Adding 6 months of androgen deprivation therapy to postoperative radiotherapy for prostate cancer: a comparison of short-course versus no androgen deprivation therapy in the RADICALS-HD randomised controlled trial.
Parker CC, Clarke NW, Cook AD, Kynaston H, Catton CN, Cross WR, Petersen PM, Persad RA, Saad F, Bower LC, Logue J, Payne H, Forcat S, Goldstein C, Murphy C, Anderson J, Barkati M, Bottomley DM, Branagan J, Choudhury A, Chung PWM, Cogley L, Goh CL, Hoskin P, Khoo V, Malone SC, Masters L, Morris SL, Nabid A, Ong AD, Raman R, Tarver KL, Tree AC, Worlding J, Wylie JP, Zarkar AM, Parulekar WR, Parmar MKB, Sydes MR, RADICALS investigators
- DOI
- 10.1016/S0140-6736(24)00548-8
- Record issued
- 2026-08-16
- Engine
- 7.39.0
- Exported
- 2026-09-21
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/7742ba99-929d-459b-93b3-343ca9463ca6 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 well-conducted and transparently reported phase 3 randomised controlled trial. The methodology is rigorous, with clear randomisation, power calculation, and intention-to-treat analysis. Minor reporting gaps include lack of explicit CONSORT reference and absence of race/ethnicity data, but these are acknowledged and do not undermine the study's integrity.
Both reviewers independently scored all dimensions as pass, with high agreement. Minor divergences on 'controls' (one marked not applicable) and 'demographics' (one marked reported but inadequate) were resolved by weighing the evidence: the trial has a control arm, and the absence of race/ethnicity data is explicitly acknowledged as a limitation, making reporting adequate. The statistics verification checked 6 tests, all consistent; the citation check flagged one reference not found in registry (the SAP).
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 6 tests: 6 consistent, 0 inconsistent; 1 recomputed directly from the reported test statistics, 5 via agent-written checks.
- CONSISTENTreported p = .350 · recomputed p = .347Recomputed HR 0.886 (95% CI 0.688–1.140), reported p=0.35
“HR 0.886 [95% CI 0.688–1.140], p=0.35”
Taken as given: 0.688–1.140 is a two-sided 95% confidence interval for the HR of 0.886, not a range, an IQR, or a different interval level; the HR is a RATIO measure, so the interval is symmetric on the log scale; p=0.35 is the p for THIS estimate, not for another comparison in the same sentenceMethod: back the two-tailed p out of the log-scale CI width and compare it against the printed pHow we recomputed it: pCI(0.886, 0.688, 1.14, 1) - CONSISTENTreported p = .350 · recomputed p = .347Reviewers 1, 2Primary outcome metastasis-free survival HR p-value
“HR 0·886 [95% CI 0·688–1·140], p=0·35”
Taken as given: The HR is a ratio (log scale).; The CI is a 95% confidence interval.Method: Two-sided p-value derived from the reported HR and 95% CI using the normal approximation.How we recomputed it: pCI(0.886, 0.688, 1.140, 1) - CONSISTENTreported p = .420 · recomputed p = .417Reviewers 1, 2Overall survival HR p-value
“HR 0·882 (0·651–1·194) | 0·42”
Taken as given: The HR is a ratio (log scale).; The CI is a 95% confidence interval.Method: Two-sided p-value derived from the reported HR and 95% CI using the normal approximation.How we recomputed it: pCI(0.882, 0.651, 1.194, 1) - CONSISTENTreported p = .240 · recomputed p = .245Reviewers 1, 2Freedom from distant metastasis HR p-value
“HR 0·816 (0·579–1·150) | 0·24”
Taken as given: The HR is a ratio (log scale).; The CI is a 95% confidence interval.Method: Two-sided p-value derived from the reported HR and 95% CI using the normal approximation.How we recomputed it: pCI(0.816, 0.579, 1.150, 1) - CONSISTENTreported p < .001 · recomputed p = <.001Reviewers 1, 2Clinical progression-free survival HR p-value
“HR 0·544 (0·433–0·684) | <0·0001”
Taken as given: The HR is a ratio (log scale).; The CI is a 95% confidence interval.Method: Two-sided p-value derived from the reported HR and 95% CI using the normal approximation.How we recomputed it: pCI(0.544, 0.433, 0.684, 1) - CONSISTENTreported p < .001 · recomputed p = <.001Reviewers 1, 2Time to non-protocol ADT HR p-value
“HR 0·543 (0·422–0·699) | <0·0001”
Taken as given: The HR is a ratio (log scale).; The CI is a 95% confidence interval.Method: Two-sided p-value derived from the reported HR and 95% CI using the normal approximation.How we recomputed it: pCI(0.543, 0.422, 0.699, 1)
Overstated conclusions
None foundConclusions 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.
