This page separates reported trial results from statistical interpretation. Numerical trial results, endpoint definitions, analysis methods, dates, and safety counts on this page are limited to the ClinicalTrials.gov record. 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
RESONATE was a randomized, open-label, parallel-group phase 3 study evaluating ibrutinib versus ofatumumab in patients with relapsed or refractory chronic lymphocytic leukemia or small lymphocytic lymphoma. The registry identifies progression-free survival assessed by an independent review committee as the primary endpoint and reports formal log-rank testing for that endpoint.
| Feature | RESONATE |
|---|---|
| Trial name | RESONATE |
| Phase | Phase 3 |
| Population | Patients with relapsed or refractory chronic lymphocytic leukemia or small lymphocytic lymphoma |
| Design | Randomized, parallel-group, unmasked |
| Allocation | Randomized |
| Primary purpose | Treatment |
| Enrollment | 391 |
| Interventions | Ofatumumab and ibrutinib |
| Primary endpoint | Progression-free survival by Independent Review Committee, limited to the time of primary analysis on 06 November 2013 |
| Results posted | Yes |
| ClinicalTrials.gov | NCT01578707 |
2. Clinical Question
The central statistical question was whether ibrutinib was superior to ofatumumab with respect to progression-free survival, as assessed by an independent review committee, in patients with relapsed or refractory chronic lymphocytic leukemia or small lymphocytic lymphoma.
Population
Patients with relapsed or refractory chronic lymphocytic leukemia or small lymphocytic lymphoma.
Intervention
Ibrutinib.
Comparator
Ofatumumab.
Primary question
Does ibrutinib produce superior progression-free survival compared with ofatumumab?
3. Trial Design
Ofatumumab
- Ofatumumab was one of the two study interventions.
- The registry identifies this group as Ofatumumab (Arm A) in the reported statistical analyses.
Ibrutinib
- Ibrutinib was one of the two study interventions.
- The registry identifies this group as Ibrutinib (Arm B) in the reported safety data.
4. Trial Timeline
Study start
The RESONATE study began in June 2012.
Primary completion
The registry lists November 2013 as the primary completion date.
Primary PFS analysis
The registered primary endpoint was limited to the time of the primary analysis on 06 November 2013. The analysis was conducted after approximately 117 PFS events, occurring about 18 months after the first subject.
Overall survival analysis
The posted OS analysis was conducted at the time of study closure, including up to 6 years of study follow-up.
5. Endpoints
| Endpoint | Registry definition / time frame | Type |
|---|---|---|
| Primary: PFS by Independent Review Committee | Progression-free survival assessed by independent review committee according to 2008 IWCLL guidelines. Limited to the time of primary analysis on 06 November 2013. Analysis was conducted after observing approximately 117 PFS events, which occurred about 18 months after the first subject. | Time-to-event |
| Secondary: Overall Response Rate | Overall Response Rate by Independent Review Committee. Time frame: about 18 months after the first subject was enrolled. | Binary |
| Secondary: Overall Survival | OS analysis conducted at the time of study closure, including up to 6 years of study follow-up. | Time-to-event |
The registry reports 6 outcome measures in total and 3 statistical analyses in the ClinicalTrials.gov record. One of those analyses corresponds to the primary endpoint and two correspond to secondary endpoints.
6. Statistical Methodology
Primary endpoint: progression-free survival
The primary endpoint is a time-to-event outcome. Each participant contributes follow-up until progression, death, or censoring according to the endpoint definition and analysis rules. The registry identifies the log-rank test as the formal method used for the primary PFS comparison.
Conceptually, the log-rank test compares the observed pattern of events between randomized groups over follow-up. The actual registry result is reported as a P-value rather than as a hazard ratio or confidence interval.
Secondary endpoint: overall response rate
Overall response rate is a binary endpoint: each participant is classified according to whether the prespecified response criterion was met. The registry reports the Fisher exact test for the comparison of the two treatment groups.
Fisher's exact test calculates the probability of tables at least as incompatible with the null hypothesis as the observed table, conditional on the table margins. It is particularly useful when cell counts may be small.
Secondary endpoint: overall survival
Overall survival is also a time-to-event endpoint. The registry identifies the log-rank test as the reported method for the OS comparison. The analysis was conducted at study closure and included up to 6 years of study follow-up.
