← Clinical Trials
Breast Cancer Phase 3 Adjuvant Therapy NCT01772472

KATHERINE: Complete Statistical Analysis of Trastuzumab Emtansine in HER2-Positive Breast Cancer

An independent statistical analysis of the randomized phase 3 KATHERINE trial comparing trastuzumab emtansine with trastuzumab as adjuvant therapy in patients with HER2-positive breast cancer who had residual tumor in the breast or axillary lymph nodes following preoperative therapy.

Trial status: COMPLETED  ·  Enrollment: 1486  ·  Primary completion: 2018-07-25
Scope of this record

This page separates reported trial results from statistical interpretation. Numerical results on this page are restricted to the ClinicalTrials.gov trial data posted on ClinicalTrials.gov for KATHERINE. ClinicalTrials.gov provides the official trial registry record.

Registry 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

KATHERINE was a randomized, parallel, phase 3 oncology trial evaluating trastuzumab emtansine versus trastuzumab as adjuvant therapy in patients with HER2-positive breast cancer who had residual tumor in the breast or axillary lymph nodes following preoperative therapy. The registry reports one primary time-to-event endpoint and a series of secondary and other pre-specified time-to-event analyses.

1486
Enrolled
2 randomized arms
3
Phase
Phase 3
0.54
Primary HR
95% CI 0.44–0.66
<0.0001
Primary P-value
Superiority hypothesis
FeatureKATHERINE
Trial nameKATHERINE
NCT identifierNCT01772472
PhasePhase 3
ConditionBreast Cancer
AllocationRandomized
Design modelParallel
MaskingNone
Primary purposeTreatment
Enrollment1486
InterventionsTrastuzumab; trastuzumab emtansine
Primary endpointInvasive Disease-free Survival (IDFS) Rate at 3 Years
Primary endpoint typeTime-to-event
Primary analysisLog-rank test with hazard ratio as the effect measure
HypothesisSuperiority
Results postedYes
Statistical analyses posted15
Lead sponsorHoffmann-La Roche
Sponsor typeIndustry

2. Clinical Question

The clinical question was whether trastuzumab emtansine, compared with trastuzumab, improved invasive disease-free survival when used as adjuvant therapy in patients with HER2-positive breast cancer who had residual tumor in the breast or axillary lymph nodes following preoperative therapy.

Population

Patients with HER2-positive breast cancer who had residual tumor in the breast or axillary lymph nodes following preoperative therapy.

Intervention

Trastuzumab emtansine.

Comparator

Trastuzumab.

Primary question

Does trastuzumab emtansine improve the time-to-event endpoint of invasive disease-free survival relative to trastuzumab?

3. Trial Design

01
Enroll1486 participants
02
Randomize2 treatment groups
03
Adjuvant therapyTrastuzumab or trastuzumab emtansine
04
Follow-upTime-to-event outcomes
05
AnalysisLog-rank and hazard ratio
Allocation
Randomized allocation in a parallel-group phase 3 design.
Masking
The registry describes the study as having no masking.
Primary purpose
Treatment.
Study status
Completed. Start date: 2013-04-03. Primary completion: 2018-07-25.
ARM 1

Trastuzumab

  • Trastuzumab
  • Comparator treatment in the randomized trial
ARM 2

Trastuzumab emtansine

  • Trastuzumab emtansine
  • Intervention treatment in the randomized trial

The registry identifies randomization, rather than treatment actually received, as the basis for the primary efficacy analysis. This distinction becomes important when interpreting the reported hazard ratio because randomization is what establishes the treatment groups for the principal comparative analysis.

4. Endpoints

Primary endpoint

EndpointTime frameTypeAnalysis
Invasive Disease-free Survival (IDFS) Rate at 3 Years At Year 3 Time-to-event Log-rank test; hazard ratio

The registry definition states that an IDFS event was defined as the first occurrence of specified events, beginning with ipsilateral invasive breast tumor recurrence and ipsilateral local-regional invasive breast cancer recurrence involving the axilla, regional lymph nodes, chest wall and/or skin of the ipsilateral breast); distant recurrence (i.e., evidence of breast cancer in any anatomic site-other than the 2 above-mentioned sites - that has either been histologically confirmed or clinically diagnosed as recurrent invasive breast cancer); contralateral invasive breast cancer; death attributable to any cause including breast cancer, non-breast cancer or unknown cause . 3-year IDFS rate in ITT population was estimated using Kaplan Meier (KM) method and the percentage of participants who were event-free 3 years after randomization was estimated.

