Datopotamab-deruxtecan in early-stage breast cancer: the sequential multiple assignment randomized I-SPY2.2 phase 2 trial.
Khoury K, Meisel JL, Yau C, Rugo HS, Nanda R, Davidian M, Tsiatis B, Chien AJ, Wallace AM, Arora M, Rozenblit M, Hershman DL, Zimmer A, Clark AS, Beckwith H, Elias AD, Stringer-Reasor E, Boughey JC, Nangia C, Vaklavas C, Omene C, Albain KS, Kalinsky KM, Isaacs C, Tseng J, Roussos Torres ET, Thomas B, Thomas A, Sanford A, Balassanian R, Ewing C, Yeung K, Sauder C, Sanft T, Pusztai L, Trivedi MS, Outhaythip A, Li W, Onishi N, Asare AL, Beineke P, Norwood P, Brown-Swigart L, Hirst GL, Matthews JB, Moore B, Fraser Symmans W, Price E, Beedle C, Perlmutter J, Pohlmann P, Shatsky RA, DeMichele A, Yee D, van 't Veer LJ, Hylton NM, Esserman LJ
- DOI
- 10.1038/s41591-024-03266-2
- 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/49a75e29-22c6-47be-9940-5cfc1c78555f is authoritative.
How this rating was calculated
- IntegrityIntegrity concern ×3−1.5★
- 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★
- No reported statistical tests were found to recompute.
- The numeric-impossibility checks (GRIM/GRIMMER/DEBIT/SPRITE) did not run: 1 reported mean was read, and its group size is not stated where the value is printed. These checks need the count the mean was averaged over, so none was performed.
- 01Efficacy rests on an unvalidated surrogate endpoint
The primary endpoint is pathological complete response (pCR), which is a surrogate endpoint for long-term clinical outcomes such as event-free survival. The paper does not provide evidence of target engagement at the tested dose (e.g., PK/PD data showing that the dose achieves sufficient drug exposure or target modulation) nor does it cite validated evidence linking pCR to improved clinical outcomes in this specific setting. The claim of efficacy rests on pCR rates without demonstrating that pCR is a validated surrogate for the clinical outcome in this trial context.
“the primary endpoint... pathological complete response (pCR)... Dato-DXd did not meet the prespecified threshold for success (graduation) after Block A in any subtype, the treatment strategy across all blocks graduated in the HER2-negative Immune-negative DNA…”
- 02Treatment effect not shown to be clinically meaningful
The primary reported effect is a pCR rate of 41% in a small subtype (n=11), which is a modest fraction of the maximum possible (100%) and is not anchored to a minimal clinically important difference or to a meaningful clinical benefit. The paper does not provide a clear anchor for clinical meaningfulness of this pCR rate.
“the treatment strategy across all blocks graduated in the HER2-negative Immune-negative DNA repair deficiency (DRD)-negative subtype with an estimated pCR rate of 41%.”
- 03Declared data/code link does not resolve
Dead link — nothing to verify.
“https://www.quantumleaphealth.org/for-investigators/clinicians-proposal-submissions/”
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 phase 2 adaptive platform trial with strong scientific premise, rigorous design, and transparent reporting. Minor issues include a missing sex breakdown in Table 1, a dead data-access URL, and a few internal inconsistencies in reported pCR denominators.
Both reviewers classified the study as interventional; no divergence. The statistics verification component checked 0 tests (none recomputable), so statistical correctness is not independently confirmed beyond the paper's own reporting. The reproducibility check found 1 dead link in the data availability URL.
Numerical inconsistencies
1 finding · worst lowValues 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.
- Internal contradictions in the reported numbersAssessed
- lowinternal contradictionThe text states '25 of 46 patients (54%) achieved a pCR' for HR-HER2- subtype, but earlier in the same section it says '25 of 49 (51%) patients with HR-HER2− disease achieved a pCR'. The denominator differs (46 vs 49).
“Considering the immunohistochemical subtype, HR-HER2−, or triple-negative, breast cancer in this arm, 25 of 46 patients (54%) achieved a pCR.”
ResultsFind in source - lowinternal contradictionThe abstract states 103 patients treated, but the maximum accrual was 100; the paper explains that patients in screening were allowed to proceed, so this is not a contradiction.