Checked — nothing surfaced.
7 major claims checked against the paper's own evidence: all adequately supported.
- supportedReviewers 1, 2Adding 6 months of ADT to postoperative radiotherapy did not improve metastasis-free survival.The primary outcome analysis shows no statistically significant difference (HR 0.886, 95% CI 0.688-1.140, p=0.35), supporting the claim.Evidence: Primary outcome result: HR 0.886 (95% CI 0.688-1.140), p=0.35.
“Adding 6 months of ADT to this radiotherapy did not improve metastasis-free survival compared with no ADT.”
AbstractFind in source - supportedReviewer 1Short-course ADT delayed time to salvage ADT.The secondary outcome time to non-protocol ADT showed a significant benefit (HR 0.543, p<0.0001), supporting the claim.Evidence: Time to non-protocol ADT: HR 0.543 (95% CI 0.422-0.699), p<0.0001.
“For time to non-protocol ADT, the HR was 0·543 (0·422–0·699; log-rank p<0·0001)”
Results ¶3Find in source - supportedReviewers 1, 2These findings do not support the use of short-course ADT with postoperative radiotherapy.Given the lack of benefit in the primary outcome and the modest benefit in delaying salvage ADT, the conclusion is supported.Evidence: Primary outcome negative; secondary outcome shows delay in salvage ADT but no improvement in metastasis-free survival.
“These findings do not support the use of short-course ADT with postoperative radiotherapy in this patient population.”
AbstractFind in source - supportedReviewers 1, 2The trial had 80% power to detect an absolute increase in 10-year metastasis-free survival from 80% to 86%.The power calculation is explicitly stated in the methods.Evidence: Methods, Statistical analysis: 'With 200 events from the 1480 participants, this final design had 80% power with two-sided α of 5% to detect an increase in 10-year metastasis-free survival from 80% to 86% (HR=0·67).'
“With 200 events from the 1480 participants, this final design had 80% power with two-sided α of 5% to detect an increase in 10-year metastasis-free survival from 80% to 86% (HR=0·67).”
Statistical analysisFind in source - supportedReviewer 1The trial is the first randomised trial of short-course ADT in men having postoperative radiotherapy that used metastasis-free survival as the primary outcome.The paper states this as a novel contribution, and the cited prior trials used progression-free survival as primary outcome.Evidence: Introduction and Discussion describe prior trials using PSA-based outcomes.
“To our knowledge, RADICALS-HD is the first randomised trial of short-course ADT in men having postoperative radiotherapy that has used a long-term, clinically meaningful, primary outcome measure of metastasis-free survival.”
Added value of this studyFind in source - supportedReviewer 2Short-course ADT delayed the time to salvage ADT.The secondary outcome 'time to non-protocol ADT' showed a significant benefit (HR 0.543, p<0.0001), supporting the claim.Evidence: Time to non-protocol ADT HR 0.543 (95% CI 0.422-0.699), p<0.0001.
“the addition of 6 months of ADT to postoperative prostate bed radiotherapy did not improve metastasis-free survival, but it did delay the time to salvage ADT.”
Discussion ¶1Find in source - supportedReviewer 2The change of primary outcome to metastasis-free survival was based on ICECaP evidence.The paper explains the change and cites ICECaP as the rationale.Evidence: Methods section describes the change and cites ICECaP.
“This change followed new evidence from the ICECaP study that metastasis-free survival was a robust early surrogate outcome measure for disease-specific survival.”
