THOR Study: A Study of Continued Herceptin (Trastuzumab) in Combination With Second Line Chemotherapy in Patients With HER2 Positive Metastatic Breast Cancer.: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for THOR Study: A Study of Continued Herceptin (Trastuzumab) in Combination With Second Line Chemotherapy in Patients With HER2 Positive Metastatic Breast Cancer. (NCT00448279), a randomized phase 3 study. The primary outcome is Progression-Free Survival (PFS) - Percentage of Participants With an Event.

2. Clinical Question

This was a comparison of treatment strategies after participants had already experienced progression on a trastuzumab-containing first-line regimen. It therefore addresses continuation of trastuzumab alongside second-line chemotherapy, rather than use of trastuzumab as initial treatment or treatment without chemotherapy. The chemotherapy choice and schedule were left to investigators, so the comparison concerns the assigned strategies as described in the registry, not one fixed chemotherapy regimen.

The registry identifies the study as randomized and open-label. It does not report the detailed randomization procedure or describe masking beyond stating that none was used. The clinical question is consequently framed around the observed between-group endpoint percentages, while interpretation also needs to account for the small recorded group sizes and the uncertainty around the comparison.

3. Trial Design

THOR was a phase 3, randomized, open-label, parallel-group study sponsored by Hoffmann-La Roche. Its official title describes a comparison of progression-free survival for participants who continued or discontinued trastuzumab while receiving second-line chemotherapy after progression on first-line chemotherapy plus trastuzumab. The registry lists the study status as completed, enrollment as 58, start as 2007-04, and primary completion as 2010-09.

The two arms were chemotherapy alone and chemotherapy plus trastuzumab. The registry records 29 participants in each arm. Chemotherapy schedule and dose were at the investigator’s discretion in both arms. In the combination arm, trastuzumab was also administered at the investigator’s discretion, using one of the registry-listed schedules: 2 milligrams per kilogram intravenously every 7 days or 6 mg/kg intravenously every 3 weeks.

The brief summary says treatment was anticipated to continue until disease progression and describes a target sample size of 100-500 individuals. The registry record does not report a more specific target or the reasons for the difference between that target range and recorded enrollment. A sample-size justification, details of the randomization sequence, allocation concealment, and protocol deviations are not reported in the registry record.

Because the study was open-label, participants and investigators were not masked to assignment according to the registry. The record does not state whether endpoint assessment was independently or blindly reviewed. It also does not report the precise dates of follow-up for individual participants or the duration of observation. These omissions limit how fully the conduct and timing of the comparison can be reconstructed from the registry record alone.

4. Participant Flow and Accounting

The registry reports how participants moved through each stage of the study. All counts below are exactly as posted; missing counts are not inferred from other stages.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: ITT population

5. Baseline Characteristics

The table reproduces the baseline characteristics posted in the registry, which describe the participants analysed at baseline and help show whether the randomized groups were comparable.

Population: Intent-to-treat (ITT) population: all randomized participants.

6. Primary Endpoint

The registry lists the primary outcome as Progression-Free Survival (PFS) - Percentage of Participants With an Event, measured over Baseline (BL) and every 8 weeks thereafter.

The primary endpoint was the percentage of participants with a progression-free survival event. The registry defines progression-free survival as the time from randomization to documented disease progression according to RECIST version 1.1, a second primary cancer, or death from any cause, whichever occurred first. Participants were censored at their last tumour evaluation. The planned assessment timeframe was baseline and every 8 weeks thereafter.

Although the endpoint name refers to progression-free survival, the posted primary outcome is expressed as a percentage with an event rather than as a median time or a hazard ratio. This distinction matters: the reported percentages describe whether an event was recorded over the study’s assessment period, but do not by themselves describe when events occurred. The registry record does not report a time-specific event estimate or a duration of follow-up that would let readers interpret the percentages at a defined time point.

General statistical context: progression-free survival is used in trials because it combines disease progression and death into an outcome that can capture deterioration before death. Censoring allows participants without an observed event to contribute information up to their last assessment, but interpretation depends on the completeness and timing of those assessments. The registry does not report further details on censoring patterns or missing tumour evaluations.

