This page separates reported trial results from statistical interpretation. Numerical results are restricted to the ClinicalTrials.gov record. ClinicalTrials.gov provides the official trial registry record.
1. Trial at a Glance
LOTUS was a randomized, double-blind, parallel phase 2 trial evaluating the addition of ipatasertib to paclitaxel in participants with breast cancer that had spread beyond the initial site and did not have certain hormonal receptors. The registry reports 124 participants, two arms, two primary endpoints, and formal statistical analyses based on the intention-to-treat population.
| Feature | LOTUS |
|---|---|
| Trial name | LOTUS |
| Phase | Phase 2 |
| Condition | Breast Neoplasms |
| Design | Randomized, double-blind, parallel |
| Primary purpose | Treatment |
| Enrollment | 124.0 |
| Primary endpoints | Progression Free Survival (PFS); PFS in Participants With Phosphatase and Tensin Homolog (PTEN)-Low Tumors |
| Primary endpoint types registered | Binary; Time-to-event |
| Results posted | Yes |
| Outcome measures posted | 20 |
| Statistical analyses posted | 6 |
| Lead sponsor | Genentech, Inc. |
| Sponsor type | Industry |
| ClinicalTrials.gov | NCT02162719 |
2. Clinical Question
The central question was whether adding ipatasertib to paclitaxel improved progression-free survival compared with placebo plus paclitaxel. The trial also prespecified PFS in participants with PTEN-low tumors as a primary endpoint.
Population
Participants with breast cancer that had spread beyond the initial site and did not have certain hormonal receptors, as described by the registry brief title.
Intervention
Ipatasertib plus paclitaxel.
Comparator
Placebo plus paclitaxel.
Primary question
Does adding ipatasertib to paclitaxel improve PFS, including PFS among participants with PTEN-low tumors?
3. Trial Design
Ipatasertib and Paclitaxel
- Ipatasertib
- Paclitaxel
Placebo and Paclitaxel
- Placebo
- Paclitaxel
The registry identifies the allocation as randomized, the design model as parallel, and masking as double. These design features matter statistically because randomization establishes the framework for a treatment comparison, while blinding can reduce the influence of treatment knowledge on trial conduct and assessment.
4. Endpoints
| Endpoint | Registry time frame | Definition | Registered type |
|---|---|---|---|
| Progression Free Survival (PFS) | Baseline up to 30 days after the last dose of study drug administration (Clinical Cut Off Date: 07 June 2016) | PFS was defined as the time from randomization to the first occurrence of disease progression, as determined by investigator review of tumor assessments by RECIST, v1.1 or death on study (≤30 days after the last dose of study treatment regimen) from any cause, whichever occurred first. | Time-to-event |
| PFS in Participants With Phosphatase and Tensin Homolog (PTEN)-Low Tumors | Baseline up to 30 days after the last dose of study drug administration (Clinical Cut Off Date: 07 June 2016) | PFS was defined as the time from randomization to the first occurrence of disease progression, as determined by investigator review of tumor assessments by RECIST, v1.1 or death on study (≤30 days after the last dose of study treatment regimen) from any cause, whichever occurred first. | Binary in the registry endpoint classification; analyzed as time-to-event in the posted statistical analysis |
The endpoint classification deserves attention. The registry's primary-endpoint type field lists Binary; Time-to-event, but the registry-reported formal analysis for both primary endpoints uses a log-rank test and a hazard ratio. The statistical analysis therefore treats both reported primary comparisons as time-to-event analyses.
5. Statistical Methodology
Intention-to-treat analysis
The primary PFS analysis used the intent-to-treat (ITT) population. The registry defines this population as all randomized participants allocated to the treatment arm to which they were randomized. For the PTEN-low endpoint, participants with PTEN-low tumors were evaluated within the ITT framework.
The key statistical advantage is preservation of the treatment comparison created by randomization. Treatment discontinuation or other post-randomization events do not automatically move a participant into the other randomized group.
Log-rank test
The registry reports Log Rank as the method for all six posted statistical analyses. The log-rank test is designed for comparing time-to-event distributions between groups while accounting for different follow-up times and right censoring.