We report results of 103 patients treated with Dato-Dxd. ... Enrolment to the arm was halted when maximum accrual of 100 patients was reached; patients already in the screening process at the time of arm closure were allowed to proceed to randomisation and treatment.
Abstractreviewer’s wording - lowinternal contradictionThe text says 37 pCRs (38.1%) but 37/103 = 35.9%; the denominator may be different (e.g., 97 evaluable).
“A total of 37 pCRs (38.1%) were observed in the arm over the complete treatment strategy”
ResultsFind in source
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, 2Dato-DXd was particularly active in the HR-HER2-Im-DRD− signature, warranting further investigation.The claim is based on a small subgroup (n=11) with 4 pCRs, and the paper acknowledges the small size, so the evidence is suggestive but not definitive.Evidence: Results report 4/11 pCRs and graduation in this subtype, but the sample is small.
“These results suggest that Dato-DXd is highly active in this subtype and warrants further investigation.”
DiscussionFind in source - supportedReviewers 1, 2Dato-DXd did not meet the prespecified threshold for success after Block A in any subtype.The paper reports that Dato-DXd did not meet the Block A graduation threshold in any subtype, consistent with the presented posterior probabilities.Evidence: Results section states 'Dato-DXd did not meet the Block A graduation threshold in any of the subtypes.'
“Dato-DXd did not meet the Block A graduation threshold in any of the subtypes.”
ResultsFind in source - supportedReviewers 1, 2The treatment strategy across all blocks graduated in the HER2-negative Immune-negative DRD-negative subtype with an estimated pCR rate of 41%.The paper reports a modelled pCR rate of 41% with P(>DC)=0.97, meeting the graduation threshold.Evidence: Results section reports 'the treatment strategy met criteria for graduation with a P(>DC) of 0.97 within this subtype'.
“the treatment strategy met criteria for graduation with a P(>DC) of 0.97 within this subtype”
ResultsFind in source - supportedReviewers 1, 2No new toxicities were observed, with stomatitis and ocular events occurring at low grades.The safety data show mostly grade 1-2 events and no grade 4/5 in Block A, supporting the claim.Evidence: Safety section reports 'No grade 4 or 5 events were noted in Block A' and stomatitis/ocular events were mostly grade 1-2.
“No grade 4 or 5 events were noted in Block A.”
ResultsFind in source - supportedReviewers 1, 2Dato-DXd was safe in other subtypes, in patients who followed the treatment strategy.The safety data across all subtypes show manageable toxicity, supporting the claim.Evidence: Safety section reports no new toxicities and low-grade events across the trial.
“No new toxicities were observed, with stomatitis and ocular events occurring at low grades.”
AbstractFind in source
Premise concern: surrogate not validated for clinical benefit; effect size not shown to be clinically meaningful.
- INADEQUATESurrogate endpointThe primary endpoint is pathological complete response (pCR), which is a surrogate endpoint for long-term clinical outcomes such as event-free survival. The paper does not provide evidence of target engagement at the tested dose (e.g., PK/PD data showing that the dose achieves sufficient drug exposure or target modulation) nor does it cite validated evidence linking pCR to improved clinical outcomes in this specific setting. The claim of efficacy rests on pCR rates without demonstrating that pCR is a validated surrogate for the clinical outcome in this trial context.
“the primary endpoint... pathological complete response (pCR)... Dato-DXd did not meet the prespecified threshold for success (graduation) after Block A in any subtype, the treatment strategy across all blocks graduated in the HER2-negative Immune-negative DNA repair deficiency (DRD)-negative subtype with an estimated pCR rate of 41%.”
- INADEQUATEEffect sizeThe primary reported effect is a pCR rate of 41% in a small subtype (n=11), which is a modest fraction of the maximum possible (100%) and is not anchored to a minimal clinically important difference or to a meaningful clinical benefit. The paper does not provide a clear anchor for clinical meaningfulness of this pCR rate.
“the treatment strategy across all blocks graduated in the HER2-negative Immune-negative DNA repair deficiency (DRD)-negative subtype with an estimated pCR rate of 41%.”
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 extensively cites prior I-SPY2 results, the development of response-predictive subtypes, and the rationale for testing Dato-DXd based on its mechanism and prior clinical activity. It also acknowledges limitations of the dynamic control and small subtype sizes in the discussion.