Secondary and other pre-specified endpoints with posted analyses

EndpointTime frameRole
IDFS Including Second Primary Non-breast Cancer (SPNBC) Rate at 3 YearsAt Year 3Secondary
IDFS Including SPNBC Rate at 7 YearsAt Year 7Secondary
IDFS Including SPNBC Rate at 8 YearsAt Year 8Secondary
Disease-free Survival (DFS) Rate at 3 YearsAt Year 3Secondary
DFS Rate at 7 YearsAt Year 7Secondary
DFS Rate at 8 YearsAt Year 8Secondary
Overall Survival (OS) Rate at 5 YearsAt Year 5Secondary
OS Rate at 7 YearsAt Year 7Secondary
OS Rate at 8 YearsAt Year 8Secondary
Distant Recurrence-free Interval (DRFI) Rate at 3 YearsAt Year 3Secondary
DRFI Rate at 7 YearsAt Year 7Secondary
DRFI Rate at 8 YearsAt Year 8Secondary
IDFS Rate at 7 YearsAt Year 7Other pre-specified
IDFS Rate at 8 YearsAt Year 8Other pre-specified

All of the posted analyses in the ClinicalTrials.gov record use the same basic comparative framework: the ITT population, comparison of trastuzumab with trastuzumab emtansine, a log-rank test, and a hazard ratio as the effect measure. That consistency makes the registry results useful for examining how one randomized time-to-event framework was applied across several related endpoints and time points.

5. Statistical Methodology

Intention-to-treat analysis

The registry defines the ITT population as including all participants who were randomized to the study regardless of whether they received any study treatment. The reported primary and secondary efficacy analyses use this population.

The statistical importance of ITT analysis is that treatment assignment remains the defining group variable after randomization. The analysis therefore preserves the treatment comparison created by the randomized design rather than redefining groups according to subsequent treatment exposure.

Time-to-event analysis

The primary endpoint and all of the statistical analyses posted on ClinicalTrials.gov are classified as time-to-event endpoints. Unlike a simple binary outcome observed at a single visit, a time-to-event analysis uses information about when an event occurs and can accommodate participants whose event status is not observed during the available follow-up.

Core time-to-event quantities
Event time  →  comparison of the time-to-event experience between randomized groups

The registry reports the treatment effect using a hazard ratio and compares the groups using a log-rank test.

Log-rank test

The reported statistical method is a log-rank test. The log-rank procedure compares the survival experience of two groups across the observed follow-up rather than reducing the analysis to a single fixed-time proportion.

Conceptually, at each observed event time the method compares the events that occurred with the numbers of participants at risk in the two groups. The evidence is then accumulated across follow-up. This makes the method particularly appropriate for randomized comparisons in which the outcome is naturally expressed as time to an event.

Hazard ratio

The registry reports the treatment effect as a hazard ratio. A hazard ratio compares the instantaneous event rates between treatment groups within the time-to-event framework.

Interpretive reference point
HR = 1  →  equal estimated hazards    |    HR < 1  →  lower estimated hazard for the numerator group

For the analyses posted on ClinicalTrials.gov, the groups are reported as trastuzumab versus trastuzumab emtansine. Thus an HR below 1 corresponds to a lower estimated hazard for trastuzumab relative to trastuzumab emtansine under that ordering.

That direction matters. The numerical value of a hazard ratio cannot be interpreted correctly without knowing which treatment is in the numerator. In this page, the registry's group comparison is explicitly Trastuzumab vs Trastuzumab Emtansine.

Confidence intervals

The primary analysis reports a 95% confidence interval of 0.44 to 0.66 around the hazard ratio of 0.54. The secondary analyses likewise provide 95% confidence intervals around their hazard-ratio estimates.