7. Results Posted in the Registry

• Chemotherapy + Trastuzumab: 58.6% (17/29) versus 58.6% (17/29) on control; risk difference 0.0 percentage points (95% CI -25.4 to 25.4), risk ratio 1.00 (0.65–1.54); unadjusted two-proportion z-test p = 1.000.

8. Secondary Efficacy Endpoints

The registry also posts results for secondary outcomes. For each outcome below, the endpoint type, the test applied to the posted summary statistics, the result and its interpretation are given in turn. Secondary endpoints are usually not adjusted for multiple testing, so they support rather than confirm the primary finding.

Overall Survival (OS) - Percentage of Participants With an Event

• Endpoint type: Binary: the percentage of participants with the event (counts rebuilt from the posted percentages and group sizes).
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • Chemotherapy + Trastuzumab: 34.5% (10/29) versus 55.2% (16/29) on control; risk difference -20.7 percentage points (95% CI -45.7 to 4.3), risk ratio 0.62 (0.34–1.14); unadjusted two-proportion z-test p = 0.113.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Chemotherapy + Trastuzumab and the control. The interval gives the range of effects that remain compatible with the data.

Percentage of Participants With a Best Overall Response of CR or PR

• Endpoint type: Binary: the percentage of participants with the event (counts rebuilt from the posted percentages and group sizes).
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • Chemotherapy + Trastuzumab: 31.0% (9/29) versus 27.6% (8/29) on control; risk difference 3.4 percentage points (95% CI -20.0 to 26.9), risk ratio 1.12 (0.50–2.51); unadjusted two-proportion z-test p = 0.773.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Chemotherapy + Trastuzumab and the control. The interval gives the range of effects that remain compatible with the data.

The secondary outcomes add information about outcomes beyond the primary progression-free survival event percentage, but they do not remove the uncertainty in the primary comparison. The registry lists overall survival event percentage, overall survival time to event, best overall response, and the percentage with complete or partial response among secondary outcomes. The posted numeric results include an overall survival event percentage and the proportion with complete or partial response. The record does not include numeric results for every listed secondary outcome; where results are absent, they are not reported in the registry record.

The posted secondary comparisons point in different descriptive directions: the overall survival event percentage was lower in the combination arm, while the complete or partial response percentage was somewhat higher in that arm. The simple comparisons are imprecise, and the reported unadjusted p-values are 0.113 and 0.773, respectively. These observations do not establish a consistent or definitive pattern across outcomes. They also do not change the fact that the primary event percentages were equal in the registry’s summary.

General statistical context: secondary outcomes can help describe a trial, but they are often interpreted after examining several endpoints. When many outcomes are assessed, the chance of seeing an apparently notable result by chance can increase. Adjustments for multiple comparisons, a prespecified hierarchy, and analysis plans can affect how much weight a secondary result deserves. The registry record does not report a multiplicity procedure or a hierarchy for interpreting these outcomes.

Accordingly, the secondary findings are best treated as descriptive and hypothesis-generating rather than as independent confirmation of a treatment difference. The record’s lack of reported analyses and the uncertainty around the posted comparisons further limit strong conclusions. Secondary results should be considered alongside, not in place of, the primary endpoint and its precision.

Secondary outcomes registered for this trial, for reference (only those above were recalculated):

• Overall Survival (OS) - Percentage of Participants With an Event
• Overall Survival - Time to Event
• Percentage of Participants by Best Overall Response (BOR)
• Percentage of Participants With a Best Overall Response of CR or PR

9. Statistical Interpretation

10. Statistical Methods Explained

What does a risk difference represent?

General statistical context: a risk difference compares the proportion of participants with an outcome in one group with the proportion in the comparator group. It is expressed on the probability scale, with zero representing equal proportions. A positive or negative value indicates the direction of the contrast according to the chosen group ordering. The primary comparison has a posted risk difference of 0.0, consistent with the equal event percentages reported for the two arms.

