This page separates reported trial results from statistical interpretation. Numerical results are taken from the publicly reported ClinicalTrials.gov record. ClinicalTrials.gov provides the official trial registry record.
1. Trial at a Glance
NETTER-2 was a randomized, open-label phase 3 trial evaluating Lutathera plus octreotide LAR 30 mg against octreotide LAR 60 mg in patients with grade 2 and grade 3 advanced gastro-enteropancreatic neuroendocrine tumor. The registered primary endpoint was progression-free survival by central assessment.
| Feature | NETTER-2 |
|---|---|
| Trial name | NETTER-2 |
| Phase | Phase 3 |
| Condition | Gastro-enteropancreatic Neuroendocrine Tumor |
| Brief title | Study to Evaluate the Efficacy and Safety of Lutathera in Patients With Grade 2 and Grade 3 Advanced GEP-NET |
| Allocation | Randomized |
| Design model | Crossover |
| Masking | None |
| Primary purpose | Treatment |
| Enrollment | 226 |
| Arms | 5 |
| Trial status | Active, not recruiting |
| Lead sponsor | Advanced Accelerator Applications |
| Sponsor type | Industry |
| ClinicalTrials.gov | NCT03972488 |
2. Clinical Question
The central statistical question was whether treatment with Lutathera plus octreotide LAR 30 mg improves progression-free survival compared with octreotide LAR 60 mg in patients with grade 2 and grade 3 advanced gastro-enteropancreatic neuroendocrine tumor.
Population
Patients with grade 2 and grade 3 advanced gastro-enteropancreatic neuroendocrine tumor.
Intervention
Lutathera plus octreotide LAR 30 mg, with 2.5% Lys-Arg sterile amino acid solution also listed among the trial interventions.
Comparator
Octreotide LAR 60 mg, identified as the control arm.
Primary question
Does the Lutathera plus octreotide LAR 30 mg strategy improve centrally assessed progression-free survival relative to octreotide LAR 60 mg?
3. Trial Design
Lutathera + Octreotide LAR 30 mg
- Lutathera
- Octreotide LAR 30 mg
- 2.5% Lys-Arg sterile amino acid solution is listed as a trial intervention
Octreotide LAR 60 mg
- Octreotide LAR 60 mg
- Identified in the registry as the control arm
The registry describes the overall study as randomized and uses a crossover design model. The trial is not masked. The registry lists 5 arms overall, while the primary statistical comparison reported for the posted primary analysis is specifically between the Lutathera plus octreotide LAR 30 mg investigational arm and the octreotide LAR 60 mg control arm.
4. Trial Timeline
Trial start
The registered study start date was January 8, 2020.
Primary completion
The registered primary completion date was July 20, 2023.
Active, not recruiting
The ClinicalTrials.gov record identifies the study status as active, not recruiting.
5. Endpoints
| Endpoint | Registry definition / time frame | Type |
|---|---|---|
| Progression Free Survival (PFS) Per Central Assessment | From randomization to the first line progression or death due to any cause, up to approx. 42 months. PFS is the time from randomization to the first line progression, centrally assessed according to RECIST 1.1, or death due to any cause. | Time-to-event |
| Overall Response Rate (ORR) Per Central Assessment (Key Secondary) | Up to approx. 42 months. | Binary |
| Time to Deterioration (TTD) Global Health Status, Diarrhea, Fatigue, Pain (EORTC QLQ-C30) (Key Secondary) | Up to approx. 42 months. | Time-to-event |
Primary endpoint definition
The primary endpoint was Progression Free Survival (PFS) Per Central Assessment. The registry defines PFS as the time from randomization to the first line progression, centrally assessed according to RECIST 1.1, or death due to any cause.
Under the registry definition, progression is defined according to RECIST 1.1 as a 20% increase in the sum of diameters of all measured target lesions, unequivocal progression of non-target lesions, or appearance of a new lesion.
6. Analysis Populations and Statistical Framework
The posted primary analysis uses the Full Analysis Set (FAS), defined in the registry as comprising all participants to whom study treatment has been assigned by randomization. This is important because the primary efficacy comparison remains anchored to the randomized treatment assignment rather than switching participants into groups based on subsequent treatment exposure.
| Component | NETTER-2 statistical framework |
|---|---|
| Primary analysis population | Full Analysis Set (FAS) |
| FAS definition | All participants to whom study treatment has been assigned by randomization |
| Primary endpoint | Progression Free Survival (PFS) Per Central Assessment |
| Endpoint type | Time-to-event |
| Primary comparison | Lutathera plus Octreotide LAR 30 mg vs Octreotide LAR 60 mg |
| Primary method | Log-rank test |
| Effect measure | Hazard ratio |
| Hypothesis | Superiority |
| P-value framework | Stratified one-sided P-value |
The analysis text also identifies intention-to-treat analysis, one-sided testing, and stratified analysis as concepts in the primary analysis. The registry method field reports a log-rank test.
