This page provides an independent statistical analysis and educational interpretation of publicly reported results. ClinicalTrials.gov provides the official trial registry record.
1. Trial at a Glance
TROPION-Lung05 was a phase 2, single-group study evaluating DS-1062a in participants with advanced or metastatic non-small cell lung cancer with actionable genomic alterations. The registered primary endpoint was the percentage of participants with objective response rate (ORR) based on blinded independent central review (BICR).
| Feature | TROPION-Lung05 |
|---|---|
| Phase | Phase 2 |
| Condition | Non-small Cell Lung Cancer |
| Study population | Participants with advanced or metastatic non-small cell lung cancer with actionable genomic alterations |
| Design model | Single group |
| Allocation | Not applicable |
| Masking | None |
| Primary purpose | Treatment |
| Enrollment | 137 |
| Intervention | DS-1062a |
| Primary endpoint type | Binary |
| Trial status | Completed |
| Lead sponsor | Daiichi Sankyo |
| ClinicalTrials.gov | NCT04484142 |
2. Clinical Question
The central registered question is descriptive and response-focused: among participants with advanced or metastatic non-small cell lung cancer with actionable genomic alterations who receive DS-1062a, what proportion have a confirmed complete response or confirmed partial response according to the registered objective response definition?
Population
Participants with advanced or metastatic non-small cell lung cancer with actionable genomic alterations.
Intervention
DS-1062a, listed in the registry as a drug intervention.
Comparator
There was no comparator arm. The study used a single-group design.
Primary question
What percentage of participants achieve an objective response based on blinded independent central review?
This distinction is important statistically. A single-group phase 2 study estimates the observed response experience in the treated study population; it does not provide a randomized between-treatment comparison. Without a comparator group, a response proportion cannot by itself establish how outcomes would differ under another treatment or under a different standard of care.
3. Trial Design
The registry identifies the study as completed, with a start date of 2021-03-30 and a primary completion date of 2023-03-10. The enrollment was 137 participants, and there was one study arm.
Single-group design and what it changes
Because all enrolled participants belong to the same study group, the primary statistical problem is estimation of a response proportion rather than comparison of two randomized proportions. The natural estimand is the proportion of participants meeting the prespecified definition of objective response during the specified assessment period.
This design can be informative when the objective is to characterize antitumor activity, but interpretation depends strongly on the external context used to judge whether the observed response proportion is clinically or statistically meaningful. That external context is not reported in the ClinicalTrials.gov record here, so this page does not introduce a historical-control response rate or an unstated benchmark.
4. Endpoints
| Endpoint | Registry definition | Time frame | Type |
|---|---|---|---|
| Percentage of Participants With Objective Response Rate (ORR) Based on Blinded Independent Central Review (BICR) | ORR is defined as the proportion of participants with a best overall response of confirmed complete response (CR) or confirmed partial response (PR) per Response Evaluation Criteria in Solid Tumors (RECIST) version 1.1. | From baseline until disease progression, death, or other protocol defined reason, up to approximately 24 months. | Binary |
The registered primary endpoint is therefore a participant-level binary classification. Each participant contributes to the numerator if their best overall response satisfies the confirmed CR or confirmed PR definition and to the denominator according to the prespecified analysis population and endpoint rules.
5. Statistical Methodology
The registry identifies the primary endpoint as binary, but the ClinicalTrials.gov record does not contain a formal statistical analysis entry, effect estimate, confidence interval, or p-value for that endpoint. The appropriate statistical framework can therefore be described without presenting an outcome that is not contained in the ClinicalTrials.gov record.
Primary estimand: response proportion
For a binary objective-response endpoint, the basic estimator is the observed proportion of participants who meet the response definition:
where x is the number of participants classified as responders and n is the number of participants in the prespecified analysis population.
The registry supplies the total enrollment of 137, but it does not supply the responder count in the ClinicalTrials.gov record. Consequently, an observed ORR cannot be calculated from the ClinicalTrials.gov record without introducing an external result.
Confidence interval for a single proportion
A confidence interval is ordinarily presented alongside a single-arm response proportion because the observed percentage is an estimate rather than a known population quantity. For a binary endpoint, exact binomial methods are one established approach, while other binomial confidence-interval methods can also be used depending on the statistical analysis plan.
The choice of interval method matters when sample sizes or event proportions produce asymmetric or otherwise non-normal sampling behavior. A confidence interval should therefore be reported together with the response estimate rather than replacing the estimate itself.