The risk difference is a direct contrast of proportions, but it does not adjust for other participant characteristics or account for the timing of events. The registry record does not report an adjusted primary analysis. Since progression-free survival is defined using time from randomization, a time-to-event analysis could retain information about when events happened; however, the posted primary measure is an event percentage. The registry record does not report a primary time-to-event estimate or its model.

What does a p-value tell the reader?

General statistical context: a p-value describes how incompatible the observed data are with a specified null model, under the assumptions of the analysis. It is not the probability that the null hypothesis is true, nor does it measure the size or clinical importance of a difference. The posted primary unadjusted p-value is 1.0. In conjunction with equal observed percentages, it indicates no evidence against equal proportions under the simple comparison represented by that calculation. It does not show that the groups are identical in the population.

For secondary outcomes, the registry lists unadjusted p-values of 0.113 for overall survival event percentage and 0.773 for best overall response of complete or partial response. These are secondary comparisons, and the record does not report adjusted analyses. Their p-values should therefore be interpreted cautiously, especially when multiple outcomes are considered. Statistical significance alone would not establish the reliability or importance of a result.

Why does uncertainty matter?

General statistical context: a confidence interval gives a range of values compatible with the observed data under the method used to calculate it. A wider interval means less precision. For the primary comparison, the recalculated interval includes zero and extends substantially in both directions. That uncertainty is important even though the observed arm percentages match. It signals that the data do not pin down the underlying difference precisely.

The primary interval and the other risk-difference intervals are simple, unadjusted recalculations from the posted summary statistics. They are not a replacement for a sponsor-reported analysis. The registry lists no reported analyses for the outcomes. The intervals also do not account for any factors not represented in the posted summaries, such as possible differences in baseline characteristics.

What do the planning illustrations mean?

General statistical context: sample-size planning asks how many participants per arm might be needed to detect a specified effect under assumptions about the statistical test and desired power. It is not a measure of the power achieved by the completed trial. The registry’s planning illustrations are explicitly based on the observed effect, equal allocation, a two-sided significance level of 0.05, and a normal approximation. They are not the trial’s own design justification.

For the overall survival event comparison, the illustration for 100% of the observed effect gives 90 participants per arm for 80% power and 120 participants per arm for 90% power. For 75% of the observed effect, the corresponding figures are 162 and 216 participants per arm; for 50%, they are 366 and 489. For the response comparison, the corresponding 100% figures are 2,735 and 3,660 participants per arm; the 75% figures are 4,820 and 6,452, and the 50% figures are 10,748 and 14,388.

These values are planning illustrations, not observed results or observed power. They depend on the observed effect estimates, which may be unstable in a small trial, and on the stated assumptions. They should not be read as evidence that the trial was designed to detect those effects. The record does not report the original sample-size calculation or its assumptions.

Why is over-dispersion relevant?

General statistical context: over-dispersion arises when observed variability is greater than a statistical model expects. It is relevant for some count outcomes and can make standard errors too small if ignored. The endpoints described in the registry are percentages of participants with events, not count-rate outcomes; over-dispersion is therefore not an issue identified for these posted comparisons. The registry does not report a model that requires an over-dispersion adjustment.

In short, the primary comparison is a simple contrast of event proportions, while the endpoint itself is defined using time-to-event concepts. The posted summaries do not provide the details needed to reconstruct a time-to-event analysis. Readers should distinguish what the registry reports—the percentages, counts, and simple comparisons—from general statistical methods that could be used for a more complete analysis.

11. Confidence Intervals and Precision

The primary arm percentages are both 58.6, and the posted risk difference is 0.0. Equal observed proportions do not imply that the population proportions are exactly equal. With 29 participants recorded in each arm, the estimate is based on a relatively small amount of information, and the uncertainty around the contrast is substantial.

A recalculated unadjusted confidence interval for the primary risk difference has limits of -0.254 and 0.254. The limits are not repeated in numeric form because they exceed the allowed display precision. The interval crosses zero, so the data are compatible with no difference as well as with differences in either direction. Its width is the more important feature for precision: the observed point estimate alone does not narrow the plausible range sufficiently to support a claim of equivalence or a precise estimate of the treatment contrast.

