with T-DXd in DESTINY-Breast03 is the longest reported OS in this disease setting (median OS, 52.6 months; median follow-up, 43 months).'
“To our knowledge, the median OS with T-DXd in DESTINY-Breast03 is the longest reported OS in this disease setting (median OS, 52.6 months; median follow-up, 43 months).”
Discussion ¶3Find in source - supportedReviewers 1, 2T-DXd demonstrated superior efficacy over T-DM1 with longer follow-up.The claim is supported by the reported HRs and CIs for PFS, OS, and PFS2.Evidence: HR for PFS 0.30 (95% CI 0.24-0.38), HR for OS 0.73 (95% CI 0.56-0.94), HR for PFS2 0.53 (95% CI 0.41-0.68).
“With longer follow-up, T-DXd continued to demonstrate superior efficacy over T-DM1 with a manageable safety profile.”
AbstractFind in source - supportedReviewers 1, 2The safety profile of T-DXd is manageable with no cumulative toxicities.The claim is supported by the safety data showing no new safety signals and manageable rates of adverse events.Evidence: No new instances of grade ≥3 ILD or pneumonitis; EAIRs lower with T-DXd for any-grade TEAEs.
“The safety profile of T-DXd continues to be manageable with no cumulative toxicities observed with longer follow-up.”
DiscussionFind in source - supportedReviewer 2T-DXd provides a clinically meaningful improvement in efficacy over T-DM1.The claim is supported by the magnitude of improvement in PFS (HR 0.30), OS (HR 0.73), ORR (78.9% vs 36.9%), and DoR (30.5 vs 17.0 months), all with non-overlapping CIs.Evidence: Table 2: PFS HR 0.30 (95% CI 0.24–0.38), OS HR 0.73 (95% CI 0.56–0.94), ORR 78.9% vs 36.9%.
“T-DXd continued to demonstrate clinically meaningful improvement in efficacy compared with T-DM1 and a manageable safety profile that was consistent with previous results”
Discussion ¶1Find in source - supportedReviewer 2The benefit of T-DXd is supported by PFS2, suggesting better outcomes when used before T-DM1.The claim is supported by the PFS2 HR of 0.53 (95% CI 0.41–0.68), showing a significant delay in progression on next-line therapy.Evidence: Results: 'Median PFS2 ... was 45.2 months ... with T-DXd and 23.1 months ... with T-DM1 (HR, 0.53; 95% CI, 0.41–0.68).'
“The median PFS2 by investigator assessment with T-DXd was approximately twice as long as that with T-DM1 in this updated analysis, which suggests that patients may derive a better clinical benefit when treated with T-DXd before T-DM1.”
Discussion ¶3Find in source
Efficacy claim is anchored to an adequate endpoint and a meaningful effect.
- ADEQUATESurrogate endpointThe primary efficacy claim is based on overall survival (OS), a hard clinical outcome, and progression-free survival (PFS), which is a validated surrogate for metastatic breast cancer. The paper reports a median OS of 52.6 vs 42.7 months (HR 0.73) and median PFS of 29.0 vs 7.2 months (HR 0.30). OS is a direct clinical outcome, and PFS is an established surrogate in this setting. The paper also reports target engagement indirectly through the mechanism of action of T-DXd, but the primary basis is OS, which is adequate.
“Median OS was 52.6 months (95% CI, 48.7 months to NE) with T-DXd and 42.7 months (95% CI, 35.4 months to NE) with T-DM1; the risk of death was reduced by 27% (HR, 0.73; 95% CI, 0.56–0.94)”
- ADEQUATEEffect sizeThe effect sizes are large and clinically meaningful: median OS improvement of ~10 months (52.6 vs 42.7 months) and median PFS improvement of ~22 months (29.0 vs 7.2 months), with HRs of 0.73 and 0.30, respectively. These are anchored to clinical meaningfulness as the authors describe them as 'clinically meaningful' and note the OS is the longest reported in this setting. The effect sizes are statistically supported with 95% CIs.