Hypothesis testing in a single-arm study
If a single-arm phase 2 study formally tests activity, the response proportion can be compared with a prespecified reference value under a binomial model. Such a test requires an explicitly defined null response rate. No such reference value is contained in the ClinicalTrials.gov record, so no hypothesis test or p-value is reported here.
Role of BICR
The primary endpoint is based on blinded independent central review. From a statistical perspective, independent central assessment provides a prespecified mechanism for classifying radiographic response separately from the treating investigator's assessment. The endpoint remains binary at the participant level even though the underlying tumor assessments are longitudinal and involve RECIST version 1.1 response classification.
6. Planned Analysis
Results are posted for the trial, and the primary endpoint is marked as having results posted, but the ClinicalTrials.gov record contains no formal statistical analyses.
What would normally be reported?
| Component | Typical statistical role | Available in the ClinicalTrials.gov record? |
|---|---|---|
| Responder count | Numerator for ORR | No |
| Analysis population size | Denominator for ORR | Total enrollment is 137; the endpoint-specific analysis population is not reported |
| ORR estimate | Observed proportion with confirmed CR or PR | No formal numerical result reported |
| Confidence interval | Quantifies uncertainty around the response proportion | No |
| Reference response rate | Required for a formal single-arm hypothesis test | No |
| P-value | Quantifies evidence against a prespecified null hypothesis | No |
This is an important distinction between results posted and formal statistical comparison posted. A registry may contain outcome-measure results without providing the inferential statistical-analysis objects needed to reconstruct a formal hypothesis test from the record alone.
7. Safety Results
The ClinicalTrials.gov record reports serious adverse events by arm as 34 affected participants out of 137 at risk for Dato DXd 6.0 mg/kg Q3W.
Serious adverse events
Dato DXd 6.0 mg/kg Q3W: affected / at risk
The corresponding crude proportion, if expressed arithmetically, would require calculation from the reported counts. The underlying counts are retained here exactly as reported in the registry rather than introducing a newly rounded percentage.
8. Statistical Methods Explained
Why is ORR treated as a binary endpoint?
The registry defines objective response as the proportion of participants whose best overall response is a confirmed complete response or confirmed partial response according to RECIST version 1.1. At the participant level, this creates a binary classification: responder or non-responder.
Why is a single proportion different from a treatment effect?
In a randomized trial, a treatment effect is commonly expressed as a contrast between treatment groups. In this single-group study, there is no concurrent randomized comparator. The primary numerical quantity is therefore the response proportion in the study population rather than a treatment-versus-control effect estimate.
What does BICR add to the analysis?
Blinded independent central review provides an independent radiologic assessment framework for the endpoint. This can reduce reliance on local investigator assessment when determining whether a participant meets the RECIST-based response definition. It does not, however, convert a single-group study into a randomized comparison.
Why does the confidence interval matter?
A response percentage calculated from a finite sample is subject to sampling uncertainty. A confidence interval communicates how precisely the observed response proportion estimates the underlying response probability under the chosen statistical framework. It should not be interpreted as a range containing the response probability of every individual participant.
Why can't a p-value be reported without a reference response rate?
A formal one-sample binomial test requires a null hypothesis, commonly expressed as a prespecified reference response probability. Without that reference value, there is no defined null against which the observed response proportion can be tested. The ClinicalTrials.gov record does not provide such a value.
Why is the approximately 24-month time frame important?
The endpoint is assessed from baseline until disease progression, death, or another protocol-defined reason, up to approximately 24 months. This prevents the endpoint from being interpreted as an unrestricted lifetime response measure. The time frame defines the observation window specified for the registered endpoint.
Why should enrollment not automatically be treated as the ORR denominator?
The trial enrolled 137 participants, but the ClinicalTrials.gov record does not explicitly state the endpoint-specific analysis population. In a complete statistical report, the denominator should follow the prespecified analysis rules for the endpoint rather than being assumed from enrollment alone.
9. Interpreting a Future ORR Result Correctly
If a formal ORR estimate is reported in a statistical analysis, several quantities should be kept conceptually separate.
The estimate
The ORR is the proportion of participants with confirmed CR or PR according to the registered RECIST version 1.1 definition.
The confidence interval
The confidence interval describes statistical uncertainty around the estimated response proportion under the selected binomial method.
The p-value
If a hypothesis test was prespecified, the p-value would describe evidence against the specified null response rate. It would not measure the size of the response.
