This page separates reported trial information from statistical interpretation. Numerical values are restricted to the ClinicalTrials.gov record. The registry reports results for seven outcome measures and one formal statistical analysis, but the ClinicalTrials.gov record is for a secondary endpoint rather than the registered primary endpoint.
1. Trial at a Glance
ACOSOG Z1071 was a completed phase 2, single-group diagnostic study in breast cancer. The trial enrolled 756 participants and evaluated sentinel lymph node biopsy and axillary lymph node surgery after systemic chemotherapy and neoadjuvant therapy.
| Feature | ACOSOG Z1071 |
|---|---|
| Trial name | ACOSOG Z1071 |
| ClinicalTrials.gov identifier | NCT00881361 |
| Phase | Phase 2 |
| Status | Completed |
| Therapeutic area | Oncology |
| Condition | Breast Cancer |
| Enrollment | 756 |
| Allocation | NA |
| Design model | Single group |
| Masking | None |
| Primary purpose | Diagnostic |
| Lead sponsor | Alliance for Clinical Trials in Oncology |
| Start | 2009-07 |
| Primary completion | 2012-11 |
2. Clinical Question
The trial addressed a diagnostic and surgical question in women with stage II, stage IIIA, or stage IIIB breast cancer undergoing chemotherapy. The registered primary endpoint focused specifically on the false-negative rate of sentinel lymph node surgery after chemotherapy when at least 2 sentinel lymph nodes were excised in women initially presenting with biopsy-proven cN1 breast cancer.
Population
Women with stage II, stage IIIA, or stage IIIB breast cancer, including the registered cN1 cohort defined by an initial presentation with biopsy-proven cN1 breast cancer.
Intervention pathway
Systemic chemotherapy and neoadjuvant therapy followed by sentinel lymph node biopsy and therapeutic conventional surgery, with axillary lymph node dissection and ultrasound imaging represented among the registered interventions.
Comparator structure
The registry identifies the overall study as a single-group design with one arm rather than a randomized parallel-group treatment comparison.
Primary question
What is the false-negative rate for sentinel lymph node surgery after chemotherapy when at least 2 sentinel lymph nodes are excised in the registered cN1 cohort?
3. Trial Design
The design has an important statistical implication: this is not a randomized treatment-versus-control efficacy trial. The principal scientific target is diagnostic performance of sentinel lymph node surgery in a defined clinical setting. Consequently, the primary false-negative-rate endpoint is conceptually different from a conventional survival endpoint or a randomized treatment-effect estimate.
4. Endpoints
| Endpoint | Registry definition / time frame | Type |
|---|---|---|
| False Negative Rate (FNR) for Sentinel Lymph Node (SLN) Surgery After Chemotherapy When at Least 2 SLNs Were Excised in Women Initially Presenting With Biopsy-proven cN1 Breast Cancer [cN1 Cohort] | At time of surgery. False negative rate (FNR) is the percentage defined as the number of patients with no positive lymph nodes by SLN and with at least one positive lymph node by axillary lymph node dissection (ALND), divided by the total number of patients initially. | Binary |
The ClinicalTrials.gov record identifies 1 registered primary endpoint and 7 posted outcome measures. The primary endpoint is binary in structure because each participant contributes to a classification based on whether the sentinel lymph node assessment was false negative according to the subsequent axillary lymph node dissection findings.
5. Statistical Methodology
Primary endpoint: false-negative rate
The registered primary endpoint is a proportion. For the cN1 cohort, the false-negative rate is defined as the number of patients whose sentinel lymph node assessment found no positive lymph nodes while axillary lymph node dissection found at least one positive lymph node, divided by the total number of patients initially defined for the endpoint.
The resulting quantity is expressed as a percentage. It describes the frequency with which sentinel lymph node surgery fails to identify nodal disease that is subsequently identified by axillary lymph node dissection under the registered definition.
The ClinicalTrials.gov record does not provide a numerical FNR estimate, confidence interval, or formal p-value for this registered primary endpoint. The record does state that results are posted, but the one statistical-analysis record registry-reported is for a secondary endpoint.