A confidence interval communicates statistical precision around the estimated effect under the analysis framework. It does not describe the range of outcomes that individual patients will experience, and it is not a probability statement that the true hazard ratio lies within the particular interval.

6. Primary Result: Invasive Disease-free Survival at 3 Years

The registry reports a formal statistical analysis of the primary endpoint, Invasive Disease-free Survival (IDFS) Rate at 3 Years, using a log-rank test in the ITT population. The effect measure is a hazard ratio and the hypothesis type is superiority.

Hazard ratio for invasive disease-free survival

0.54

95% CI: 0.44–0.66   ·   P < 0.0001

Comparison: Trastuzumab vs Trastuzumab Emtansine   ·   ITT population

Primary endpointAnalysis populationMethodEffect measureEstimate95% CIP-value
IDFS Rate at 3 Years ITT Log-rank test Hazard ratio 0.54 0.44–0.66 <0.0001
Clinical Biostats interpretation

The reported HR of 0.54 means that, with trastuzumab as the numerator group in the reported comparison, the estimated instantaneous IDFS event hazard was 54% of that estimated for trastuzumab emtansine under the fitted time-to-event comparison. Expressed as a simple relative complement, this corresponds to a 46% lower estimated hazard for trastuzumab relative to trastuzumab emtansine.

It does not mean that 46% of participants avoided an event, that 46% of participants benefited, or that each participant experienced a 46% reduction in personal risk. A hazard ratio is a relative time-to-event measure, not an absolute risk difference.

The 95% CI of 0.44–0.66 indicates the statistical uncertainty surrounding the estimated hazard ratio. Its entire range is below 1, so the reported interval is consistent with a lower estimated hazard for the numerator group throughout the interval. The width of the interval also matters: it communicates that the estimate is not an exact quantity.

The p-value of <0.0001 addresses evidence against the relevant null hypothesis under the stated statistical test. It does not measure the size of the treatment effect, the probability that the treatment is beneficial, or the clinical importance of the result.

Because the endpoint is time-to-event and the effect measure is a hazard ratio, interpretation should remain tied to the survival-analysis framework. The registry extract in the ClinicalTrials.gov record does not provide additional model details about proportional-hazards diagnostics, so no stronger claim about that assumption is made here.

7. Secondary Results: IDFS Including Second Primary Non-breast Cancer

Three-year analysis

IDFS Including SPNBC at 3 Years

HR 0.57

95% CI: 0.46–0.69   ·   P < .0001

ITT population   ·   Log-rank test   ·   Superiority

Clinical Biostats interpretation

The HR of 0.57 indicates a lower estimated event hazard for trastuzumab relative to trastuzumab emtansine in the reported comparison, corresponding to a 43% lower estimated hazard on the relative hazard scale.

The 95% CI of 0.46–0.69 describes uncertainty around that estimate and remains below 1 throughout the reported interval. The p-value of <.0001 is evidence from the reported log-rank analysis against the relevant null hypothesis; it should not be interpreted as a measure of effect magnitude.

Seven- and eight-year analyses

EndpointTime frameHR95% CIP-value
IDFS Including SPNBC Rate at 7 YearsAt Year 70.570.46–0.69<0.0001
IDFS Including SPNBC Rate at 8 YearsAt Year 80.570.46–0.69<.0001

The ClinicalTrials.gov record reports the same hazard-ratio estimate and confidence interval at Years 7 and 8 as at Year 3 for this endpoint. These entries should be read as the registry's posted analyses at those specified time frames; they should not be interpreted as proof that the underlying survival curves, event counts, or follow-up distributions are identical.

8. Secondary Results: Disease-free Survival

EndpointTime frameHR95% CIP-value
Disease-free Survival (DFS) Rate at 3 YearsAt Year 30.580.48–0.70<0.0001
DFS Rate at 7 YearsAt Year 70.580.48–0.70<0.0001
DFS Rate at 8 YearsAt Year 80.580.48–0.70<.0001

Each DFS analysis was performed in the ITT population using a log-rank test, with hazard ratio as the effect measure and a superiority hypothesis. The registry-reported estimates are below 1, indicating a lower estimated hazard for trastuzumab in the reported trastuzumab-versus-trastuzumab-emtansine ordering.