7. Results
Primary Endpoint: PFS by Independent Review Committee
Reported log-rank result
Superiority hypothesis · Log-rank test
Primary analysis: 06 November 2013, after approximately 117 PFS events.
| Endpoint | Comparison | Method | Reported result |
|---|---|---|---|
| PFS by Independent Review Committee | Ofatumumab (Arm A) vs Ibrutinib (Arm B) | Log-rank test | P < 0.0001 |
The reported P < 0.0001 indicates that the observed time-to-event data were highly incompatible with the null hypothesis used for the superiority comparison under the reported log-rank analysis. The registry therefore reports a statistically significant primary comparison in favor of the alternative hypothesis.
The P-value does not quantify the size of the treatment effect. In particular, the ClinicalTrials.gov record does not report a hazard ratio, confidence interval, median PFS, or absolute PFS estimates, so the magnitude and precision of the effect cannot be quantified from this registry analysis record alone.
A log-rank P-value also does not mean that every patient experienced a benefit or that the event hazard was reduced by a particular percentage. Time-to-event analyses incorporate the timing of events and censoring, and interpretation of a single hazard-based treatment summary would additionally depend on the proportional-hazards framework if a Cox model were used.
Because the ClinicalTrials.gov record does not identify the analysis population beyond the registered endpoint, this page does not assume an ITT, per-protocol, or other specific analysis population.
Secondary Endpoint: Overall Response Rate
Reported Fisher exact result
Superiority hypothesis · Fisher exact test
Time frame: about 18 months after the first subject was enrolled.
| Endpoint | Comparison | Method | Reported result |
|---|---|---|---|
| Overall Response Rate by Independent Review Committee | Ofatumumab (Arm A) vs Ibrutinib | Fisher exact test | P < .0001 |
The Fisher exact result of P < .0001 indicates strong statistical evidence against the null hypothesis of no difference in the underlying categorical response comparison, under the reported superiority analysis.
The P-value does not provide the response-rate difference, a risk ratio, an odds ratio, or a confidence interval. The ClinicalTrials.gov record therefore do not establish how large the numerical difference in response rates was.
Because ORR is binary, its interpretation differs fundamentally from PFS and OS. ORR classifies response status, whereas time-to-event endpoints incorporate both whether an event occurred and when it occurred. A statistically significant ORR comparison therefore should not be translated into a particular survival effect.
Secondary Endpoint: Overall Survival
Reported log-rank result
Superiority hypothesis · Log-rank test
OS analysis at study closure, including up to 6 years of study follow-up.
| Endpoint | Comparison | Method | Reported result |
|---|---|---|---|
| Overall Survival | Ofatumumab (Arm A) vs Ibrutinib | Log-rank test | P = 0.1653 |
The reported P = 0.1653 does not provide conventional statistical evidence against the null hypothesis at commonly used two-sided significance levels. It should be read as a statement about the evidence reported by this particular OS comparison, not as proof that the two treatments have identical survival distributions.
The ClinicalTrials.gov record does not report an OS hazard ratio, confidence interval, median survival, or absolute survival probabilities. Consequently, the numerical size and precision of any OS difference cannot be determined from the posted analysis entry alone.
The P-value is not a measure of effect size. A larger P-value can occur with a small effect, substantial statistical uncertainty, limited event information, or other features of the data and design. Conversely, a small P-value would not by itself establish clinical magnitude.
Because OS was analyzed after study closure with up to 6 years of follow-up, the analysis has a different time frame from the primary PFS analysis. These results should therefore not be treated as if they arose from the same data cutoff.
8. What the Three Reported Analyses Tell Us
| Endpoint | Role | Endpoint type | Method | P-value |
|---|---|---|---|---|
| PFS by IRC | Primary | Time-to-event | Log-rank test | <0.0001 |
| Overall Response Rate by IRC | Secondary | Binary | Fisher exact test | <.0001 |
| Overall Survival | Secondary | Time-to-event | Log-rank test | 0.1653 |
The pattern illustrates why endpoint-specific interpretation matters. The primary PFS and secondary ORR analyses have very small reported P-values, while the OS analysis has a P-value of 0.1653. These are different endpoints, measured over different time frames, and analyzed with methods appropriate to their respective outcome structures.