7. Statistical Methodology
Log-rank test
The primary PFS comparison uses a log-rank test. The log-rank test is designed for comparing survival or time-to-event distributions between treatment groups while accounting for the timing of events and censoring.
Rather than reducing each participant to a simple event/no-event indicator, the analysis uses the information contained in when progression or death occurs. Participants who remain event-free at their last evaluable follow-up contribute information up to that point.
The log-rank framework evaluates whether the observed distribution of events over time differs systematically between the randomized treatment groups.
Hazard ratio
The primary treatment effect is expressed as a hazard ratio (HR). For a conventional interpretation, an HR below 1 indicates a lower estimated instantaneous event rate in the investigational group relative to the comparator under the fitted time-to-event framework.
Equivalently, 1 − 0.276 = 0.724, so the estimate corresponds to an approximately 72.4% lower estimated hazard relative to the control under the model. This is a relative hazard interpretation, not a statement that 72.4% of participants avoid progression or death.
Stratified analysis
The primary analysis is explicitly described as a stratified analysis, and the posted analysis notes identify the P-value as a stratified one-sided P-value. Stratification allows the treatment comparison to account for prespecified or relevant grouping factors rather than treating all participants as belonging to one completely homogeneous risk set.
One-sided superiority testing
The primary analysis uses a one-sided P-value under a superiority hypothesis. Conceptually, the test asks whether the observed treatment effect is sufficiently incompatible with the null hypothesis in the prespecified favorable direction.
Intention-to-treat principle
The analysis text identifies intention-to-treat analysis as an analysis concept. The registry-reported FAS definition similarly anchors the efficacy population to randomized treatment assignment. This preserves the principal advantage of randomization: baseline treatment groups are defined before subsequent outcomes and treatment changes occur.
8. Primary Result: Progression-Free Survival
The primary endpoint analysis compares centrally assessed PFS between Lutathera plus octreotide LAR 30 mg and octreotide LAR 60 mg in the Full Analysis Set.
Hazard ratio for progression or death
95% CI: 0.182–0.418 · P < 0.0001
Stratified one-sided P-value · Superiority hypothesis
| Primary endpoint | Lutathera + Octreotide LAR 30 mg | Octreotide LAR 60 mg | Effect estimate |
|---|---|---|---|
| Progression Free Survival (PFS) Per Central Assessment | Investigational arm | Control arm | HR 0.276 (95% CI 0.182–0.418); P < 0.0001 |
The estimated hazard ratio of 0.276 indicates that, under the time-to-event analysis, the estimated instantaneous rate of first-line progression or death in the Lutathera plus octreotide LAR 30 mg group was about 27.6% of the corresponding rate in the octreotide LAR 60 mg group. Expressed as a relative reduction in estimated hazard, this is approximately 72.4%.
The HR does not mean that 72.4% of participants were protected from progression, that 72.4% of participants benefited, or that each participant experienced exactly the same proportional reduction in risk. It is a population-level relative time-to-event measure.
The two-sided 95% confidence interval of 0.182–0.418 describes uncertainty around the estimated hazard ratio under the statistical framework. It does not give the range of individual patient outcomes or guarantee that the true hazard ratio is somewhere in that interval for every population or future study.
The P-value of <0.0001 addresses evidence against the null hypothesis under the specified one-sided, stratified testing framework. It does not measure the size of the treatment effect. Effect size is communicated by the hazard ratio and its confidence interval.
Because this is a time-to-event analysis, interpretation also depends on censoring and on the assumptions underlying the hazard-ratio framework. In particular, a single HR is most naturally interpreted when the relative hazards are reasonably represented by the model over the relevant follow-up. The ClinicalTrials.gov record does not provide a separate assessment of the proportional-hazards assumption.
What is and is not established by the reported PFS result
The posted analysis establishes that the registry reports a strong statistically significant superiority comparison for the primary PFS endpoint under its stated testing framework. It does not, by itself, provide a median PFS, a specific landmark PFS rate, or an absolute difference in months. Those quantities are not included in the ClinicalTrials.gov record and therefore are not reproduced here.