Reading the DFS estimates

The HR of 0.58 corresponds to a 42% lower estimated hazard on the relative hazard scale for the numerator group. The 95% CI of 0.48–0.70 communicates uncertainty around that estimate. The p-values of <0.0001 or <.0001, depending on the registry entry, quantify statistical evidence under the reported test rather than the magnitude of the observed effect.

9. Secondary Results: Overall Survival

EndpointTime frameHR95% CIP-value
Overall Survival (OS) Rate at 5 YearsAt Year 50.700.53–0.910.0082
OS Rate at 7 YearsAt Year 70.700.53–0.910.0082
OS Rate at 8 YearsAt Year 80.700.53–0.910.0082

Overall survival hazard ratio

0.70

95% CI: 0.53–0.91   ·   P = 0.0082

Reported at Years 5, 7, and 8   ·   ITT population

Clinical Biostats interpretation

An HR of 0.70 means the estimated instantaneous death hazard for trastuzumab was 70% of that for trastuzumab emtansine in the reported comparison. On the relative hazard scale, that corresponds to a 30% lower estimated hazard for the numerator group.

The 95% CI of 0.53–0.91 gives the reported uncertainty interval around the HR. It is narrower than a point estimate alone can communicate and remains below 1 across the interval. The p-value of 0.0082 provides evidence under the reported log-rank analysis but does not say that the effect size is 0.0082, nor does it quantify clinical importance.

OS is a time-to-death endpoint, so the hazard ratio should not be confused with a difference in mortality percentages at a particular time point. The ClinicalTrials.gov record does not provide the corresponding absolute survival percentages, median survival times, event counts, or Kaplan-Meier estimates needed for those additional interpretations.

10. Secondary Results: Distant Recurrence-free Interval

EndpointTime frameHR95% CIP-value
Distant Recurrence-free Interval (DRFI) Rate at 3 YearsAt Year 30.600.47–0.76<0.0001
DRFI Rate at 7 YearsAt Year 70.600.47–0.76<0.0001
DRFI Rate at 8 YearsAt Year 80.600.47–0.76<.0001

The DRFI analyses again use the ITT population and log-rank testing. The HR of 0.60 corresponds to a 40% lower estimated hazard for the numerator group on the relative hazard scale. The 95% CI of 0.47–0.76 describes uncertainty around that estimate.

Why the endpoint-specific estimate matters

Although several KATHERINE endpoints use the same log-rank and hazard-ratio framework, the endpoints are not interchangeable. IDFS, DFS, OS, and DRFI represent different clinical event definitions. A hazard ratio from one endpoint should therefore be interpreted for that endpoint rather than transferred to another outcome.

11. Other Pre-specified IDFS Results

EndpointTime frameHR95% CIP-valueRole
IDFS Rate at 7 YearsAt Year 70.540.44–0.66<0.0001Other pre-specified
IDFS Rate at 8 YearsAt Year 80.540.44–0.66<.0001Other pre-specified

The registry reports the same HR and 95% confidence interval for IDFS at Years 7 and 8 as for the primary IDFS analysis at Year 3. These results reinforce the importance of distinguishing the endpoint definition from the time frame named in the outcome measure: the registry labels the primary endpoint at Year 3, while the other pre-specified entries are separately labeled at Years 7 and 8.

12. Results Across the Time Frames

Reported hazard ratios by endpoint family
IDFS
0.54
IDFS + SPNBC
0.57
DFS
0.58
OS
0.70
DRFI
0.60

This visual is a direct representation of the hazard-ratio estimates reported in the ClinicalTrials.gov record; it is not a reconstruction of Kaplan-Meier curves. The estimates span from 0.54 for IDFS to 0.70 for OS, with the other reported endpoint families between those values.