The clinical context
A response proportion does not by itself describe duration of response, progression-free survival, overall survival, or comparative effectiveness against another treatment.
This separation is especially important for a single-arm study. A large observed response proportion may be clinically interesting, but statistical interpretation still depends on the analysis population, the precision of the estimate, the prespecified reference value if hypothesis testing is used, and the clinical context in which the result is judged.
10. Binary Endpoint Mechanics
The registered ORR is the proportion of participants whose best overall response satisfies either of these two categories under RECIST version 1.1.
This structure is statistically different from endpoints such as progression-free survival, where each participant contributes a time-to-event observation that may be censored. The registered primary endpoint for TROPION-Lung05 is binary, so the fundamental analysis is based on counts and proportions.
The approximately 24-month assessment window nevertheless matters operationally because response classification is not simply a baseline characteristic. Participants are followed from baseline until disease progression, death, or another protocol-defined reason, and the best overall response observed during the specified period determines the endpoint classification.
11. Analysis Population and Missing Data Considerations
The ClinicalTrials.gov record does not specify a separate analysis population definition for the primary endpoint, nor do they provide missing-data or imputation rules. Therefore, no trial-specific imputation strategy is asserted here.
For a binary response endpoint, missing or unevaluable assessments can affect both the numerator and denominator depending on the prespecified analysis rules. Common approaches include defining non-responders under particular missingness rules, restricting the denominator to an evaluable population, or using other protocol-defined conventions. These approaches can produce different estimates and should never be inferred after seeing the observed results.
12. Blinding and Central Review
The study is listed as having no masking, while the primary endpoint is based on blinded independent central review. These statements describe different aspects of the study.
| Feature | Registry information | Statistical implication |
|---|---|---|
| Overall study masking | None | The trial itself is not described as masked. |
| Primary response assessment | Blinded independent central review | The primary response classification uses an independent central-review process. |
| Endpoint type | Binary | The final participant-level response classification is summarized as a proportion. |
The important analytical point is that study-level masking and endpoint-assessment procedures are not synonymous. An unmasked treatment study can still specify an independent blinded review process for radiographic response assessment.
13. No Randomized Comparator: Consequences for Interpretation
The single-group design is one of the most important statistical characteristics of TROPION-Lung05.
No between-arm hazard ratio
There is no randomized comparator arm in the registry-reported design, so a treatment-versus-control hazard ratio is not an appropriate primary result for this study.
No randomized risk difference
A risk difference comparing DS-1062a with another treatment cannot be estimated from the single-group design alone.
External comparisons require assumptions
Comparing a single-arm response rate with results from another study requires attention to differences in eligibility, assessment, follow-up, patient characteristics, and endpoint definitions.
Primary estimand is descriptive
The directly supported quantity is the response proportion under the registered ORR definition.
This does not make a single-arm study statistically uninformative. Rather, the inferential target is different. The design is naturally suited to estimating activity within the studied population and, when a prespecified external benchmark exists, testing whether the observed response proportion is sufficiently different from that benchmark.
14. Design Features Not Supported by the Supplied Data
The ClinicalTrials.gov record does not provide evidence for several design features that are sometimes central to clinical-trial statistical interpretation. They should therefore not be attributed to TROPION-Lung05 on this page.
| Design topic | What can be stated from the ClinicalTrials.gov record? |
|---|---|
| Non-inferiority margin | No non-inferiority margin is provided. |
| Crossover | No crossover information is provided. |
| Factorial design | The study is listed as single group, not as a factorial design. |
| Interim analysis | No interim-analysis plan or result is provided. |
| Multiplicity | No multiplicity-adjustment strategy is provided. |
| Missing-data imputation | No trial-specific imputation strategy is provided. |
| Stratification | No stratification factors are provided. |
| Bayesian methods | No Bayesian analysis is provided. |
These omissions are not evidence that such procedures were absent from every underlying trial document; they mean only that they are not supported by the ClinicalTrials.gov record and therefore are not presented as trial-specific methods here.
15. Trial Timeline
Study start
The TROPION-Lung05 study began on 2021-03-30.
Primary completion
The registry lists 2023-03-10 as the primary completion date.
Trial status
The study is listed as completed, with results posted for the registered primary endpoint.