Hazard ratio
The primary effect measure was a Hazard Ratio (Stratified Analysis). The hazard ratio summarizes the relative event rate between the randomized groups within the model used for the analysis. An HR below 1 indicates a lower estimated instantaneous event rate in the ipatasertib-plus-paclitaxel group than in the placebo-plus-paclitaxel group.
The hazard ratio is not an absolute risk difference, a probability of response, a median survival difference, or a statement that every participant experiences the same proportional change in risk.
Kaplan-Meier framework
The registry's endpoints are defined as times from randomization to progression or death, and the reported method is the log-rank test. For this type of endpoint, Kaplan-Meier estimation is the standard descriptive framework for representing the event-time distribution while retaining information from participants who are censored before experiencing the event.
Confidence intervals
Each posted analysis supplies a two-sided confidence interval around its hazard ratio. The interval describes statistical uncertainty around the estimated relative treatment effect under the specified analysis framework. It does not describe the range of effects that must occur for individual participants.
6. Primary Results: Progression Free Survival
The first primary endpoint was PFS from baseline through 30 days after the last dose of study drug administration, with a clinical cutoff date of 07 June 2016. The analysis compared ipatasertib and paclitaxel with placebo and paclitaxel in the ITT population.
Hazard ratio for progression or death
90% CI: 0.40–0.91 · P = 0.0372
Method: stratified log-rank analysis; superiority hypothesis
| Primary PFS analysis | Reported result |
|---|---|
| Comparison | Ipatasertib and Paclitaxel vs Placebo and Paclitaxel |
| Analysis population | Intent-to-treat population |
| Method | Log Rank |
| Effect measure | Hazard Ratio (Stratified Analysis) |
| Hazard ratio | 0.60 |
| Confidence interval | 90% CI 0.40–0.91, two-sided |
| P-value | 0.0372 |
| Hypothesis type | Superiority |
The reported HR of 0.60 means that, under the time-to-event analysis used by the registry, the estimated instantaneous rate of progression or death was 0.60 times that in the comparator group. Expressed as a simple relative interpretation, 0.60 corresponds to a 40% lower estimated hazard in the ipatasertib-plus-paclitaxel group.
That does not mean that 40% of participants avoided progression, that 40% lived longer, or that every participant experienced exactly a 40% reduction in risk. A hazard ratio is a relative time-to-event measure, not an absolute probability.
The 90% CI of 0.40–0.91 communicates uncertainty around the estimated hazard ratio. It is substantially narrower than an interval extending across very different directions of effect, but it still permits a range of plausible relative treatment effects. The interval should be read together with the point estimate rather than treating 0.60 as an exact treatment effect.
The P-value of 0.0372 is a measure of compatibility with the statistical null hypothesis under the analysis framework. It is not a measure of effect size, clinical importance, or the probability that the treatment works. The size of the treatment effect is conveyed by the hazard ratio and its confidence interval.
The registry reports this as a stratified analysis, but the ClinicalTrials.gov record does not identify the stratification variables. The page therefore does not infer them. As with any Cox-type hazard-ratio interpretation, proportional hazards is a relevant modeling consideration; the ClinicalTrials.gov record does not report an assessment of that assumption.
7. Primary Results: PFS in PTEN-Low Tumors
The second primary endpoint evaluated PFS among participants with Phosphatase and Tensin Homolog (PTEN)-Low Tumors. The time frame was the same baseline-to-30-days-after-last-dose period, with the clinical cutoff date of 07 June 2016.
Hazard ratio for progression or death in PTEN-low tumors
90% CI: 0.30–1.16 · P = 0.1753
Method: stratified log-rank analysis; superiority hypothesis
| Primary PTEN-low PFS analysis | Reported result |
|---|---|
| Comparison | Ipatasertib and Paclitaxel vs Placebo and Paclitaxel |
| Analysis population | ITT population; participants with PTEN-low tumors evaluated for this endpoint |
| Method | Log Rank |
| Effect measure | Hazard Ratio (Stratified Analysis) |
| Hazard ratio | 0.59 |
| Confidence interval | 90% CI 0.30–1.16, two-sided |
| P-value | 0.1753 |
| Hypothesis type | Superiority |
The reported HR of 0.59 corresponds to an estimated instantaneous rate of progression or death approximately 0.59 times that of the comparator group under the reported analysis. As a simple relative interpretation, this is a 41% lower estimated hazard.