The most important statistical point is that these estimates should not be treated as five measurements of the same outcome. Each endpoint has its own event definition and time frame. The fact that several estimates are numerically similar does not establish equivalence among the endpoints.

13. Safety Results

The ClinicalTrials.gov record reports serious adverse events by randomized treatment arm as affected participants over participants at risk.

Safety measureTrastuzumabTrastuzumab emtansine
Serious adverse events58 / 72094 / 740

Trastuzumab

58 of 720 participants were reported as affected by serious adverse events.

Trastuzumab emtansine

94 of 740 participants were reported as affected by serious adverse events.

These figures are presented as reported counts over the corresponding numbers at risk. They should not be converted into a new derived percentage for this page because the data rule for this record is to report numbers exactly as reported in the registry. The safety data also represent a different analytical question from the efficacy hazard ratios: serious adverse events describe an adverse-event outcome, whereas the reported efficacy estimates describe time-to-event treatment comparisons.

Safety interpretation: the ClinicalTrials.gov record contains serious adverse events by arm but do not provide a formal comparative statistical analysis of those safety counts. Accordingly, this page does not assign a hazard ratio, confidence interval, or p-value to the serious-adverse-event comparison.

14. Statistical Methods Explained

Why was a log-rank test used?

The primary endpoint is a time-to-event outcome, and the registry reports the log-rank test as the analysis method. A log-rank test is designed to compare the event-time distributions of two groups across follow-up while incorporating the timing of observed events and the information contributed by participants who remain event-free during follow-up.

What does the primary hazard ratio of 0.54 mean?

Because the reported comparison is trastuzumab versus trastuzumab emtansine, an HR of 0.54 means that the estimated instantaneous event hazard for the trastuzumab group was 54% of that for the trastuzumab emtansine group under the reported analysis. The corresponding relative complement is a 46% lower estimated hazard. It does not mean a 46% absolute reduction in events.

Why is the analysis population important?

The registry explicitly states that the ITT population includes all participants randomized to the study regardless of whether they received any study treatment. This keeps participants attached to their randomized treatment assignment and preserves the principal comparison created by randomization.

What does the 95% confidence interval tell us?

For the primary HR of 0.54, the 95% CI is 0.44–0.66. The interval describes statistical uncertainty around the estimated hazard ratio under the analysis framework. A confidence interval is more informative than the point estimate alone because it shows the range of effect estimates compatible with the statistical uncertainty represented by the interval.

Why doesn't the p-value measure effect size?

The primary p-value is <0.0001. A p-value quantifies evidence against a null hypothesis under the specified statistical model and test. It is affected by both the magnitude of the observed effect and the amount of information in the data, so it cannot be interpreted as a direct measure of how large or clinically important the treatment effect is.

Why is this a time-to-event analysis rather than a simple percentage comparison?

The registry classifies the primary endpoint as time-to-event. Time-to-event methods preserve information about when events occur rather than treating every participant as if they had exactly the same follow-up duration. The log-rank test and hazard ratio are therefore aligned with the structure of the endpoint.

Why should different endpoints not be collapsed into one result?

IDFS, IDFS including SPNBC, DFS, OS, and DRFI are separately defined outcomes in the registry. Even when they use the same statistical method, their events and clinical meanings differ. A hazard ratio of 0.54 for IDFS and a hazard ratio of 0.70 for OS are therefore separate statistical findings, not interchangeable estimates of one underlying quantity.

15. Understanding the Primary Confidence Interval

Primary estimate
HR = 0.54    |    95% CI = 0.44–0.66

The point estimate is 0.54, while the interval communicates the statistical precision around that estimate. The interval should be read as a range describing uncertainty about the estimated treatment effect, not as a range of individual patient outcomes.

The interval is particularly useful because it shows that the primary estimate is not being presented as an exact property of the trial population. Sampling variability means that a future study conducted under the same general design could produce a different estimate. The confidence interval gives a quantitative indication of that uncertainty for the observed analysis.

It is also important that the confidence interval is attached to the hazard ratio. It therefore does not directly answer questions about absolute event rates, absolute risk differences, number needed to treat, or survival probabilities at a particular time. Those require corresponding absolute outcome estimates, which are not reported in the ClinicalTrials.gov record used for this page.