Statistical analysisFind in source
Efficacy claim is anchored to an adequate endpoint and a meaningful effect.
- ADEQUATESurrogate endpointThe primary outcome is metastasis-free survival, a hard clinical outcome (distant metastasis or death from any cause), not a surrogate. The paper explicitly justifies this choice based on ICECaP evidence that metastasis-free survival is a robust surrogate for disease-specific survival, and the trial was powered on this endpoint.
“The primary outcome measure was metastasis-free survival, defined as distant metastasis arising from prostate cancer or death from any cause.”
- ADEQUATEEffect sizeThe primary result is a null finding: HR 0.886 (95% CI 0.688–1.140), p=0.35, with 10-year metastasis-free survival 79.2% vs 80.4%. The effect is not statistically significant and the paper concludes no benefit, which is clinically meaningful in the context of a non-inferiority/negative trial. The effect size is anchored to clinical meaningfulness by the trial's design (powered to detect an absolute increase from 80% to 86%).
“There was no evidence that metastasis-free survival was improved in patients allocated to short-course ADT compared with those allocated to no ADT (HR 0·886 [95% CI 0·688–1·140], p=0·35).”
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 cites two prior randomised trials and discusses their limitations, particularly the use of PSA progression as an outcome. The rationale for the study and the choice of metastasis-free survival is logically linked to the limitations of PSA-based outcomes. The paper also addresses how the study design mitigates prior limitations by using a clinically meaningful primary outcome.
“Two randomised controlled trials, GETUG-AFU 16 and RTOG 0534, have previously tested the addition of short-course ADT to salvage radiotherapy to the prostate bed.”
“Given the limitations of PSA-based outcome measures in any trial of ADT, we used metastasis-free survival as the primary outcome measure.”
“Our view was that short-course ADT would be expected to delay PSA progression after postoperative radiotherapy, and yet such a positive finding should not be sufficient to justify its use.”
“Two randomised controlled trials, GETUG-AFU 16 and RTOG 0534, have previously tested the addition of short-course ADT to salvage radiotherapy to the prostate bed.”
“This change followed new evidence from the ICECaP study that metastasis-free survival was a robust early surrogate outcome measure for disease-specific survival.”
“ADT will inevitably delay PSA progression, and that alone should not be sufficient to change practice.”
Randomisation was centralised using minimisation with a random element, stratified by key prognostic factors. The trial was open-label, which is stated. A power calculation was provided (80% power to detect HR 0.67). Inclusion/exclusion criteria were pre-specified. The analysis followed intention-to-treat. Outlier handling is not explicitly described, but the analysis population is defined. Controls are inherent in the randomised comparison. Independent replication is not applicable for a single pivotal trial.
“Recruitment and random allocation of participants was implemented centrally using minimisation with a random element, stratified by Gleason score, positive margins, radiotherapy timing, planned radiotherapy schedule, and planned type of ADT, in a computerised system.”
“With 200 events from the 1480 participants, this final design had 80% power with two-sided α of 5% to detect an increase in 10-year metastasis-free survival from 80% to 86% (HR=0·67).”
“The allocated treatment was open label.”
“Recruitment and random allocation of participants was implemented centrally using minimisation with a random element, stratified by Gleason score, positive margins, radiotherapy timing, planned radiotherapy schedule, and planned type of ADT, in a computerised system.”
“The allocated treatment was open label.”
“With 200 events from the 1480 participants, this final design had 80% power with two-sided α of 5% to detect an increase in 10-year metastasis-free survival from 80% to 86% (HR=0·67).”
The paper reports age, PSA, Gleason score, T stage, lymph node involvement, positive margins, CAPRA-S score, Charlson Comorbidity Index, and country. Sex is not reported as a variable because all participants are male (prostate cancer), which is appropriate. Age and health status are reported. Demographics such as race/ethnicity are not collected, which is acknowledged as a limitation.
“Data on race and ethnicity were not collected.”
“Age, years | 66 (61–69) | 66 (61–69) | 66 (61–69)”
“data were not collected on ethnicity and race, so we cannot comment on how well the participants reflect the underlying population”
The paper states that appropriate ethical review was in place for each participating country and that all participants gave written informed consent. The trial is registered with ISRCTN and ClinicalTrials.gov. Regulatory compliance is implied through the ethical review process.