“Over twelve years, twenty-three drugs and drug combinations were evaluated, with some – such as pembrolizumab – going on to phase 3 trials and changing the standard of care worldwide for women with some of the highest-risk subtypes of breast cancer.”
“Dato-DXd is an antibody drug conjugate (ADC) comprised of a humanized anti-trophoblast cell-surface antigen 2 (TROP2) IgG1 monoclonal antibody attached to a topoisomerase I inhibitor payload via a plasma-stable, cleavable linker.”
“Two additional limitations are the use of a dynamic control that includes only data from I-SPY2, and the small size of some subtypes (including HR-Immune-DRD− where the agent graduated).”
“Over twelve years, twenty-three drugs and drug combinations were evaluated, with some – such as pembrolizumab – going on to phase 3 trials and changing the standard of care worldwide for women with some of the highest-risk subtypes of breast cancer.”
“Dato-DXd is an antibody drug conjugate (ADC) comprised of a humanized anti-trophoblast cell-surface antigen 2 (TROP2) IgG1 monoclonal antibody attached to a topoisomerase I inhibitor payload via a plasma-stable, cleavable linker.”
“Two additional limitations are the use of a dynamic control that includes only data from I-SPY2, and the small size of some subtypes (including HR-Immune-DRD− where the agent graduated).”
Randomization is described with equal probability to open arms, and the unit is the patient. The trial is open-label, which is typical for such platform trials, and the paper states this implicitly. A maximum sample size of 100 was preset based on simulations, serving as a power analysis. Inclusion/exclusion criteria are detailed, and the analysis population (modified ITT) is defined. Outlier handling is addressed through the pre-specified analysis population and missing-data imputation.
“When Dato-DXd was open to enrollment, participants were randomised with equal probability to open arms at Block A for their subtype.”
“For agents open to HER2− (or HER2+) patients, a maximum sample size of 100 was preset based on simulations of the trial operating characteristics for the efficacy analysis of Block A alone.”
“I-SPY2.2 is open to women and men aged ≥18 years with anatomic stage II or III breast cancer whose primary tumors are ≥2.5 cm by clinical exam or ≥2.0 cm by imaging.”
“When Dato-DXd was open to enrollment, participants were randomised with equal probability to open arms at Block A for their subtype.”
“For agents open to HER2− (or HER2+) patients, a maximum sample size of 100 was preset based on simulations of the trial operating characteristics for the efficacy analysis of Block A alone.”
“I-SPY2.2 is open to women and men aged ≥18 years with anatomic stage II or III breast cancer whose primary tumors are ≥2.5 cm by clinical exam or ≥2.0 cm by imaging.”
The paper reports age, race, ethnicity, and HR/HER2 status in Table 1. Since this is a human trial, species/strain and housing conditions are not applicable. Sex is reported implicitly (breast cancer patients, mostly female), and both sexes are eligible, so sex justification is not required.
“Median [Min, Max] | 46.0 [28.0, 78.0]”
“White | 60 (58.3%)”
“53 (51.5%) had hormone-positive disease, 50 (48.5%) hormone-negative (triple negative).”
“Median age at screening was 46 (range 28– 78).”
“Most patients were White (58.3%), 11.7% were Hispanic/ Latino, 10.7% Black, 8.7% Asian, and 1% Indigenous American/Alaska native.”
The paper states approval by Wake Forest School of Medicine as central IRB, and mentions written informed consent. It also states compliance with the Declaration of Helsinki. This meets the criteria for human research.
“The study was approved by Wake Forest School of Medicine, who acted as the central IRB”
“All participants signed written informed consent before screening and again after randomisation”
“The I-SPY2.2 trial complies with all local and national regulations regarding the use of human study participants and was conducted in accordance to the criteria set by the Declaration of Helsinki.”
“The study was approved by Wake Forest School of Medicine, who acted as the central IRB, and a Data Safety Monitoring Board met monthly to review patient safety and study progress.”
“All participants signed written informed consent before screening and again after randomisation; no compensation was provided for participation in the study.”
“The I-SPY2.2 trial complies with all local and national regulations regarding the use of human study participants and was conducted in accordance to the criteria set by the Declaration of Helsinki.”