“The median PFS and ORR by investigator assessment reinforced the clinical benefit of T-DXd over T-DM1 and were consistent with the analysis at the previous data cutoff. Median OS was reached in both treatment groups in this updated analysis, with an approximate 10-month improvement over T-DM1 observed with T-DXd and a reduction in the risk of death by approximately 27%”
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 paper cites numerous prior studies establishing HER2 as a therapeutic target and the development of T-DXd and T-DM1. It acknowledges strengths and weaknesses of prior work, including the EMILIA trial and CLEOPATRA trial. The rationale for comparing T-DXd to T-DM1 is logically presented based on their differing mechanisms and drug-to-antibody ratios. Limitations of prior analyses (e.g., median OS not reached) are explicitly addressed as motivation for this updated analysis.
“We report on an exploratory analysis of DESTINY-Breast03 (data cutoff, 20 November 2023), with updated efficacy, including median OS, and safety data with longer follow-up.”
“Longer follow-up is needed to determine a more precise estimate of the median OS in the T-DXd group due to the number of patients censored; more mature data are expected at the next data cutoff as the study continues.”
“The discovery of HER2 alterations led to the development of treatments that specifically target HER2, resulting in improved prognosis for patients with this subtype of breast cancer”
“T-DXd has a high, homogeneous drug-to-antibody ratio of approximately 8, while T-DM1 has a drug-to-antibody ratio of approximately 3.5”
“However, median OS was not reached in either treatment group at the primary analysis or the second OS interim analysis”
The study is an open-label, multicenter, phase 3 trial with balanced block randomization and stratification factors. The randomization method and unit are described. Blinding of patients and investigators was not possible, but tumor assessments were performed by BICR, which is an acceptable alternative. The sample size was pre-specified (approximately 500 patients), and inclusion/exclusion criteria are detailed. The analysis populations (full analysis set and safety analysis set) are defined, and the handling of missing data is implicit in the survival analysis methods.
“Randomization of patients involved balanced block randomization with a 1:1 allocation ratio for T-DXd and T-DM1.”
“The study aimed to enroll approximately 500 patients, with random assignment determined using EAST software version 6.4.”
“Randomization of patients involved balanced block randomization with a 1:1 allocation ratio for T-DXd and T-DM1.”
“Due to distinct administration protocols and adverse event profiles of the treatments, blinding of patients and investigators was not possible. However, tumor assessments were performed by BICR, which were previously reported”
“The study aimed to enroll approximately 500 patients, with random assignment determined using EAST software version 6.4.”
Sex is reported (99.6% female in both groups). Age is reported with median and range. Demographics including race, ethnicity, region, ECOG PS, HER2 IHC status, hormone receptor status, and CNS/liver/lung metastases are all detailed in Table 1. Species/strain/source and housing conditions are not applicable for a human trial. Sex justification is not applicable as both sexes are enrolled.
“Female | 260 (99.6) | 262 (99.6) | | Male | 1 (0.4) | 1 (0.4)”
“The median age was 54.3 years (range, 27.9–83.1 years) in the T-DXd group and 54.2 years (range, 20.2–83.0 years) in the T-DM1 group.”
“Race, n (%) | | White | 71 (27.2) | 72 (27.4) | | Black or African American | 10 (3.8) | 9 (3.4) | | Asian | 152 (58.2) | 162 (61.6)”
“The median age was 54.3 years (range, 27.9–83.1 years) in the T-DXd group and 54.2 years (range, 20.2–83.0 years) in the T-DM1 group.”
The trial protocol was approved by ethical bodies or institutional review boards at each site, and all patients provided written informed consent. The study was conducted in accordance with the Declaration of Helsinki and ICH-GCP. This meets the criteria for adequate ethics reporting.
“Before initiation of the study, the trial protocol was approved by the ethical bodies or institutional review boards at each site.”
“All participating patients provided their informed consent in writing before enrollment.”
“The study was conducted in accordance with the standards set by the Declaration of Helsinki, the International Conference on Harmonization Guideline for Good Clinical Practice, any local regulations and the study protocol.”
“Before initiation of the study, the trial protocol was approved by the ethical bodies or institutional review boards at each site.”
“All participating patients provided their informed consent in writing before enrollment.”
“The study was conducted in accordance with the standards set by the Declaration of Helsinki, the International Conference on Harmonization Guideline for Good Clinical Practice, any local regulations and the study protocol.”
The study uses two antibody-drug conjugates, T-DXd and T-DM1, which are identified by name and dose (5.4 mg/kg and 3.6 mg/kg). The manufacturer is implied (Daiichi Sankyo). Statistical software (SAS, R) is identified. No other biological or chemical resources are used, so other sub-criteria are not applicable.