16. Reading the P-values Correctly

Endpoint familyReported P-valueWhat it addresses
Primary IDFS at 3 Years<0.0001Evidence from the reported log-rank comparison under a superiority hypothesis
IDFS including SPNBC<.0001 or <0.0001Evidence from the reported log-rank comparison
DFS<.0001 or <0.0001Evidence from the reported log-rank comparison
OS0.0082Evidence from the reported log-rank comparison
DRFI<.0001 or <0.0001Evidence from the reported log-rank comparison

The repeated presence of small p-values does not mean that all endpoints have the same effect size. For example, the OS HR is 0.70, whereas the primary IDFS HR is 0.54. The p-values answer a different question from the hazard ratios.

A useful rule

Hazard ratio → magnitude and direction of relative effect.
Confidence interval → uncertainty and precision around the effect estimate.
P-value → evidence against the null hypothesis under the specified test.

17. Randomization and the ITT Principle

Randomization is the central design feature that allows the trial to compare the two treatment strategies while reducing systematic differences in treatment assignment. The registry identifies the study as randomized and parallel, with two interventions: trastuzumab and trastuzumab emtansine.

The ITT definition posted on ClinicalTrials.gov for the statistical analyses is explicit: all participants randomized to the study are included regardless of whether they received any study treatment. The same ITT framework is reported for the primary endpoint and the registry-reported secondary analyses.

What ITT preserves

The treatment comparison generated by randomization, because participants remain classified according to their randomized assignment.

What ITT does not guarantee

It does not guarantee equal numbers, equal follow-up, or equal event experience between randomized groups.

For time-to-event outcomes, ITT also means that the statistical analysis is not restricted only to participants who completed treatment. That distinction helps prevent post-randomization treatment behavior from redefining the principal efficacy comparison.

18. Why the Hazard Ratio Direction Matters

The analyses posted on ClinicalTrials.gov consistently specify the groups compared as Trastuzumab vs Trastuzumab Emtansine. This ordering is essential when interpreting every HR on the page.

Endpoint familyHRRelative interpretation for trastuzumab vs trastuzumab emtansine
IDFS0.5446% lower estimated hazard
IDFS including SPNBC0.5743% lower estimated hazard
DFS0.5842% lower estimated hazard
OS0.7030% lower estimated hazard
DRFI0.6040% lower estimated hazard

These are simple interpretations of the reported HRs using 1 − HR. They do not transform the hazard ratios into absolute risk reductions. The distinction is important because a relative hazard reduction and an absolute probability difference are fundamentally different measures.

19. Interpreting the Repeated Long-term Estimates

The ClinicalTrials.gov record contains several endpoint families measured at multiple time points. Some estimates remain numerically identical across those time points. For example, the IDFS including SPNBC HR is reported as 0.57 at Years 3, 7, and 8, with a 95% CI of 0.46–0.69. The DFS HR is likewise reported as 0.58 at Years 3, 7, and 8.

These repeated numbers should be handled carefully. A repeated registry estimate is a reported statistical result; it should not be reverse-engineered into an assumption about the underlying event process. Without individual participant event and censoring information, one cannot reconstruct the Kaplan-Meier curves or independently verify how the estimates were generated.

Educational note: this page does not draw fabricated Kaplan-Meier curves. A valid Kaplan-Meier reconstruction requires appropriate event and censoring information or sufficiently detailed source data.

20. Limitations

21. Why This Trial Matters Statistically

KATHERINE is a useful teaching example because its registry record connects a randomized phase 3 design to a coherent family of time-to-event analyses. The primary endpoint, secondary endpoints, and other pre-specified outcomes are all analyzed within an ITT framework using the log-rank test and hazard ratio.