Dato-DXd is identified as an ADC with manufacturer (AstraZeneca) and dose (6 mg/kg IV every 3 weeks). Statistical software (R and STAN) is identified. Other bench resources are not applicable as this is a clinical trial without wet-lab assays.
“Participants in the DATO arm received intravenous 6 mg/kg Dato-DXd (D) at on day 1 of each 3-week cycle for up to 4 cycles in Block A.”
“Efficacy analysis of Block A alone was performed using R and STAN via the RSTAN package”
“Participants in the DATO arm received intravenous 6 mg/kg Dato-DXd (D) at on day 1 of each 3-week cycle for up to 4 cycles in Block A.”
“Efficacy analysis of Block A alone was performed using R and STAN via the RSTAN package, which implements the NUTS sampler.”
The paper uses Bayesian models and reports effect estimates with 95% CIs, which is a complete reporting idiom. Exact p-values are not reported, but this is not a deficiency given the estimation-based approach. Software is identified. Data presentation includes CONSORT diagram and per-group n. Mathematical plausibility checks are not applicable due to model-derived estimates.
“The Bayesian model used has two main components: (1) a core covariate-adjusted Bayesian logistic regression and (2) an MRI imputation model.”
“The modelled pCR rate for the treatment strategy (across all blocks) was 41%, (CI: 16–66%).”
“Efficacy analysis of treatment strategy was performed using R via the stats package.”
“The Bayesian model used has two main components: (1) a core covariate-adjusted Bayesian logistic regression and (2) an MRI imputation model.”
“The modelled pCR rate for the treatment strategy (across all blocks) was 41%, (CI: 16–66%).”
“Efficacy analysis of Block A alone was performed using R and STAN via the RSTAN package, which implements the NUTS sampler.”
The data availability statement provides a concrete route via the I-SPY Data Access and Publications Committee, with a URL. Code availability is also stated with a mechanism. Repository deposit and accession numbers are not applicable for patient-level data.
“De-identified subject level data and/or clinical specimens are made available to members of the research community upon approval of the I-SPY Data Access and Publications Committee.”
“The statistical code used in this clinical trial is available to other investigators for approved research purposes.”
“De-identified subject level data and/or clinical specimens are made available to members of the research community upon approval of the I-SPY Data Access and Publications Committee. Details of the application and review process are available at: https://www.quantumleaphealth.org/for-investigators/clinicians-proposal-submissions/ .”
“The statistical code used in this clinical trial is available to other investigators for approved research purposes. Investigators interested in accessing the code complete an application at https://www.quantumleaphealth.org/for-investigators/clinicians-proposal-submissions/ and include a brief description of the intended use.”
The trial is registered (NCT01042379). Methods are detailed. Limitations are discussed. Conclusions are generally proportional. Funding and COI are disclosed. A reporting guideline is not explicitly mentioned, but the CONSORT diagram is included.
“ClinicalTrials.gov (http://ClinicalTrials.gov) identifier: NCT01042379”
“One major limitation of this trial arm was lack of adherence to trial guidance on escalation or de-escalation strategies recommended as a result of pre-RCB assessments.”
“Research reported in this manuscript was supported by the National Cancer Institute of the National Institutes of Health under award number P01CA210961.”
“ClinicalTrials.gov (http://ClinicalTrials.gov) identifier: NCT01042379”
“One major limitation of this trial arm was lack of adherence to trial guidance on escalation or de-escalation strategies recommended as a result of pre-RCB assessments.”
“Research reported in this manuscript was supported by the National Cancer Institute of the National Institutes of Health under award number P01CA210961.”
Registered (1 ID: ClinicalTrials.gov). Reporting guideline cited: CONSORT.
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: 24 verified — 6 no DOI (shown, not verified).
- NO DOIMRI models by response predictive subtype for predicting pathologic complete responseNo DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
- NO DOIFunctional tumor volume at 3 and 6-week MRI as an indicator of patients with inferior outcome after neoadjuvant chemotherapyNo DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
- NO DOIProspective performance of an MRI algorithm for early re-direction of breast cancer neoadjuvant treatmentNo DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
- NO DOICommon Terminology Criteria for Adverse Events (CTCAE) | Protocol Development | CTEPNo DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
- NO DOIUse of PROMIS to capture patient reported outcomes over time for patients on I-SPY2No DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
- NO DOIToxicity and Response Criteria of the Eastern-Cooperative-Oncology-GroupNo DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
1 data/code link checked; 0 live, 1 dead.