The confidence interval is 0.30–1.16. Unlike the primary PFS estimate of 0.60, this interval extends above 1.00. That means the reported estimate is compatible with a range that includes both a lower hazard and a higher hazard under the confidence-interval framework.
The P-value of 0.1753 does not establish a statistically significant superiority finding under a conventional 0.05 threshold. More importantly, the P-value should not be used to conclude that there is no treatment effect. The wide interval shows that the estimate is uncertain, and the point estimate alone should not be treated as definitive evidence of the magnitude of an effect in this subgroup.
This endpoint also concerns a selected tumor characteristic rather than the full randomized population. Subset analyses generally contain less information than the overall ITT comparison, and the ClinicalTrials.gov record does not provide the number of PTEN-low participants contributing to this analysis. No conclusion about treatment-effect heterogeneity between PTEN-low and other tumors should be drawn from these data alone.
8. Secondary Endpoint Results
The registry also reports four secondary time-to-event analyses: PFS in participants with PIK3CA/AKT1/PTEN-altered tumors, overall survival, OS in participants with PTEN-low tumors, and OS in participants with PIK3CA/AKT1/PTEN-altered tumors.
| Secondary endpoint | Effect | Confidence interval | P-value | Analysis |
|---|---|---|---|---|
| PFS in Participants With PIK3CA/AKT1/PTEN-altered Tumors | HR 0.76 | 90% CI 0.46–1.27 | 0.3636 | Unstratified log-rank analysis |
| Overall Survival (OS) | HR 0.80 | 95% CI 0.50–1.28 | 0.3607 | Stratified log-rank analysis |
| OS in Participants With PTEN-Low Tumors | HR 0.73 | 95% CI 0.32–1.65 | 0.4422 | Stratified log-rank analysis |
| OS in Participants With PIK3CA/AKT1/PTEN-altered Tumors | HR 1.13 | 95% CI 0.52–2.47 | 0.7599 | Unstratified log-rank analysis |
PFS in PIK3CA/AKT1/PTEN-altered tumors
The secondary PFS analysis among participants with PIK3CA/AKT1/PTEN-altered tumors reported an HR of 0.76, with a two-sided 90% CI of 0.46–1.27 and P = 0.3636. The analysis used the unstratified form of the hazard-ratio analysis.
Overall survival
For OS, the reported analysis used a clinical cutoff date of 31 August 2019. The stratified hazard ratio was 0.80, with a two-sided 95% CI of 0.50–1.28 and P = 0.3607.
OS in PTEN-low tumors
Among participants with PTEN-low tumors, the reported OS hazard ratio was 0.73, with a two-sided 95% CI of 0.32–1.65 and P = 0.4422. The analysis was stratified.
OS in PIK3CA/AKT1/PTEN-altered tumors
The OS analysis in participants with PIK3CA/AKT1/PTEN-altered tumors reported an HR of 1.13, with a two-sided 95% CI of 0.52–2.47 and P = 0.7599. This analysis was unstratified.
9. Comparing the Primary and Secondary Hazard Ratios
Looking across the six formal analyses shows why the hazard ratio should always be interpreted together with its confidence interval and analysis population.
| Endpoint | HR | CI | P-value | Stratification |
|---|---|---|---|---|
| Primary PFS | 0.60 | 90% CI 0.40–0.91 | 0.0372 | Stratified |
| Primary PFS, PTEN-low | 0.59 | 90% CI 0.30–1.16 | 0.1753 | Stratified |
| Secondary PFS, PIK3CA/AKT1/PTEN-altered | 0.76 | 90% CI 0.46–1.27 | 0.3636 | Unstratified |
| Secondary OS | 0.80 | 95% CI 0.50–1.28 | 0.3607 | Stratified |
| Secondary OS, PTEN-low | 0.73 | 95% CI 0.32–1.65 | 0.4422 | Stratified |
| Secondary OS, PIK3CA/AKT1/PTEN-altered | 1.13 | 95% CI 0.52–2.47 | 0.7599 | Unstratified |
The point estimates range from 0.59 to 1.13, but the estimates cannot be compared as though they were measurements of the same quantity. They concern different endpoints, different analysis populations, different follow-up cutoffs, and, for some analyses, different stratification specifications.