Posted statistical analysis
The single formal statistical analysis reported by the registry concerns the secondary outcome measure Node Status After Preoperative Chemotherapy, as Measured by the Total Number of Positive Nodes (SLN+ALND). The analysis included patients with post-chemotherapy AUS results and compared patients classified as having a post-chemotherapy AUS result of Normal versus Suspicious.
| Element | Posted analysis |
|---|---|
| Outcome measure | Node Status After Preoperative Chemotherapy, as Measured by the Total Number of Positive Nodes (SLN+ALND) |
| Time frame | At time of surgery |
| Analysis population | Patients with post-chemotherapy AUS results |
| Groups compared | Post-chemotherapy AUS Normal vs Post-chemotherapy AUS Suspicious |
| Method | t-test, 2 sided |
| Hypothesis type | Superiority |
| P-value | 0.003 |
6. Results
Primary endpoint: false-negative rate
The registry identifies the false-negative rate as the primary endpoint and indicates that results were posted for it. However, the ClinicalTrials.gov record does not contain a numerical FNR estimate, confidence interval, or p-value. Because the numerical result is not included in the ClinicalTrials.gov record, no numerical primary-endpoint result is reported here.
The important statistical distinction is between an endpoint having posted results and a formal statistical-analysis record containing an estimate and inferential statistics.
For a diagnostic endpoint such as FNR, the estimate itself is the central descriptive quantity. A confidence interval would add information about precision, while a formal hypothesis test would address a prespecified null hypothesis if one had been defined and reported. Neither a numerical estimate nor such inferential quantities are posted on ClinicalTrials.gov for the primary endpoint here.
Accordingly, the absence of a numerical primary result in this page should not be interpreted as evidence for or against any particular FNR value. It is a limitation of the ClinicalTrials.gov record for this analysis.
Secondary endpoint: total number of positive nodes by post-chemotherapy AUS category
Two-sided t-test
Comparison: Post-chemotherapy AUS Normal vs Post-chemotherapy AUS Suspicious
Outcome: Total number of positive nodes (SLN+ALND)
The posted analysis reports a two-sided t-test comparing the total number of positive nodes between the post-chemotherapy AUS Normal and Suspicious groups among patients with post-chemotherapy AUS results.
A two-sided P = 0.003 indicates that, under the statistical test as specified in the registry, the observed difference in the analyzed outcome is inconsistent with the null hypothesis of no difference at conventional significance levels.
The p-value does not quantify the size of the difference in the number of positive nodes. It also does not give the probability that the null hypothesis is true, nor does it tell us how clinically important the observed difference is.
Precision cannot be evaluated from the ClinicalTrials.gov record because no group means, mean difference, standard error, or confidence interval are provided. The p-value therefore supplies evidence against the null hypothesis but does not by itself describe the magnitude or practical importance of the group difference.
The analysis is also observational with respect to the post-chemotherapy AUS classification: the study record identifies a single-group design, and the two groups arise from the AUS result rather than from randomized assignment to Normal versus Suspicious ultrasound status. Thus the comparison should not be interpreted as a randomized treatment effect.
7. Statistical Methods Explained
Why is a t-test reported for this endpoint?
The registry reports a two-sided t-test for the total number of positive nodes, comparing the post-chemotherapy AUS Normal and Suspicious groups. A two-sample t-test is designed to compare the means of a quantitative outcome between two groups under assumptions concerning the sampling distribution and variance structure. The registry does not provide the group means or the exact variance specification, so those details cannot be reconstructed from the ClinicalTrials.gov record.
What does P = 0.003 mean?
Under the null hypothesis represented by the reported two-sided t-test, a result at least as inconsistent with that null as the observed result has a reported p-value of 0.003. This is evidence against the null hypothesis under the test framework. It is not a measure of effect size.
Why doesn't the p-value tell us how large the difference was?
The p-value combines information about the observed difference with information about variability and sample information. A small p-value can accompany a modest effect when the estimate is precise, while a large effect can be estimated imprecisely. To understand magnitude, one would want the group-specific estimates and their difference, ideally with a confidence interval.
Why is the primary endpoint different from the t-test endpoint?