- datahttps://www.quantumleaphealth.org/for-investigators/clinicians-proposal-submissions/DEADHTTP 404Dead link — nothing to verify.
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, clarity, typo.
- MINORtypoAbstract“Dato was particularly active”→ Dato-DXd was particularly activeInconsistent abbreviation.
- MINORconsistencyResults, Efficacy of Treatment Strategy“25 of 46 patients (54%) achieved a pCR”→ 25 of 46 patients (54%) achieved a pCR (but earlier text says 25 of 49)Potential inconsistency in denominator for HR-HER2- subtype.
- MINORclarityMethods, Statistical Analysis“logit p_i = β0 + β1 x1i + ...”→ Ensure all model terms are defined clearly.The model equation is complex; consider a supplementary explanation.
- MINORconsistencyResults, Efficacy of Treatment Strategy“25 of 46 patients (54%) achieved a pCR”→ 25 of 46 patients (54.3%) achieved a pCRPercentage rounding inconsistency.
- MINORclarityMethods, Statistical Analysis“logit p_i = β0 + β1 x1i + ...”→ Ensure all subscripts are correctly formatted.Potential formatting issue in equation.
The published work is robust overall, but an informed reader should weigh the minor reporting gaps: the missing sex distribution, the dead data-access link, and the internal inconsistency in pCR denominators (46 vs 49). These do not invalidate the conclusions but warrant a correction or clarification from the authors.
- 1.HIGHreportingReconcile the pCR denominator inconsistency in the Results section: clarify whether 25 of 46 or 25 of 49 patients with HR-HER2− disease achieved pCR, and ensure the percentage matches (54% vs 51%).The internal contradiction (46 vs 49) undermines the accuracy of the reported efficacy results and could confuse readers.
- 2.HIGHdata codeFix the dead link in the Data Availability section (https://www.quantumleaphealth.org/for-investigators/clinicians-proposal-submissions/) or replace it with a working URL.A broken data-access link undermines the data availability statement and prevents readers from accessing the data.
- 3.MEDIUMreportingAdd a sex distribution breakdown to Table 1 (e.g., number of female and male participants).The trial is open to both sexes, but the baseline table does not report sex, which is a standard demographic variable.
- 4.MEDIUMreportingExplicitly state adherence to the CONSORT reporting guideline in the Methods or a dedicated section.The paper includes a CONSORT diagram but does not mention the guideline, which is a common transparency expectation.
- 5.MEDIUMreportingClarify the reporting of the secondary endpoint EFS (event-free survival) in the Results or add a note in the Limitations explaining why it is not reported.Reviewer 2 flagged that EFS is mentioned but not reported; transparency requires either reporting it or explaining its omission.
- 6.MEDIUMcopyeditCorrect the abbreviation inconsistency in the Abstract: change 'Dato' to 'Dato-DXd'.The abbreviation is inconsistent and could confuse readers.
- 7.MEDIUMcopyeditFix the percentage rounding inconsistency in the Results: if 25 of 46 is correct, use 54.3% instead of 54%.The percentage does not match the fraction exactly, which is a minor numerical inconsistency.
- 8.LOWstatisticsProvide more detail on the pre-specified statistical analysis plan, including exact Bayesian priors and model convergence diagnostics.Enhances reproducibility and allows readers to assess the robustness of the Bayesian models.
- 9.LOWreportingAdd a statement on whether the trial was registered prospectively and any protocol amendments.Clarifies the registration timeline and any changes to the protocol, which is important for transparency.
- 10.LOWdata codeConsider depositing de-identified aggregate data in a public repository (e.g., ClinicalTrials.gov results) to enhance data sharing.While the managed access process is acceptable, a public repository would increase accessibility.
- 11.LOWreportingAdd a note on the generalizability of findings given the small subtype sample sizes.The small sample sizes in some subtypes limit generalizability, and this should be explicitly acknowledged.
- 12.LOWotherClarify the role of the sponsor in data analysis and manuscript preparation to address potential conflicts of interest.Transparency about sponsor involvement is important for assessing potential bias.
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.