The subgroup confidence intervals are also relatively broad. This is important statistically: a point estimate such as 0.73 does not necessarily provide strong evidence of a precise treatment effect when the confidence interval spans a wide range.
10. Safety Results
The ClinicalTrials.gov record reports serious adverse events by randomized arm as affected participants divided by participants at risk.
| Safety measure | Ipatasertib and Paclitaxel | Placebo and Paclitaxel |
|---|---|---|
| Serious adverse events, affected / at risk | 18 / 61 | 12 / 62 |
The ClinicalTrials.gov record does not provide a formal statistical comparison for serious adverse events. Accordingly, the counts are presented descriptively rather than assigned a P-value or relative effect estimate.
11. Statistical Methods Explained
Why was a log-rank test used?
PFS and OS are time-to-event endpoints. Some participants can remain event-free at the end of observation and therefore contribute censored follow-up rather than an observed event time. The log-rank test is designed for comparing the event-time distributions of two groups while accounting for this censoring structure.
What does an HR of 0.60 mean?
An HR of 0.60 means that the estimated instantaneous rate of the analyzed event is 0.60 times that of the comparator under the fitted time-to-event framework. A convenient relative description is a 40% lower estimated hazard. It does not mean a 40% absolute reduction in the probability of progression or death.
Why does the confidence interval matter?
The confidence interval communicates the precision of the estimated hazard ratio. For the primary PFS analysis, the reported 90% CI is 0.40–0.91. For the PTEN-low primary analysis, the 90% CI is 0.30–1.16. The second interval is wider and crosses 1.00, illustrating how a point estimate below 1 can coexist with substantial statistical uncertainty.
Why doesn't the P-value measure the size of the treatment effect?
The P-value addresses evidence against a null hypothesis under the statistical model; it is not a scale for measuring treatment benefit. An effect estimate such as HR 0.60 tells us the estimated relative magnitude, while its confidence interval tells us about uncertainty around that estimate. The P-value answers a different question.
Why is intention-to-treat important?
The ITT definition reported by the registry keeps randomized participants in the treatment group to which they were assigned. This preserves the comparison generated by randomization. It also means the efficacy analysis is not restricted only to participants who completed treatment exactly as planned.
Why are the PTEN-low and PIK3CA/AKT1/PTEN-altered results different from the overall PFS result?
They answer different questions in different analysis populations. A subgroup estimate can differ from the overall estimate simply because the underlying participants and number of observed events differ. A different point estimate does not by itself establish a biologically or statistically different treatment effect between subgroups; formal interaction testing would be needed for that question, and no such test is reported here.
What is the importance of stratified versus unstratified analysis?
The registry labels four of the analyses as stratified and two as unstratified. Stratification can account for prespecified factors in a time-to-event comparison, whereas an unstratified analysis does not incorporate such factors in the same way. Because the ClinicalTrials.gov record does not identify the stratification variables, no specific factors are attributed to the stratified analyses on this page.
12. Confidence Intervals in Context
The LOTUS registry results provide an instructive contrast between point estimates and precision.
Primary PFS
HR 0.60 with a two-sided 90% CI of 0.40–0.91. The interval remains below 1.00.
PTEN-low PFS
HR 0.59 with a two-sided 90% CI of 0.30–1.16. The interval crosses 1.00.
Overall survival
HR 0.80 with a two-sided 95% CI of 0.50–1.28.
Altered-tumor OS
HR 1.13 with a two-sided 95% CI of 0.52–2.47, illustrating considerable uncertainty around the point estimate.
A confidence interval that crosses 1.00 does not prove equivalence between treatment groups. It means that the interval includes the null hazard ratio as well as other possible relative effects. Conversely, an interval below 1.00 should not be interpreted as proof that the treatment has an important clinical effect without considering the magnitude, endpoint, population, and broader evidence.
13. Time-to-Event Analysis: What Is Being Estimated?
The registry definition of PFS combines two possible events: investigator-determined disease progression under RECIST v1.1 and death on study within the specified post-treatment window. Whichever qualifying event occurs first determines the PFS event time.
Participants who do not experience a qualifying event during observable follow-up may be censored according to the applicable analysis rules. The ClinicalTrials.gov record does not specify the complete censoring algorithm beyond the registered endpoint definition.