“Patients with advanced HER2-positive metastatic breast cancer previously treated with taxane and trastuzumab were randomized to T-DXd (5.4 mg per kg (261 patients)) or T-DM1 (3.6 mg per kg (263 patients)).”
“Statistical analysis used SAS version 9.3 or later and R 4.2.0 for the RPSFTM.”
“Patients with advanced HER2-positive metastatic breast cancer previously treated with taxane and trastuzumab were randomized to T-DXd (5.4 mg per kg (261 patients)) or T-DM1 (3.6 mg per kg (263 patients)).”
“Statistical analysis used SAS version 9.3 or later and R 4.2.0 for the RPSFTM.”
The paper reports hazard ratios with 95% CIs for PFS, OS, and PFS2, which is the standard for clinical trials. The statistical tests are named (stratified log-rank test, Cox proportional hazards model, Cochran-Mantel-Haenszel test). Exact p-values are not reported for the updated analyses, but this is acceptable as the analyses are exploratory and effect estimates with CIs are provided. Software is identified. Data presentation includes Kaplan-Meier curves and tables with per-group n. Mathematical plausibility checks were not performed due to the nature of the data (survival analysis).
“Analysis of PFS and OS between treatment groups employed a stratified log-rank test, considering randomization factors.”
“median OS was 52.6 versus 42.7 months (HR, 0.73; 95% CI, 0.56–0.94) with T-DXd versus T-DM1, respectively.”
“Statistical analysis used SAS version 9.3 or later and R 4.2.0 for the RPSFTM.”
“Analysis of PFS and OS between treatment groups employed a stratified log-rank test, considering randomization factors.”
“median PFS by investigator assessment was 29.0 versus 7.2 months (hazard ratio (HR), 0.30; 95% confidence interval (CI), 0.24–0.38)”
“median OS was 52.6 versus 42.7 months (HR, 0.73; 95% CI, 0.56–0.94) with T-DXd versus T-DM1, respectively.”
The data availability statement provides a concrete route for accessing anonymized individual participant data through the Vivli platform, with a link to the specific member page. This meets the criteria for adequate data availability. Since the data are patient-level and identifiable, repository deposit and accession numbers are not applicable. No custom code is mentioned, so code sharing is not applicable.
The trial is registered (NCT03529110). Methods are detailed enough for replication. The paper discusses limitations, including the exploratory nature of the analysis and the lack of BICR for updated endpoints. Conclusions are proportional to the evidence, with appropriate caveats about cross-trial comparisons. Funding and competing interests are disclosed.
“This study was funded by Daiichi Sankyo and AstraZeneca.”
“ClinicalTrials.gov registration: NCT03529110”
“Potential limitations of the DESTINY-Breast03 trial have been published”
“This study was funded by Daiichi Sankyo and AstraZeneca.”
Registered (3 IDs: ClinicalTrials.gov). No reporting guideline cited.
Broken references and links
1 finding · worst lowReferences 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.
- References not resolvable to a published paperRecomputed
Checked 25 references by DOI: 24 verified — 1 DOI unresolved.
- UNRESOLVED10.1016/s1470-2045(19ENHERTU® Approved in the U.S. for Patients with HER2 Positive Metastatic Breast Cancer Treated with a Prior Anti-HER2-Based RegimenCited DOI does not resolve to any Crossref record.
2 data/code links checked; 2 live.
- datahttps://vivli.org/LIVEHTTP 200Resolves, but the content could not be matched to the paper.
- datahttps://vivli.org/ourmember/daiichi-sankyo/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 consistency, typo, clarity.
- MINORconsistencyAbstract“median OS was 52.6 versus 42.7 months (HR, 0.73; 95% CI, 0.56–0.94)”→ Consider adding 'respectively' for clarity.Minor clarity issue.
- MINORtypoDiscussion“sequalae”→ Change to 'sequelae'.Spelling error.
- MINORconsistencyTable 2“HR (95% CI) | 0.73 (0.56–0.94)”→ Ensure consistent formatting of HRs across tables.Formatting consistency.
- MINORclarityAbstract“Trastuzumab deruxtecan (T-DXd) demonstrated significantly improved efficacy over trastuzumab emtansine (T-DM1) in DESTINY-Breast03 (median follow-up, 28 months).”→ Consider adding 'at the primary analysis' after 'DESTINY-Breast03' for clarity, as the current analysis has longer follow-up.Minor clarity issue; the sentence could be misinterpreted as referring to the current analysis.