Statistical conceptHow it appears in KATHERINE
RandomizationThe study uses randomized allocation with two parallel treatment groups.
Intention-to-treat analysisThe reported efficacy population includes all randomized participants regardless of whether they received study treatment.
Time-to-event endpointThe primary endpoint is classified as time-to-event.
Log-rank testThe reported comparative method for the primary and registry-reported secondary analyses.
Hazard ratioThe reported effect measure across the statistical analyses reported.
Confidence interval95% intervals quantify uncertainty around the reported hazard ratios.
Superiority testingThe primary analysis is explicitly classified under a superiority hypothesis.
Multiple endpointsThe registry reports a primary endpoint plus multiple secondary and other pre-specified outcomes.
Safety analysisSerious adverse events are reported by randomized treatment arm as affected/at-risk counts.

The statistical lesson is broader than the individual HR of 0.54. A clinical trial result becomes interpretable only when the endpoint, analysis population, comparison direction, statistical method, effect measure, confidence interval, and hypothesis framework are read together.

22. Clinical Interpretation vs Statistical Interpretation

Statistical interpretation

The primary randomized comparison reports an HR of 0.54 with a 95% CI of 0.44–0.66 and P < 0.0001 using a log-rank analysis in the ITT population. The registry classifies the hypothesis as superiority.

Clinical interpretation

The statistical result concerns invasive disease-free survival as defined by the registry. It should be interpreted alongside the endpoint definition, follow-up time, absolute outcome measures when available, and safety information rather than as a standalone measure of overall treatment value.

The distinction is deliberate. Statistical significance describes evidence under a specified model and test. Clinical interpretation requires understanding what outcome was measured, how large the estimated effect was, how precisely it was estimated, and what other evidence is relevant to the treatment comparison.

23. A Compact Statistical Reading of KATHERINE

QuestionAnswer from the ClinicalTrials.gov record
What was randomized?Participants were randomized between trastuzumab and trastuzumab emtansine in a parallel phase 3 study.
What was the primary endpoint?IDFS Rate at 3 Years, a time-to-event endpoint.
Who was analyzed?The ITT population, including all randomized participants regardless of whether they received study treatment.
How was the primary comparison performed?Log-rank test.
What was the effect measure?Hazard ratio.
What was the primary HR?0.54.
What was the primary 95% CI?0.44–0.66.
What was the primary P-value?<0.0001.
What was the hypothesis type?Superiority.
What safety information is reported?Serious adverse events: 58/720 for trastuzumab and 94/740 for trastuzumab emtansine.

24. Related Tutorials

Learn more about the methods used in this trial:

25. Related Calculators

26. Sources

Continue with the statistical methods

Explore the core survival-analysis concepts that connect randomized trials with hazard ratios, confidence intervals, log-rank tests, and time-to-event endpoints.

27. Record Summary

KATHERINE is a randomized phase 3 trial with 1486 enrolled participants and two parallel treatment groups, trastuzumab and trastuzumab emtansine. Its primary endpoint is a time-to-event measure, Invasive Disease-free Survival (IDFS) Rate at 3 Years, analyzed in the ITT population using a log-rank test and a hazard ratio under a superiority hypothesis.

The primary reported HR is 0.54 with a 95% CI of 0.44–0.66 and P < 0.0001. The same general statistical framework is used for the registry-reported secondary analyses of IDFS including SPNBC, DFS, OS, and DRFI, as well as the other pre-specified IDFS analyses. The reported HRs range from 0.54 to 0.70 across these endpoint families.

The statistical interpretation depends on keeping several concepts distinct: the hazard ratio describes relative event hazard, the confidence interval describes uncertainty around that estimate, and the p-value addresses evidence under the specified hypothesis test. None of these quantities alone describes absolute clinical risk or the experience of an individual participant.

The registry also supplies serious adverse-event counts of 58/720 for trastuzumab and 94/740 for trastuzumab emtansine. These safety figures should be considered separately from the efficacy hazard ratios because they represent a different outcome and are not accompanied in the ClinicalTrials.gov record by a formal comparative statistical analysis.

Clinical Biostats methodology: A trial-results page should not merely reproduce a statistical estimate. The goal is to show how the endpoint, randomized comparison, analysis population, statistical test, effect measure, confidence interval, p-value, and limitations fit together while clearly separating reported evidence from educational interpretation.