9. Secondary Endpoint Results
Overall Response Rate
Overall Response Rate per central assessment was a key secondary endpoint measured up to approximately 42 months. The posted comparison uses a stratified one-sided P-value and reports a stratified odds ratio.
Odds ratio for objective response
95% CI: 3.32–18.40 · P < 0.0001
Stratified one-sided P-value · Superiority hypothesis
An odds ratio of 7.81 means that the estimated odds of the response outcome were 7.81 times as high in the Lutathera plus octreotide LAR 30 mg group as in the octreotide LAR 60 mg group under the reported analysis.
An odds ratio is not the same as a risk ratio or a percentage-point difference in response probability. Without the underlying response counts or response percentages, the odds ratio cannot be converted into an absolute response-rate difference from the ClinicalTrials.gov record alone.
The 95% confidence interval of 3.32–18.40 indicates substantial uncertainty in the magnitude of the estimated odds ratio even though the entire interval is above 1. The P-value of <0.0001 addresses statistical evidence under the stated one-sided framework; it does not quantify how clinically large the response difference is.
Time to Deterioration
The key secondary endpoint Time to Deterioration (TTD) Global Health Status, Diarrhea, Fatigue, Pain (EORTC QLQ-C30) was analyzed as a time-to-event endpoint up to approximately 42 months.
Hazard ratio for time to deterioration
95% CI: 0.570–1.283 · P = 0.2222
Global Health Status stratified one-sided P-value
The estimated hazard ratio of 0.856 is below 1, corresponding to an estimated hazard approximately 85.6% of the comparator hazard under the reported analysis. In relative terms, the point estimate corresponds to approximately a 14.4% lower estimated hazard.
However, the two-sided 95% confidence interval of 0.570–1.283 spans 1. This means the data are compatible with a range of relative effects that includes both a lower and a higher hazard under the model.
The stratified one-sided P-value of 0.2222 does not provide evidence against the null hypothesis under the reported superiority-testing framework. Importantly, a nonsignificant P-value does not prove that the treatment effects are identical. It indicates that the observed data did not meet the specified statistical criterion for superiority for this endpoint.
The endpoint also illustrates why the HR and confidence interval should be considered together: the point estimate alone might suggest a lower deterioration hazard, whereas the interval indicates considerable uncertainty about the direction and magnitude of the underlying effect.
10. Results Summary
| Endpoint | Type | Effect measure | Estimate | 95% CI | P-value |
|---|---|---|---|---|---|
| Progression Free Survival (PFS) Per Central Assessment | Time-to-event | Hazard ratio | 0.276 | 0.182–0.418 | <0.0001 |
| Overall Response Rate (ORR) Per Central Assessment (Key Secondary) | Binary | Odds ratio | 7.81 | 3.32–18.40 | <0.0001 |
| Time to Deterioration (TTD) Global Health Status, Diarrhea, Fatigue, Pain (EORTC QLQ-C30) (Key Secondary) | Time-to-event | Hazard ratio | 0.856 | 0.570–1.283 | 0.2222 |
The three posted statistical analyses therefore illustrate two different patterns. The primary PFS analysis reports a hazard ratio substantially below 1 with a confidence interval entirely below 1 and a stratified one-sided P-value below 0.0001. The ORR analysis reports an odds ratio substantially above 1 with a confidence interval entirely above 1 and a P-value below 0.0001. The TTD analysis has a point estimate below 1, but its confidence interval includes 1 and its reported P-value is 0.2222.
11. Safety
The ClinicalTrials.gov record reports serious adverse events by treatment pathway. These figures should be interpreted as counts of affected participants over the corresponding number at risk; they are not the same as the efficacy analysis population and should not be substituted for a complete adverse-event table.
| Arm / treatment pathway | Serious adverse events affected | At risk | Affected / at risk |
|---|---|---|---|
| Lutathera® Plus Octreotide LAR 30 mg | 30 | 147 | 30/147 |
| Octreotide LAR 60 mg (Control Arm) | 15 | 73 | 15/73 |
| Crossover Lutathera | 2 | 29 | 2/29 |
| Re-Treatment Lutathera | 0 | 8 | 0/8 |
The reported safety data demonstrate why the crossover structure matters for safety interpretation as well as efficacy. The registry distinguishes the randomized treatment arms from crossover and re-treatment Lutathera pathways. These groups should therefore not be pooled into a single randomized comparison without an explicitly defined analysis population and exposure framework.