16. What This Trial Can and Cannot Establish Statistically
| Question | What the design can address | What requires additional evidence |
|---|---|---|
| How often did participants meet the ORR definition? | A single-group response proportion can directly address this. | A numerical estimate is not reported in the ClinicalTrials.gov record. |
| How precise is the ORR estimate? | A confidence interval can quantify uncertainty. | The confidence interval is not reported. |
| Is ORR different from a prespecified benchmark? | A one-sample binomial test can address this if a null response rate was prespecified. | No reference response rate or p-value is reported. |
| Is DS-1062a better than another treatment? | Not through randomized comparison in this single-group design. | A concurrent comparator or an appropriately justified external comparison would be needed. |
| Does ORR establish longer survival? | No. ORR measures response status. | Time-to-event endpoints such as progression-free or overall survival would be needed for those questions. |
17. Limitations
- Single-group design: there is no randomized comparator, so observed response cannot be interpreted as a randomized treatment effect.
- No formal statistical-analysis results reported: the ClinicalTrials.gov record contain no statistical-analysis entries despite indicating that results are posted.
- No responder count reported: the numerator needed to calculate the registered ORR is not included in the ClinicalTrials.gov record.
- No confidence interval reported: the precision of the primary response estimate cannot be reconstructed without the underlying response count and analysis population.
- No reference response rate reported: a formal one-sample hypothesis test cannot be reconstructed from the ClinicalTrials.gov record.
- Endpoint-specific denominator not reported: total enrollment is 137, but the ClinicalTrials.gov record does not establish the denominator used for the primary endpoint analysis.
- Assessment window: the registered ORR is assessed from baseline until disease progression, death, or another protocol-defined reason, up to approximately 24 months.
- Safety interpretation: serious adverse events are reported descriptively for the single study group and cannot be converted into a randomized comparative safety effect.
- Unreported design details: the ClinicalTrials.gov record does not specify interim-analysis, multiplicity, stratification, imputation, Bayesian, crossover, or non-inferiority procedures.
18. Why This Trial Matters Statistically
TROPION-Lung05 is a useful teaching case because it illustrates a statistical situation that is easy to misunderstand: a phase 2 study can have a clearly defined binary efficacy endpoint and substantial enrollment while still requiring a fundamentally different interpretation from a randomized controlled trial.
| Concept | How it appears in TROPION-Lung05 |
|---|---|
| Single-arm design | One study group with 137 participants and no comparator arm. |
| Binary endpoint | ORR is defined by whether the best overall response is confirmed CR or confirmed PR. |
| RECIST classification | Response is defined according to RECIST version 1.1. |
| Independent assessment | The primary endpoint is based on blinded independent central review. |
| Proportion estimation | The natural primary summary is the percentage of participants meeting the response definition. |
| Confidence interval | Precision around a single response proportion should ordinarily be reported. |
| One-sample hypothesis testing | A formal test would require a prespecified reference response rate. |
| Analysis population | The endpoint-specific denominator should follow prespecified rules rather than being assumed from enrollment. |
| Safety description | Serious adverse events are reported as affected participants out of those at risk. |
The broader lesson is that the statistical design determines what an observed result means. A percentage from a single group is not interchangeable with a randomized treatment effect, and a p-value cannot be interpreted without knowing the null hypothesis and analysis framework that produced it.
19. Related Tutorials
Learn more about the methods used in this trial:
20. Related Statistical Calculators
21. Sources
- ClinicalTrials.gov: TROPION-Lung05 (NCT04484142).
- Linked publication: PubMed record for PMID 40516821.
- Linked publication: PubMed record for PMID 39761483.
Continue through the Clinical Biostats statistical pathway
Explore the statistical concepts behind binary endpoints, single-arm studies, response assessment, and confidence intervals.
22. Record Summary
TROPION-Lung05 is a completed phase 2, single-group study of DS-1062a in advanced or metastatic non-small cell lung cancer with actionable genomic alterations. The registry reports 137 participants, one study arm, no masking, and a primary binary endpoint defined as the percentage of participants with objective response rate based on blinded independent central review. ORR is defined as the proportion of participants whose best overall response is a confirmed complete response or confirmed partial response according to RECIST version 1.1, assessed from baseline until disease progression, death, or another protocol-defined reason, up to approximately 24 months.
The key statistical feature is the single-group design. The primary analysis is therefore naturally framed around estimation of a response proportion rather than a randomized treatment effect. Although results are posted for the primary endpoint, the ClinicalTrials.gov record contains no formal statistical-analysis entries and no responder count, confidence interval, reference response rate, or p-value. A complete inferential interpretation therefore requires those additional quantities rather than reconstructing them from unsupported assumptions.