“All participants gave written informed consent.”
“This study is registered with the ISRCTN registry, ISRCTN40814031, and with ClinicalTrials.gov (https://ClinicalTrials.gov) , NCT00541047 .”
“All participants gave written informed consent.”
The ADT regimens are described in detail (gonadotropin-releasing hormone analogue, bicalutamide, degarelix) with dosing and administration. The radiotherapy schedule is specified. Statistical software (Stata version 17.0) is identified. No antibodies, cell lines, or other bench reagents are used, so those criteria are not applicable.
“Participants were given ADT using gonadotropin-releasing hormone analogue therapy according to site choice using their indicated route, usually subcutaneously, with once-monthly injections recommended.”
“All analyses followed the intention-to-treat principle, with analyses according to allocated group, and were conducted in Stata version 17.0.”
“Participants were given ADT using gonadotropin-releasing hormone analogue therapy according to site choice using their indicated route, usually subcutaneously, with once-monthly injections recommended.”
“All analyses followed the intention-to-treat principle, with analyses according to allocated group, and were conducted in Stata version 17.0.”
The paper names the statistical tests used (log-rank test, Cox regression, competing risks regression, chi-square test). Assumptions are addressed via the Grambsch–Therneau test for proportional hazards. Exact p-values are reported for primary and secondary outcomes. Effect sizes are reported with 95% confidence intervals. Statistical 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 (time-to-event, large N).
“For time-to-event outcome measures, the statistical significance of differences between groups was evaluated with the log-rank test, stratified by randomisation stratification factors.”
“There was no evidence that metastasis-free survival was improved in patients allocated to short-course ADT compared with those allocated to no ADT (HR 0·886 [95% CI 0·688–1·140], p=0·35; , ).”
“For time-to-event outcome measures, the statistical significance of differences between groups was evaluated with the log-rank test, stratified by randomisation stratification factors.”
“HR 0·886 [95% CI 0·688–1·140], p=0·35”
“The Grambsch–Therneau test was used to test the proportional hazards assumption”
The paper provides a data sharing statement describing a controlled access approach with a formal application process. This is appropriate for individual patient data. No repository deposit or accession numbers are applicable for patient-level data. No custom code is mentioned, so code sharing is not applicable.
“The RADICALS trial data are held at the MRC Clinical Trials Unit at UCL, which encourages optimal use of data by using a controlled access approach to data sharing (https://www.mrcctu.ucl.ac.uk/our-research/other-research-policy/data-sharing/) .”
“The RADICALS trial data are held at the MRC Clinical Trials Unit at UCL, which encourages optimal use of data by using a controlled access approach to data sharing”
The trial is registered with ISRCTN and ClinicalTrials.gov. Methods are detailed enough for replication. Limitations are explicitly discussed, including the lack of race/ethnicity data and the low event rate. Conclusions are appropriately cautious. Funding and conflicts of interest are declared.
“This study is registered with the ISRCTN registry, ISRCTN40814031, and with ClinicalTrials.gov (https://ClinicalTrials.gov) , NCT00541047 .”
“Grant funding in the UK was provided by the Clinical Trials Advisory Award Committee on behalf of Cancer Research UK (UK/C7829/A6381).”
“This study is registered with the ISRCTN registry, ISRCTN40814031, and with ClinicalTrials.gov (https://ClinicalTrials.gov) , NCT00541047 .”
“RADICALS-HD has several limitations.”
“Grant funding in the UK was provided by the Clinical Trials Advisory Award Committee on behalf of Cancer Research UK (UK/C7829/A6381).”
Registered (3 IDs: ClinicalTrials.gov, ISRCTN, PROSPERO). 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 19 references by DOI: 18 verified — 1 DOI unresolved.
- UNRESOLVED10.5281/zenodo.6586525RADICALS trial statistical analysis planCited DOI does not resolve to any Crossref record.
1 data/code link checked; 1 live.
- datahttps://www.mrcctu.ucl.ac.uk/our-research/other-research-policy/data-sharing/LIVEHTTP 200Resolves, but the content could not be matched to the paper.