12. Crossover and Its Statistical Importance
NETTER-2 is registered with a crossover design model. This is a particularly important statistical feature because treatment assignment at randomization and treatment actually received later in the study can diverge.
Randomized comparison
The primary efficacy analysis remains defined by the randomized treatment assignment through the Full Analysis Set.
Subsequent treatment
The registry separately identifies a crossover Lutathera pathway and a re-treatment Lutathera pathway.
Why ITT matters
Analyzing participants according to randomized assignment preserves the causal comparison created by randomization, even when subsequent treatment changes occur.
What the posted HR does not do
The reported primary HR is not identified as a crossover-adjusted treatment effect. No crossover-adjusted HR is reported in the ClinicalTrials.gov record used for this page.
A crossover can complicate interpretation because the outcome observed after randomization can reflect more than the initial treatment assignment. For an overall survival endpoint, this can dilute differences between randomized groups if participants in the control pathway later receive the investigational therapy. For PFS, the precise effect depends on when crossover occurs relative to progression and how the protocol defines treatment exposure and endpoint ascertainment.
13. Statistical Methods Explained
Why was a log-rank test used for the primary endpoint?
PFS is a time-to-event endpoint. The log-rank test compares the timing of events across treatment groups while accounting for participants who are censored before experiencing progression or death. A simple comparison of proportions would discard important information about when events occurred.
What does an HR of 0.276 mean?
An HR of 0.276 means that the estimated instantaneous rate of progression or death was about 27.6% of the comparator rate under the reported time-to-event framework. It does not mean that 27.6% of patients experienced an event or that every patient's individual risk was reduced by the same proportion.
Why is the confidence interval important?
The 95% CI of 0.182–0.418 shows the uncertainty surrounding the primary HR estimate. The interval is about the statistical estimate, not about the range of individual patient responses. A point estimate should always be interpreted together with its interval because two estimates of the same size can have very different levels of precision.
Why is the primary P-value one-sided?
The registry explicitly identifies the primary analysis as using a stratified one-sided P-value under a superiority hypothesis. A one-sided test concentrates its rejection region in the prespecified direction. Its validity depends on that directional hypothesis being specified before the results are examined.
Why does the ORR odds ratio differ conceptually from the PFS hazard ratio?
ORR is a binary endpoint: participants either meet the prespecified response definition or they do not. Its odds ratio compares the odds of response between groups. PFS is a time-to-event endpoint, so its hazard ratio compares event rates over time. The two effect measures therefore cannot be interpreted as interchangeable measures of treatment benefit.
What does the TTD HR of 0.856 tell us?
The point estimate corresponds to an estimated deterioration hazard about 85.6% of the comparator hazard. But the 95% CI of 0.570–1.283 includes 1, and the reported P-value is 0.2222. The appropriate interpretation is therefore one of statistical uncertainty rather than treating the point estimate as definitive evidence of a treatment effect.
Why does the Full Analysis Set matter?
The FAS is defined as all participants to whom study treatment was assigned by randomization. This keeps the primary efficacy comparison tied to the randomized groups. That distinction is especially important in a crossover design because later treatment received can differ from the initial assignment.
14. Understanding the Primary PFS Result
The primary HR of 0.276 is a relative measure. It summarizes the difference between the treatment groups in the estimated hazard of progression or death under the analysis model.
The ClinicalTrials.gov record does not provide a median PFS, landmark PFS percentages, or an absolute difference in PFS time. The HR therefore cannot be translated from the ClinicalTrials.gov record into a statement such as "X additional months of PFS."
The reported stratified one-sided P-value is <0.0001. This measures evidence against the specified null hypothesis; it does not tell us that the effect is clinically important, nor does its magnitude directly measure the treatment effect.
The 95% CI of 0.182–0.418 provides the uncertainty interval for the estimated HR. The interval remains below 1 throughout, but it spans a substantial range of possible relative hazard reductions.
15. Hazard Ratios, Censoring, and Time-to-Event Endpoints
Time-to-event endpoints require different statistical thinking from ordinary continuous or binary outcomes. The key information is not merely whether a participant experienced progression or death, but when the event occurred and whether the participant remained event-free through the available observation period.
The survival function describes the probability that the event time T exceeds a particular time t. In a PFS analysis, the event is defined by the protocol as progression or death.
Censoring allows participants without an observed event by their last evaluable follow-up to contribute partial information rather than being discarded entirely. Correct handling of censoring is therefore fundamental to Kaplan-Meier and related time-to-event analyses.