The registered primary endpoint is a false-negative rate, which is a binary diagnostic-performance measure. The posted t-test instead concerns the total number of positive nodes and compares that outcome across two post-chemotherapy AUS categories. These are different statistical questions and should not be conflated.
Why does the single-group design matter?
In a single-group study, participants are not randomized between the post-chemotherapy AUS categories. A comparison between AUS Normal and AUS Suspicious therefore describes differences associated with those observed categories rather than a randomized intervention effect. This distinction is important when interpreting a statistically significant p-value.
What does a false-negative rate measure?
The FNR measures the proportion of relevant cases in which the sentinel lymph node procedure is negative while the reference axillary lymph node dissection identifies at least one positive lymph node, using the definition registered for this study. It is therefore a diagnostic-performance measure, not a treatment-effect measure.
8. Safety Results
The ClinicalTrials.gov record reports serious adverse events by cohort rather than by randomized treatment arm. Because ACOSOG Z1071 is recorded as a single-group study, these figures should not be interpreted as comparative treatment-arm safety rates.
| Cohort | Participants with serious adverse events / at risk |
|---|---|
| cN1 Cohort | 6/676 |
| cN2 Cohort | 0/39 |
These counts describe the serious-adverse-event information posted on ClinicalTrials.gov for the two named cohorts. The ClinicalTrials.gov record does not include a larger adverse-event table, event-specific breakdown, severity distribution, or randomized arm comparison.
9. Design Topics and What the Registry Does Not Establish
Randomization
The registry identifies allocation as NA and the design model as single group. No randomized treatment comparison is reported.
Factorial design
No factorial design is identified in the ClinicalTrials.gov record.
Crossover
No crossover design or crossover analysis is identified in the ClinicalTrials.gov record.
Non-inferiority margin
No non-inferiority hypothesis or margin is reported. The registry-reported hypothesis type is superiority.
Bayesian methods
No Bayesian method is reported. The registry-reported formal analysis is a two-sided t-test.
Interim analysis
No interim-analysis method or alpha-spending procedure is reported in the ClinicalTrials.gov record.
Missing-data methods
No imputation method or missing-data strategy is reported in the ClinicalTrials.gov record.
Multiplicity
No multiplicity-adjustment procedure is posted on ClinicalTrials.gov for the posted analysis.
10. Analysis Population and Cohort Structure
The primary endpoint is explicitly restricted to women initially presenting with biopsy-proven cN1 breast cancer and requires that at least 2 sentinel lymph nodes were excised. The posted secondary statistical analysis uses a different population definition: patients with post-chemotherapy AUS results.
| Analysis | Population definition reported | Statistical role |
|---|---|---|
| Primary FNR endpoint | Women initially presenting with biopsy-proven cN1 breast cancer; at least 2 SLNs excised | Registered primary endpoint |
| Secondary node-status analysis | Patients with post-chemotherapy AUS results | Formal statistical analysis posted |
| Safety information | cN1 and cN2 cohorts | Serious adverse events by cohort |
This distinction matters because the denominator and eligibility criteria determine what a statistic means. A false-negative rate calculated in the registered cN1 cohort cannot automatically be transferred to the broader enrollment population, and a p-value from the AUS comparison cannot be treated as a test of the primary FNR endpoint.
11. Interpreting the Diagnostic Endpoint
The registered FNR counts these discordant cases relative to the total number of patients initially evaluated for the endpoint.
The diagnostic structure can be understood as a comparison between a sentinel lymph node assessment and the subsequent axillary lymph node dissection finding used in the registry definition. A false negative occurs when the sentinel procedure does not identify positive nodes but the axillary dissection does identify at least one positive node.
That construction is different from a conventional treatment comparison. There is no treatment hazard ratio to interpret for the primary endpoint, and there is no randomized control arm against which the sentinel procedure is compared. The statistical target is instead the frequency of an important diagnostic classification error under the specified surgical conditions.
A point estimate of an FNR would summarize the observed frequency of false negatives, but the estimate alone would not indicate how precisely the trial measured that quantity. A confidence interval would provide a sampling-based description of uncertainty around the observed proportion. Its width would depend on the amount of information contributing to the endpoint and the observed event frequency.