This structure explains why PFS cannot be reduced to a simple proportion of patients who progressed. It is a time-to-event endpoint: both whether an event occurred and when it occurred contribute information to the analysis.
14. Analysis Cutoffs and Follow-Up
Trial start
The registry profile lists 2014-08-19 as the study start date.
Primary PFS cutoff
The primary PFS endpoints use a clinical cutoff date of 07 June 2016 and define the observation window as baseline through 30 days after the last dose of study drug administration.
Primary completion
the ClinicalTrials.gov record lists 2016-06-07 as the primary completion date.
Overall-survival cutoff
The registry-reported OS analyses use a later clinical cutoff date of 31 August 2019.
The different cutoff dates are statistically important. The primary PFS analyses and the later OS analyses do not represent the same observation period. Results from different cutoffs should therefore not be combined as though they came from a single synchronized analysis.
15. Multiplicity, Subgroups, and Multiple Analyses
The registry supplies two primary endpoints and four secondary statistical analyses, for six posted formal statistical analyses in total. The two primary endpoints concern overall PFS and PFS among participants with PTEN-low tumors.
| Analysis family | Number in the ClinicalTrials.gov record | Statistical issue |
|---|---|---|
| Primary endpoints | 2 | Multiple primary questions are evaluated |
| Secondary analyses | 4 | Additional treatment-effect estimates and P-values |
| Total statistical analyses posted | 6 | Results should be interpreted according to their prespecified roles |
The ClinicalTrials.gov record does not state an alpha-allocation strategy, multiplicity-adjustment method, hierarchical testing procedure, or interim-analysis rule. Therefore, the P-values are reported exactly as provided, without claiming that a particular familywise error rate was controlled.
Subgroup analyses require additional caution. PTEN-low and PIK3CA/AKT1/PTEN-altered tumor analyses concern selected populations. Their estimates can be more variable than the overall randomized comparison, and a difference between subgroup estimates is not itself evidence of treatment-effect modification.
16. What the Hazard Ratio Does — and Does Not — Mean
The primary PFS hazard ratio of 0.60 indicates a lower estimated instantaneous rate of progression or death in the ipatasertib-plus-paclitaxel group under the reported time-to-event analysis. The corresponding two-sided 90% confidence interval was 0.40–0.91.
This does not mean that the probability of progression or death was reduced by exactly 40%, nor does it mean that every participant experienced the same proportional effect.
The P-value of 0.0372 indicates the result's statistical compatibility with the null hypothesis under the reported testing framework. It should not be interpreted as a 3.72% probability that the null hypothesis is true, nor as a 96.28% probability that ipatasertib is effective.
A hazard ratio is a relative measure. Absolute survival probabilities, median event times, and event counts can provide additional clinical context. Those quantities are not reported in the ClinicalTrials.gov record used for this page, so they are not reconstructed or inferred.
17. Limitations
- Limited registry detail: the ClinicalTrials.gov record does not include the full statistical analysis plan, sample-size assumptions, power calculation, randomization ratio, or detailed censoring rules.
- No median event times: median PFS and median OS are not reported and therefore are not reported.
- No event counts: the number of PFS or OS events contributing to each efficacy analysis is not reported.
- Subgroup uncertainty: the PTEN-low and PIK3CA/AKT1/PTEN-altered analyses concern selected tumor populations, and their sample sizes are not provided.
- Multiplicity: two primary endpoints and multiple secondary analyses are reported, but the ClinicalTrials.gov record does not specify how multiplicity was controlled.
- Stratification details: several analyses are labeled stratified, but the ClinicalTrials.gov record does not identify the stratification factors.
- Unstratified analyses: the PIK3CA/AKT1/PTEN-altered PFS and OS analyses are explicitly labeled unstratified, so their estimates should not be treated as interchangeable with the stratified estimates.
- Hazard-ratio assumptions: the registry does not provide an assessment of the proportional-hazards assumption. A single hazard ratio can be an incomplete summary if relative hazards vary substantially over time.
- Safety inference: serious adverse-event counts are provided, but no formal comparative safety analysis is reported in the ClinicalTrials.gov record used here.