The ClinicalTrials.gov record identifies the log-rank test and hazard ratio but do not provide a separate Kaplan-Meier estimate table or a formal assessment of proportional hazards.
16. Odds Ratios and Response Endpoints
The ORR analysis provides a useful contrast with the primary PFS analysis. ORR is binary, whereas PFS is time-to-event.
| Feature | PFS | ORR |
|---|---|---|
| Endpoint type | Time-to-event | Binary |
| Reported effect measure | Hazard ratio | Odds ratio |
| Reported estimate | 0.276 | 7.81 |
| 95% CI | 0.182–0.418 | 3.32–18.40 |
| P-value | <0.0001 | <0.0001 |
| Interpretive focus | Timing of progression or death | Odds of meeting the response definition |
The two endpoints answer related but different questions. A response endpoint focuses on whether a prespecified tumor-response state occurs, whereas PFS incorporates the timing of progression and death. An odds ratio therefore cannot be read as though it were a hazard ratio.
17. One-Sided Testing and Confidence Intervals
The registry reports one-sided P-values for the primary PFS analysis and for the key secondary analyses. At the same time, the posted effect estimates are accompanied by two-sided 95% confidence intervals.
P-value
Addresses statistical evidence against the specified null hypothesis under the stated one-sided testing framework.
Confidence interval
Describes uncertainty around the estimated effect and shows a range of effect sizes compatible with the statistical model and data.
Effect estimate
Communicates the magnitude and direction of the observed treatment comparison.
Not interchangeable
A very small P-value does not imply a large effect, while a confidence interval is not itself a probability statement about the true effect.
This distinction is particularly visible in the TTD analysis. Its HR is below 1, but the confidence interval includes 1 and the P-value is 0.2222. Looking only at the point estimate would omit the uncertainty that is central to interpretation.
18. Multiplicity and Endpoint Interpretation
The ClinicalTrials.gov record identifies one primary endpoint and two posted key secondary endpoint analyses. The primary hypothesis is a superiority hypothesis with a stratified one-sided P-value.
| Endpoint | Role | Analysis | Effect measure |
|---|---|---|---|
| Progression Free Survival (PFS) Per Central Assessment | Primary | Stratified log-rank | Hazard ratio |
| Overall Response Rate (ORR) Per Central Assessment | Key secondary | Stratified one-sided P-value | Stratified odds ratio |
| Time to Deterioration (TTD) Global Health Status, Diarrhea, Fatigue, Pain (EORTC QLQ-C30) | Key secondary | Log-rank | Hazard ratio |
When a trial evaluates multiple endpoints, the statistical interpretation depends on the prespecified hierarchy and error-control strategy. The ClinicalTrials.gov record identifies the endpoint roles and reported testing approaches, but do not provide a complete multiplicity-adjustment procedure. This page therefore does not infer a familywise error allocation or claim that every secondary P-value has the same confirmatory status as the primary analysis.
19. Stratification and Randomization
Randomization is the foundation of the treatment comparison because treatment assignment is established before subsequent outcomes occur. The primary analysis also identifies stratified analysis as part of its statistical framework.
Stratification can improve alignment between the randomization structure and the statistical comparison. The exact stratification factors used for NETTER-2 are not included in the ClinicalTrials.gov record, so none are specified here.
The absence of reported stratification factors is important. It would be inappropriate to infer them from typical neuroendocrine-tumor trial practice or from other publications. The registry's registry-reported analysis text supports the existence of stratified analysis, but not the identities of the strata.
20. Planned Analysis vs Posted Results
NETTER-2 is a results-posted trial, so a separate "Planned Analysis" section is not required in place of results. The registry provides formal statistical analyses for the primary PFS endpoint and two key secondary endpoints.
| Analysis status | What the ClinicalTrials.gov record shows |
|---|---|
| Results posted | Yes |
| Outcome measures posted | 8 |
| Statistical analyses posted | 3 |
| Primary-endpoint analyses posted | 1 |
| Primary analyses with estimate + CI | 1 |
| Analysis methods | Log-rank test; other methods as reported in the registry |
| Effect measures | Hazard ratio; odds ratio |
| Hypothesis type | Superiority |
21. Important Limitations and Interpretation Issues
- Limited numerical detail: the ClinicalTrials.gov record provides effect estimates and confidence intervals for three analyses but do not provide median PFS, median TTD, absolute response percentages, event counts for PFS, or landmark survival estimates.
- Crossover design: later treatment pathways can complicate interpretation of treatment effects, particularly for outcomes influenced by subsequent therapy.