9. Safety Results
The ClinicalTrials.gov record provides serious adverse-event counts by arm. The affected/at-risk counts are reported as 59/191 for ofatumumab and 141/195 for ibrutinib.
| Safety measure | Ofatumumab (Arm A) | Ibrutinib (Arm B) |
|---|---|---|
| Serious adverse events, affected / at risk | 59 / 191 | 141 / 195 |
The affected/at-risk counts show that serious adverse events were reported in both randomized arms, with the registry-reported counts being 59/191 for ofatumumab and 141/195 for ibrutinib. The ClinicalTrials.gov record does not provide a formal statistical comparison of serious adverse events, confidence intervals, or a standardized risk measure.
10. Statistical Methods Explained
Why was a log-rank test used for PFS?
PFS is a time-to-event endpoint. Some participants may experience progression or death during follow-up while others may remain event-free at the end of their available observation period. The log-rank test is designed to compare the survival distributions of randomized groups while using the timing of observed events and accommodating right censoring.
What does a log-rank P-value of <0.0001 mean?
It means that, under the null hypothesis represented by the reported log-rank analysis, the observed separation in event-time information would be very unusual. It does not tell us how many additional months of PFS occurred, what the hazard ratio was, or what percentage of patients benefited.
Why was Fisher exact test used for ORR?
ORR is a categorical endpoint. Each participant can be classified according to whether the response criterion was met. Fisher's exact test evaluates the treatment-by-response table directly and does not rely on a large-sample approximation to the same extent as some other categorical tests.
Why is the OS P-value different from the PFS P-value?
PFS and OS measure different events and have different follow-up structures. The primary PFS analysis was limited to the 06 November 2013 analysis and followed approximately 117 PFS events. The OS analysis was conducted at study closure and included up to 6 years of follow-up. Different endpoints can therefore produce different statistical results without contradiction.
Why does P = 0.1653 not prove that there is no OS difference?
A non-small P-value means the registry-reported analysis does not provide strong evidence against its null hypothesis. It does not establish that the two survival distributions are exactly equal. To characterize the magnitude and precision of an OS difference, an effect estimate such as a hazard ratio together with a confidence interval would be informative, but those quantities are not present in the ClinicalTrials.gov record.
What does censoring mean in a time-to-event analysis?
Censoring occurs when the event of interest has not been observed by the end of a participant's usable follow-up. The participant can still contribute information up to that point. This is one reason time-to-event analysis differs from simply comparing the proportion of participants who experienced an event by a fixed date.
Why should the P-values not be treated as effect sizes?
A P-value describes evidence against a null hypothesis under a specified statistical procedure. It does not directly measure the magnitude of a treatment effect. Two studies can have similar P-values with different effect sizes, and the same effect size can produce different P-values under different sample sizes and event information.
11. Primary Analysis and Event-Driven Time Frames
The primary endpoint was explicitly event-driven: the analysis was conducted after observing approximately 117 PFS events. This is a natural feature of time-to-event trial design because the amount of information for comparing survival distributions depends heavily on the number of observed events rather than only on the number of enrolled participants.
The registry description indicates that approximately 117 PFS events triggered the primary analysis. The registry does not provide a hazard ratio or confidence interval for that analysis in the ClinicalTrials.gov record.
The distinction between enrollment and events is important. RESONATE enrolled 391 participants, but the primary PFS analysis was defined around approximately 117 observed PFS events. Enrollment establishes the randomized study population; events determine how much time-to-event information has accumulated for the primary comparison.
12. Understanding the Log-Rank Test
The log-rank test compares the observed number of events in each treatment group with the number expected under the null hypothesis at each relevant event time. Conceptually, it asks whether the pattern of event occurrence over follow-up differs systematically between the randomized groups.
The actual calculation uses the risk sets at successive event times and incorporates censoring. The registry reports the resulting method and P-value, but not the underlying event-time table.
A log-rank test is therefore different from comparing two fixed-time proportions. A participant who experiences progression early contributes information at a different point in the risk-set sequence from a participant who remains event-free for a longer period. This makes the method well suited to PFS and OS.