- MINORconsistencyTable 3 footnote“T-DXd, 5.4 mg/kg Q3W, n = 257”→ Ensure consistency: the safety analysis set is n=257 for T-DXd, but the full analysis set is n=261. This is correctly noted in the text but could be clarified in the table title.Minor consistency issue; the table title could explicitly state 'Safety analysis set'.
The published work is robust and well-reported; an informed reader should weigh the minor reporting gaps (explicit power analysis, outlier handling, exact p-values for exploratory analyses, and statistical assumption verification) as limitations but not validity threats. The one reference not found in the registry should be verified or corrected, as it may be a fabrication signal. No erratum is warranted for the statistical results, as all recomputed tests were consistent.
- 1.HIGHreportingVerify or correct the reference 'ENHERTU® Approved in the U.S. for Patients with HER2 Positive Metastatic Breast Cancer Treated with a Prior Anti-HER2-Based Regimen' (DOI 10.1016/s1470-2045(19) — incomplete DOI) that was not found in any registry; if it cannot be verified, remove or replace it.A reference that cannot be located in any registry is a potential fabrication signal and must be resolved before publication.
- 2.HIGHreportingAdd an explicit a priori power analysis to the Methods (Statistical analysis) section, stating the assumed effect size, alpha, and power that justified the target enrollment of ~500 patients.Reviewer 2 flagged the absence of an explicit power calculation as a reporting gap; providing it strengthens the study design reporting.
- 3.HIGHstatisticsReport exact p-values for the updated exploratory analyses (e.g., OS HR) in the Abstract and Results, or state that p-values are intentionally omitted for exploratory analyses.Both reviewers noted that exact p-values are not reported for the current exploratory analysis; adding them would facilitate interpretation.
- 4.HIGHstatisticsAdd a statement in the Methods (Statistical analysis) on how outliers were handled, even if none were identified or standard methods were used.Reviewer 2 flagged outlier handling as not reported; a brief statement would close this reporting gap.
- 5.HIGHstatisticsExplicitly state that assumptions of the statistical models (e.g., proportional hazards for the Cox model) were verified, or note that standard methods were used and assumptions are considered met.Reviewer 2 flagged assumptions verification as inadequate; adding this statement improves transparency.
- 6.MEDIUMreportingClarify the reporting guideline used (e.g., CONSORT) in the Methods or reporting summary, rather than only referencing the Nature Portfolio reporting summary.Reviewer 1 suggested clarifying the reporting guideline to enhance transparency.
- 7.MEDIUMstatisticsProvide more detail on the RPSFTM sensitivity analysis in the Methods, including assumptions and potential biases.Reviewer 1 suggested this to aid reproducibility.
- 8.MEDIUMreportingProvide a more detailed description of the randomization algorithm (e.g., block sizes) in the Methods (Trial design).Reviewer 2 suggested this to strengthen the design reporting.
- 9.MEDIUMreportingAdd a statement on whether any data were excluded from analysis and the reasons for exclusion, beyond the pre-specified criteria.Reviewer 2 suggested this to enhance transparency.
- 10.MEDIUMstatisticsConsider adding a supplementary table with exact p-values for all secondary and exploratory endpoints.Reviewer 2 suggested this to enhance transparency.
- 11.LOWcopyeditFix the typo 'sequalae' to 'sequelae' in the Discussion.Copyedit pass flagged this spelling error.
- 12.LOWcopyeditAdd 'respectively' to the Abstract sentence 'median OS was 52.6 versus 42.7 months (HR, 0.73; 95% CI, 0.56–0.94)' for clarity.Copyedit pass flagged this as a minor clarity issue.
- 13.LOWcopyeditClarify the Abstract sentence 'Trastuzumab deruxtecan (T-DXd) demonstrated significantly improved efficacy over trastuzumab emtansine (T-DM1) in DESTINY-Breast03 (median follow-up, 28 months)' by adding 'at the primary analysis' to avoid confusion with the current longer follow-up.Copyedit pass flagged this as potentially misleading.
- 14.LOWcopyeditEnsure consistent formatting of HRs across tables (e.g., Table 2) and clarify that Table 3 uses the safety analysis set (n=257 for T-DXd) in the table title.Copyedit pass flagged formatting and consistency issues in tables.
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.