- Primary analysis population: the reported PFS analysis uses the Full Analysis Set defined by randomized treatment assignment. Safety pathways have separate denominators and include crossover and re-treatment groups.
- Hazard-ratio interpretation: a single HR is a relative time-to-event measure and should not be interpreted as an absolute risk reduction or a guarantee of proportional benefit for every participant.
- Proportional-hazards assumption: the ClinicalTrials.gov record does not provide a formal diagnostic or test for the proportional-hazards assumption.
- Multiplicity: the ClinicalTrials.gov record identifies primary and key secondary endpoints but do not provide a complete multiplicity-adjustment strategy. Secondary P-values should therefore not automatically be treated as having identical confirmatory status.
- One-sided testing: the primary and key secondary analyses use or report one-sided testing, so interpretation depends on the prespecified direction of the superiority hypothesis.
- Safety denominators: serious adverse-event counts are reported for randomized and subsequent-treatment pathways with different denominators. They should not be pooled without a defined exposure-based analysis.
- Missing data: the ClinicalTrials.gov record does not specify a detailed missing-data or imputation strategy for the endpoints. No imputation method is inferred here.
- Stratification details: the analysis identifies stratified methods, but the ClinicalTrials.gov record does not specify the actual stratification factors.
22. Why This Trial Matters Statistically
NETTER-2 is a useful teaching example because several core clinical-trial concepts appear in a single analysis framework: randomization, crossover, time-to-event endpoints, stratified log-rank testing, hazard ratios, one-sided superiority testing, confidence intervals, binary response analysis, odds ratios, and distinct efficacy and safety analysis pathways.
| Concept | How it appears in NETTER-2 |
|---|---|
| Randomization | The study is registered as randomized, and the primary FAS comprises participants assigned treatment by randomization. |
| Crossover design | The registry identifies the design model as crossover and separately reports a crossover Lutathera pathway. |
| Time-to-event endpoint | The primary endpoint is centrally assessed PFS from randomization to first-line progression or death. |
| Log-rank test | Used for the primary PFS analysis and reported for the TTD key secondary endpoint. |
| Hazard ratio | Used to quantify the relative treatment effect for PFS and TTD. |
| Confidence interval | 95% two-sided intervals accompany the reported HR and OR estimates. |
| One-sided testing | The primary P-value is explicitly described as a stratified one-sided P-value. |
| Odds ratio | The key secondary ORR analysis reports a stratified odds ratio of 7.81. |
| Stratified analysis | The primary PFS analysis explicitly identifies stratified analysis. |
| Intention-to-treat principle | Intention-to-treat analysis is identified among the concepts in the primary analysis text. |
| Safety analysis | Serious adverse events are reported separately for randomized, crossover, and re-treatment pathways. |
23. Statistical Concepts in This Trial
Learn more about the methods used in this trial:
24. Related Statistical Calculators
25. Sources
- ClinicalTrials.gov: NETTER-2, NCT03972488. Official trial registry record and source for the trial data summarized on this page.
- PubMed: PubMed record: PMID 38851203.
- PubMed: PubMed record: PMID 42604001.
- PubMed: PubMed record: PMID 40473457.
- PubMed: PubMed record: PMID 39570708.
- PubMed: PubMed record: PMID 33973550.
Continue through the Clinical Biostats statistical tutorials
Explore the statistical concepts that underlie randomized clinical-trial analysis, including time-to-event endpoints, hazard ratios, confidence intervals, odds ratios, and hypothesis testing.
26. Record Summary
NETTER-2 provides a clear example of how a randomized phase 3 trial can combine a time-to-event primary endpoint with key secondary efficacy measures and a crossover treatment structure. The primary PFS analysis reports a hazard ratio of 0.276 with a two-sided 95% CI of 0.182–0.418 and a stratified one-sided P-value of <0.0001. The key secondary ORR analysis reports a stratified odds ratio of 7.81 with a 95% CI of 3.32–18.40 and P < 0.0001, while the key secondary TTD analysis reports an HR of 0.856 with a 95% CI of 0.570–1.283 and P = 0.2222.
The statistical lesson is broader than any individual number. The PFS result must be interpreted through its randomized Full Analysis Set, stratified log-rank framework, one-sided superiority test, hazard-ratio effect measure, confidence interval, and time-to-event structure. The ORR result requires a different interpretation because its odds ratio describes a binary outcome. The TTD result demonstrates why point estimates, confidence intervals, and P-values should be considered together rather than separately.