13. Fisher Exact Test for Overall Response Rate
The ORR analysis has a different mathematical structure from PFS and OS. Rather than modeling the time until an event, the analysis classifies participants according to response status and compares the resulting categorical distributions between treatment groups.
| Response | No response | |
|---|---|---|
| Ofatumumab | Observed count | Observed count |
| Ibrutinib | Observed count | Observed count |
The table above illustrates the structure of a Fisher exact analysis only; the ClinicalTrials.gov record does not provide the underlying response counts, so no numerical cells are populated. The registry nevertheless identifies Fisher exact testing as the formal method and reports P < .0001.
This distinction is useful when reading clinical-trial results: a response-rate comparison can establish evidence of a difference in the proportion classified as responders, while a PFS analysis asks a different question about time to progression or death.
14. Multiplicity, Interim Analysis, and Other Design Issues
The ClinicalTrials.gov record identifies a superiority hypothesis and three posted statistical analyses: one primary PFS analysis and two secondary analyses for ORR and OS. They do not specify a multiplicity-adjustment procedure, alpha allocation across endpoints, an interim-analysis boundary, or an alpha-spending strategy.
| Design topic | What the ClinicalTrials.gov record establishes |
|---|---|
| Hypothesis type | Superiority. |
| Primary endpoint | One registered primary endpoint: PFS by IRC. |
| Secondary analyses | ORR by IRC and OS. |
| Multiplicity adjustment | Not specified in the ClinicalTrials.gov record. |
| Interim analysis | Not specified as a formal interim-monitoring design in the ClinicalTrials.gov record. |
| Alpha spending | Not specified. |
| Missing-data / imputation strategy | Not specified. |
| Stratification factors | Not specified. |
| Bayesian methods | Not reported. |
| Crossover | Not specified in the ClinicalTrials.gov record. |
This absence of detail is itself important for statistical interpretation. A registry-level analysis record can identify the endpoint and test while leaving the complete prespecified statistical-analysis plan unspecified. The page therefore does not infer unreported error-control procedures or analysis conventions.
15. Analysis Populations and Missing Data
The ClinicalTrials.gov record identifies the total enrollment as 391 and provide serious-adverse-event denominators of 191 for ofatumumab and 195 for ibrutinib. They do not explicitly define the efficacy analysis population, safety analysis population, per-protocol population, or handling of missing endpoint assessments.
Why analysis population matters
A randomized efficacy comparison can be interpreted differently depending on whether all randomized participants, treated participants, or another defined population is analyzed.
Why missing data matter
For response endpoints, missing assessments can affect classification. For time-to-event endpoints, loss to follow-up creates censoring and makes the censoring assumptions relevant to interpretation.
What can be reported here
The registry provides the reported methods and P-values, but not enough information to reconstruct the full analysis populations or missing-data strategy.
What should not be inferred
It would be inappropriate to label the analyses ITT, per-protocol, as-treated, or otherwise without supporting information in the ClinicalTrials.gov record.
16. Statistical Interpretation of P-values
The primary PFS analysis reports P < 0.0001. Under the reported superiority log-rank analysis, this is strong evidence against the corresponding null hypothesis. It does not quantify the size of the PFS difference.
The secondary ORR analysis reports P < .0001 using Fisher exact testing. This provides strong evidence against the corresponding null categorical comparison. It does not reveal the actual response percentages or their difference.
The secondary OS analysis reports P = 0.1653. The registry-reported analysis does not provide conventional statistical evidence against its null hypothesis at commonly used significance levels. This does not prove equality of survival outcomes and does not provide an estimate of the size of any difference.
17. Limitations
- Registry-level statistical detail: the ClinicalTrials.gov record identifies the principal endpoint methods and P-values but does not provide a complete statistical analysis plan.
- No reported effect estimates in the analyses posted on ClinicalTrials.gov: hazard ratios, confidence intervals, median PFS, median OS, and response percentages are not provided in the ClinicalTrials.gov record.
- Different endpoint time frames: the primary PFS analysis occurred on 06 November 2013 after approximately 117 PFS events, whereas the OS analysis occurred at study closure with up to 6 years of follow-up.
- Incomplete multiplicity information: the ClinicalTrials.gov record does not specify how type I error was allocated across the primary and secondary analyses.
- Analysis populations not specified: the ClinicalTrials.gov record does not define the efficacy or safety populations beyond the reported serious-adverse-event denominators.
- Missing-data strategy not specified: the ClinicalTrials.gov record does not describe imputation, censoring rules in detail, or sensitivity analyses.