Copyediting
4 minorWording, consistency and formatting errors that need correcting before submission.
No major wording or formatting errors. 4 minor suggestions below.
4 copyedit issues flagged: mostly consistency, typo, clarity.
- MINORtypoAbstract, Methods“The trial had 80% power with two-sided α of 5% to detect an absolute increase in 10-year metastasis-free survival from 80% to 86% (hazard ratio [HR] 0·67).”→ Consider rephrasing for clarity: 'The trial had 80% power, with a two-sided α of 5%, to detect an absolute increase...'Minor punctuation issue.
- MINORconsistencyResults, paragraph 1“1267 (86%) of 1478 had a Gleason score of 7 or higher”→ Ensure the denominator (1478) is consistent with the total number of participants with available data.The denominator is not 1480, which is explained by missing data, but the text could clarify.
- MINORconsistencyAbstract, Findings“1480 patients (median age 66 years [IQR 61–69]) were randomly assigned”→ Consider adding 'to receive' for clarity: 'were randomly assigned to receive'Minor grammatical improvement.
- MINORclarityMethods, Statistical analysis“The full statistical analysis plan is published elsewhere and the details are summarised here.”→ Consider providing a reference or link to the published SAP.Clarity improvement for readers.
The published work is robust and methodologically sound. An informed reader should weigh the minor reporting gaps (no explicit CONSORT reference, lack of race/ethnicity data) and the unresolved SAP reference, but none of these undermine the primary conclusions. No erratum or re-analysis is warranted based on the rigor review.
- 1.HIGHreportingAdd an explicit statement in the Methods or a dedicated section that the trial is reported in accordance with the CONSORT statement, and include the CONSORT flow diagram as a supplementary figure.Both reviewers noted the absence of an explicit reporting guideline reference, which is a standard expectation for RCTs and would enhance transparency.
- 2.HIGHdata codeVerify the reference to the statistical analysis plan (DOI 10.5281/zenodo.6586525) and correct it if the DOI is wrong, or provide a working link to the published SAP.The citation check flagged this reference as not found in any registry, which could indicate a fabrication or an incorrect DOI; a correct, verifiable reference is essential.
- 3.MEDIUMreportingAdd a sentence in the Discussion or Methods acknowledging the lack of race/ethnicity data and discussing the generalizability of findings to non-White populations.While the absence is already noted, a more explicit discussion of generalizability would strengthen the manuscript's transparency.
- 4.MEDIUMdata codeProvide more detail on the data sharing application process, such as expected response times or criteria for approval, in the Data Sharing section.A more detailed data access process would help readers understand how to request the data and improve reproducibility.
- 5.MEDIUMreportingAdd a link or reference to the full statistical analysis plan in the Methods where it says 'The full statistical analysis plan is published elsewhere'.Providing a direct link to the SAP would improve transparency and allow readers to verify the pre-specified analyses.
- 6.MEDIUMstatisticsReport exact p-values for secondary outcomes that are currently reported as '<0.0001' instead of threshold values.Exact p-values provide more information and allow readers to assess the strength of evidence more precisely.
- 7.MEDIUMreportingClarify the denominator in the Results section where it says '1267 (86%) of 1478 had a Gleason score of 7 or higher' by noting that data were missing for some participants.The copyedit flagged this as a potential inconsistency; clarifying the denominator would avoid confusion.
- 8.LOWcopyeditRephrase the power calculation sentence in the Abstract for clarity: 'The trial had 80% power, with a two-sided α of 5%, to detect an absolute increase...'Minor punctuation and clarity improvement flagged by the copyedit pass.
- 9.LOWcopyeditAdd 'to receive' in the Abstract Findings: '1480 patients were randomly assigned to receive'.Minor grammatical improvement for clarity.
- 10.LOWreportingConsider adding a forest plot for subgroup analyses to visually present interaction p-values.A forest plot would make the subgroup results easier to interpret and is a common practice in RCT reports.
- 11.LOWdata codeAdd a statement about the availability of the statistical analysis code, even if not publicly shared, to enhance reproducibility.Even a statement that code is available on request would improve transparency.
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.