Because the ClinicalTrials.gov record does not contain the numerical FNR estimate or confidence interval, it would be inappropriate to substitute the secondary t-test p-value for that missing diagnostic information. The two statistics answer different questions.
12. Understanding the t-Test in Context
The posted secondary analysis provides a useful example of how a familiar statistical test can coexist with a more specialized primary endpoint.
| Question | What the registry-reported analysis tells us | What it does not tell us |
|---|---|---|
| Are the two AUS groups statistically distinguishable on the analyzed outcome? | The two-sided t-test reports P = 0.003. | The p-value alone does not give the size of the difference. |
| How many positive nodes differed between groups? | The outcome was the total number of positive nodes (SLN+ALND). | Group means and the mean difference are not reported. |
| How precise was the estimated difference? | Not determinable from the ClinicalTrials.gov record. | No confidence interval or standard error is reported. |
| Was AUS category randomized? | No; the overall study is identified as single group. | The comparison should not be interpreted as a randomized treatment effect. |
The phrase superiority in the registry's hypothesis-type field should also be interpreted carefully. It describes the hypothesis framework attached to the posted analysis, but it does not transform the observationally defined AUS categories into randomized intervention groups.
13. Limitations
- Primary numerical result limitation: the ClinicalTrials.gov record identifies the FNR as the registered primary endpoint but does not provide its numerical estimate, confidence interval, or p-value in the ClinicalTrials.gov record.
- Single-group design: the registry identifies a single-group design, so the study should not be interpreted as a randomized comparison of two treatments.
- Secondary analysis population: the t-test includes patients with post-chemotherapy AUS results, which is not the same population definition as the registered cN1 primary endpoint.
- Incomplete effect-size information: the posted t-test p-value of 0.003 is not accompanied in the ClinicalTrials.gov record by group means, a mean difference, a standard error, or a confidence interval.
- Assumptions of the t-test: the ClinicalTrials.gov record names the test but does not provide diagnostics or details sufficient to evaluate distributional or variance assumptions.
- Diagnostic interpretation: the FNR is a diagnostic-performance measure and should not be interpreted as a treatment efficacy statistic.
- Safety detail: serious adverse events are reported only as cohort-level affected/at-risk counts; more detailed safety characterization is not available in the ClinicalTrials.gov record.
- Multiplicity: no multiplicity-adjustment strategy is posted on ClinicalTrials.gov for the posted statistical analysis, so no broader familywise-error interpretation should be inferred.
14. Why This Trial Matters Statistically
ACOSOG Z1071 is useful as a teaching case because it illustrates a statistical problem that is easy to confuse with a conventional treatment trial. The study asks a diagnostic-performance question in a clinically defined cohort, while one of the posted analyses uses a two-group comparison of a quantitative node-count outcome.
| Concept | How it appears in ACOSOG Z1071 |
|---|---|
| Diagnostic accuracy | The primary endpoint is a false-negative rate for sentinel lymph node surgery. |
| Binary endpoint | The registered primary endpoint is classified as binary. |
| Reference procedure | False-negative classification uses axillary lymph node dissection findings in the registered definition. |
| Single-group design | The registry identifies the design model as SINGLE_GROUP. |
| t-test | The posted secondary analysis uses a two-sided t-test. |
| P-value | The formal analysis reports P = 0.003. |
| Analysis population | The posted t-test includes patients with post-chemotherapy AUS results. |
| Endpoint distinction | The formal t-test is secondary and is not the registered primary FNR endpoint. |
| Cohort-specific safety | Serious adverse events are reported for cN1 and cN2 cohorts. |
15. Statistical Interpretation vs Clinical Interpretation
Statistical interpretation
The registry-reported formal statistical analysis reports a two-sided t-test with P = 0.003 for the comparison of total positive nodes between post-chemotherapy AUS Normal and Suspicious groups.
Clinical interpretation
The clinical meaning of that comparison depends on the magnitude of the difference in positive nodes, the precision of the estimates, and the diagnostic context. Those quantitative effect-size details are not reported in the ClinicalTrials.gov record.