- Registry endpoint classification: the second primary endpoint is classified as binary in the registry-reported endpoint metadata even though its formal statistical analysis is a time-to-event log-rank/hazard-ratio analysis.
18. Why This Trial Matters Statistically
LOTUS is a useful teaching case because the ClinicalTrials.gov record illustrate how a randomized clinical trial can generate several layers of statistical evidence from the same treatment comparison.
| Concept | How it appears in LOTUS |
|---|---|
| Randomization | Participants were randomly allocated to two treatment arms. |
| Double blinding | The trial is registered as double masked. |
| Parallel design | The two treatment groups were evaluated as parallel arms. |
| ITT analysis | Primary efficacy analyses used all randomized participants according to assigned treatment. |
| Time-to-event endpoint | PFS and OS are analyzed using event times from randomization. |
| Log-rank test | The registry reports Log Rank for all six formal statistical analyses. |
| Hazard ratio | Relative treatment effects are reported as hazard ratios. |
| Confidence intervals | Every formal analysis includes a reported two-sided confidence interval. |
| Subgroup analysis | PTEN-low and PIK3CA/AKT1/PTEN-altered tumor populations are separately analyzed. |
| Different cutoffs | Primary PFS uses 07 June 2016, whereas OS analyses use 31 August 2019. |
| Stratified analysis | Primary PFS, PTEN-low PFS, OS, and PTEN-low OS are labeled stratified. |
| Unstratified analysis | PIK3CA/AKT1/PTEN-altered PFS and OS are labeled unstratified. |
| Safety reporting | Serious adverse events are reported as affected participants over participants at risk by arm. |
19. A Practical Reading of the LOTUS Results
A statistically disciplined reading of the registry results proceeds in layers rather than beginning with the P-value.
Step 1 · Identify the endpoint
The primary PFS endpoint measures time from randomization to progression or death, not simply whether progression occurred.
Step 2 · Identify the population
The primary analysis uses the ITT population; the PTEN-low analysis evaluates the relevant tumor subset.
Step 3 · Read the effect estimate
The primary PFS HR is 0.60, indicating a lower estimated event hazard in the ipatasertib group.
Step 4 · Read the interval
The 90% CI of 0.40–0.91 shows uncertainty around the primary PFS estimate.
Step 5 · Read the P-value
P = 0.0372 provides evidence against the null under the reported testing framework; it does not measure effect size.
Step 6 · Check the design
Randomization, double blinding, ITT analysis, stratification, subgroup selection, and multiple endpoints all affect interpretation.
This sequence prevents a common statistical error: treating a P-value as the conclusion and the effect estimate as supporting decoration. In a time-to-event trial, the hazard ratio, confidence interval, endpoint definition, analysis population, and censoring framework all contribute to what the result actually says.
20. Related Tutorials
Learn more about the methods used in this trial:
21. Related Statistical Calculators
22. Sources
- ClinicalTrials.gov: LOTUS, NCT02162719.
- Linked publication: PubMed record for PMID 28800861.
Continue through Clinical Biostats
Use the related statistical tutorials and calculators to explore randomized-trial design, survival analysis, hazard ratios, confidence intervals, and time-to-event methods in greater depth.
23. Record Summary
LOTUS provides a compact example of randomized time-to-event analysis. The trial enrolled 124.0 participants in a two-arm, randomized, double-blind, parallel phase 2 design. Its primary PFS analysis reported a hazard ratio of 0.60 with a two-sided 90% CI of 0.40–0.91 and P = 0.0372. The second primary analysis, restricted to participants with PTEN-low tumors, reported HR 0.59 with a two-sided 90% CI of 0.30–1.16 and P = 0.1753.
The secondary analyses illustrate why treatment effects must be interpreted in context. Reported hazard ratios were 0.76 for PFS in PIK3CA/AKT1/PTEN-altered tumors, 0.80 for OS, 0.73 for OS in PTEN-low tumors, and 1.13 for OS in PIK3CA/AKT1/PTEN-altered tumors. These estimates were accompanied by confidence intervals of varying width and by both stratified and unstratified analyses.
The most important statistical lesson is that a clinical-trial result is not a single number. The endpoint definition, randomized population, analysis method, effect measure, confidence interval, P-value, analysis cutoff, subgroup definition, and missing design details all determine how the reported evidence should be understood.