- Safety comparison not formally reported: serious adverse-event counts by arm are available, but the ClinicalTrials.gov record does not provide a formal statistical comparison.
- No hazard-ratio interpretation: because no hazard ratios are reported in the ClinicalTrials.gov record, a numerical relative reduction in event hazard cannot be calculated or inferred from the P-values.
18. Why This Trial Matters Statistically
RESONATE is a useful statistical teaching case because the same randomized comparison generates evidence across different endpoint structures. PFS and OS require time-to-event methods, whereas ORR is categorical. The registry therefore illustrates why the statistical method should follow the structure of the endpoint rather than applying one test indiscriminately to every outcome.
| Concept | How it appears in RESONATE |
|---|---|
| Randomization | The trial is described as randomized. |
| Parallel-group design | Two treatment arms are evaluated in a parallel design. |
| Superiority testing | The primary and posted secondary analyses use a superiority hypothesis. |
| Time-to-event endpoint | PFS and OS are registered as time-to-event outcomes. |
| Log-rank test | Used for the reported PFS and OS comparisons. |
| Fisher exact test | Used for the reported ORR comparison. |
| Event-driven analysis | The primary PFS analysis was conducted after approximately 117 PFS events. |
| Independent review | The primary PFS and secondary ORR endpoints were assessed by an Independent Review Committee. |
| Different follow-up windows | The primary PFS analysis and later OS analysis use different specified time frames. |
| Safety denominators | Serious adverse events are reported as affected participants over participants at risk in each arm. |
19. A Statistical Reading of RESONATE
The cleanest way to read the registry-reported RESONATE analysis is to begin with the prespecified primary endpoint rather than with individual P-values. PFS was a time-to-event endpoint assessed by an Independent Review Committee, and the primary comparison used a log-rank test after approximately 117 PFS events. The reported P < 0.0001 indicates strong evidence against the null hypothesis under that analysis.
The secondary ORR analysis uses a different statistical framework because response is binary. Fisher exact testing again produced a very small P-value, <.0001. This supports a statistical difference in response classification under the reported test, but the ClinicalTrials.gov record does not reveal the underlying response rates.
OS provides an instructive contrast. The analysis was performed at study closure with up to 6 years of follow-up, and the reported log-rank P-value was 0.1653. That result should not be transformed into a statement that the treatments have identical survival. It indicates that the registry-reported analysis did not provide strong evidence against its null hypothesis.
Most importantly, the three P-values cannot be placed on a single numerical scale of treatment benefit. PFS, ORR, and OS measure different outcomes, use different statistical structures, and have different follow-up descriptions. A complete quantitative comparison would require the corresponding effect estimates, confidence intervals, event counts, and analysis-population definitions.
20. Statistical Concepts in This Trial
Learn more about the methods used in this trial:
21. Related Statistical Calculators
22. Sources
- ClinicalTrials.gov: RESONATE, NCT01578707. Official trial registry record and source for the trial data summarized on this page.
- PubMed: PMID 40266025.
- PubMed: PMID 31512258.
- PubMed: PMID 31196847.
- PubMed: PMID 30842083.
- PubMed: PMID 29470582.
Continue with the underlying statistical methods
Explore the statistical concepts that connect RESONATE's randomized design, categorical response analysis, and time-to-event endpoints.
23. Record Summary
RESONATE was a randomized phase 3 trial of ibrutinib versus ofatumumab in patients with relapsed or refractory chronic lymphocytic leukemia or small lymphocytic lymphoma. The registry identifies PFS by Independent Review Committee as the primary time-to-event endpoint and reports a log-rank P-value of <0.0001 after approximately 117 PFS events. The secondary ORR analysis used Fisher exact testing and reported P < .0001, while the secondary OS analysis used a log-rank test and reported P = 0.1653 at study closure with up to 6 years of follow-up.
The statistical lesson is not simply that some P-values are small and another is not. The more important lesson is that PFS, ORR, and OS require different interpretations because they represent different endpoint structures. The primary PFS result is event-driven and time-to-event based; ORR is categorical; and OS represents a later time-to-event analysis. The ClinicalTrials.gov record does not provide the effect estimates and confidence intervals needed to quantify the magnitude or precision of the treatment differences, so this page deliberately does not infer them.