The distinction is important: statistical evidence against a null hypothesis does not automatically establish the magnitude, clinical importance, or causal interpretation of an association. Those questions require the effect estimate, uncertainty, study design, and endpoint context.
16. Longitudinal Trial History
Trial begins
The ClinicalTrials.gov record lists July 2009 as the study start.
Primary completion
The ClinicalTrials.gov record lists November 2012 as the primary completion date.
Results posted
The trial is identified as completed, with results posted and seven outcome measures reported in the ClinicalTrials.gov record.
17. What the P-Value Does — and Does Not — Mean
The reported P = 0.003 comes from a two-sided t-test comparing the total number of positive nodes between the post-chemotherapy AUS Normal and Suspicious groups. Under the null hypothesis represented by that test, the observed result provides evidence against equality of the compared group means.
It does not mean that there is a 0.3% probability that the null hypothesis is true. It also does not mean that the difference is 0.3%, nor does it quantify the clinical magnitude of the difference.
A p-value is incomplete as a description of a quantitative comparison. To understand the result fully, a reader would ideally see the estimated mean number of positive nodes in each group, their difference, a confidence interval, and the assumptions underlying the t-test.
Because the AUS categories were not created through randomized assignment in the registry-reported single-group design, the t-test should be read as a comparison between observed groups rather than evidence that changing an intervention from one condition to another would cause a change in node status.
18. Primary Endpoint vs Posted Statistical Analysis
| Feature | Registered primary endpoint | Posted statistical analysis |
|---|---|---|
| Clinical question | False-negative rate of SLN surgery after chemotherapy | Total number of positive nodes by post-chemotherapy AUS category |
| Time frame | At time of surgery | At time of surgery |
| Endpoint type | Binary | Binary in the registry's outcome classification |
| Population | Initially biopsy-proven cN1 breast cancer with at least 2 SLNs excised | Patients with post-chemotherapy AUS results |
| Groups | Diagnostic classification within the cN1 cohort | Post-chemotherapy AUS Normal vs Suspicious |
| Formal statistical method reported | None in the ClinicalTrials.gov record | Two-sided t-test |
| Reported p-value | Not reported | 0.003 |
This side-by-side view is one of the most important statistical features of the record. The formal p-value should remain attached to the secondary analysis for which it was reported. It should not be reassigned to the primary FNR question.
19. Related Tutorials
Learn more about the methods used in this trial:
20. Related Statistical Calculators
21. Sources
- ClinicalTrials.gov: NCT00881361 — ACOSOG Z1071.
- PubMed: PMID 24101169.
- PubMed: PMID 25646192.
- PubMed: PMID 29980907.
- PubMed: PMID 28657941.
- PubMed: PMID 25664534.
Continue through the Clinical Biostats statistical pathway
Explore the statistical concepts that connect diagnostic endpoints, hypothesis testing, p-values, and t-tests to broader clinical-trial analysis.
22. Record Summary
ACOSOG Z1071 provides a useful statistical example of a phase 2 diagnostic study in which the principal registered question is a false-negative rate rather than a randomized treatment effect. The registry defines the primary endpoint as the percentage of patients with no positive lymph nodes by sentinel lymph node surgery but at least one positive lymph node by axillary lymph node dissection, among women initially presenting with biopsy-proven cN1 breast cancer when at least 2 sentinel lymph nodes were excised.
The ClinicalTrials.gov record also contains one formal statistical analysis for a secondary endpoint: a two-sided t-test comparing total positive nodes between patients with post-chemotherapy AUS results classified as Normal versus Suspicious. That analysis reports P = 0.003. The p-value provides evidence against the null hypothesis represented by the test, but the ClinicalTrials.gov record does not include the group-specific estimates, difference, confidence interval, or standard error needed to quantify the magnitude and precision of the finding.
The most important analytical lesson is therefore to keep the endpoints and populations separate. The primary FNR question is a diagnostic-performance problem, while the posted t-test addresses a secondary quantitative comparison. Treating the secondary p-value as though it were evidence about the primary FNR endpoint would change the statistical question being answered.