General statistical context: confidence intervals reflect sampling uncertainty under the method used to calculate them. They do not automatically account for missing observations, bias, endpoint misclassification, or unreported features of study conduct. The registry does not report enough detail to assess all of those issues. The interval is also an unadjusted recalculation rather than a reported sponsor analysis.

The secondary outcome intervals are also broad. The response percentages have reported 95% confidence intervals of 12.7 to 47.2 for chemotherapy alone and 15.3 to 50.8 for chemotherapy plus trastuzumab. These intervals overlap considerably, but overlap by itself is not a formal test of the difference between groups. The registry’s risk-difference comparison is a separate contrast and is also imprecise. Overall, the recorded estimates should be read as uncertain rather than as precise estimates of group differences.

12. Multiplicity, Missing Data and Analysis Populations

The registry lists one primary outcome and several secondary outcomes, including overall survival, best overall response, and response categories. The record does not report a statistical plan describing how these outcomes were ordered, which comparisons were confirmatory, or whether any adjustment for multiple testing was planned. A multiplicity correction is not reported in the registry record. Without that information, unadjusted secondary p-values should not be treated as if they had been protected against the increased chance of false-positive findings when multiple comparisons are examined.

The registry identifies the primary analysis population as the ITT population. It does not provide a detailed definition of that population, explain how participants were handled after treatment discontinuation, or report whether all randomized participants were included in each endpoint analysis. The registry’s outcome arm counts are 29 per arm, while the serious adverse event risk sets are 26 and 28. The reason those safety denominators differ from the outcome arm counts is not reported in the registry record.

For progression-free survival, participants were censored at their last tumour evaluation, according to the endpoint description. The record does not report how many participants were censored, the reasons for censoring, the extent of missing tumour assessments, or any sensitivity analyses for missing data. It also does not describe how missing outcome values were handled for the percentage-based summaries.

The registry lists zero reported analyses for the primary outcome and no reported analyses for the posted secondary outcomes. It therefore does not supply the analysis details needed to assess model choice, covariate adjustment, or sensitivity to missing observations. These gaps do not invalidate the descriptive summaries, but they limit how confidently readers can evaluate the robustness of inferential claims. The posted recalculations are unadjusted and should be distinguished from a complete prespecified trial analysis.

13. Safety Results

The registry reports serious adverse events as affected participants divided by participants at risk.

Other (non-serious) adverse events

Participants with at least one non-serious event at or above the registry's 0% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: febrile neutropenia (3.8%), general malaise (3.8%), acute renal failure (3.8%), hospitalisation for intrapleuric chemotherapy and thoracentesis (3.8%), gastric volvulus (3.6%).

The registry reports serious adverse event information for 26 participants at risk in the chemotherapy-alone group, with 4 affected, and 28 at risk in the chemotherapy-plus-trastuzumab group, with 1 affected. For other adverse events, 26 participants were at risk in each group, with 26 affected in each. These risk sets do not match the outcome arm counts, and the reason for the denominator differences is not reported in the registry record.

The listed serious adverse event terms include febrile neutropenia, general malaise, acute renal failure, hospitalisation for intrapleuric chemotherapy and thoracentesis, and gastric volvulus. The registry gives percentages of 3.8% for each of the first four terms and 3.6% for gastric volvulus. It does not provide event counts for each listed term in this summary, so these percentages should not be converted into counts or treated as a complete description of each event category.

General statistical context: adverse event comparisons can be difficult to interpret when denominators differ or when event definitions and follow-up are not fully described. The registry does not report deaths in the adverse-event summary. The figures describe recorded safety information and do not, by themselves, establish a causal difference between treatment groups.

14. What the Result Does, and Does Not, Mean

The registry record describes a randomized comparison of second-line chemotherapy with or without continued trastuzumab after progression on first-line chemotherapy plus trastuzumab. The primary posted outcome is the percentage with a progression-free survival event. That percentage was 58.6 in each arm, and the simple risk difference was 0.0. The data therefore show equal recorded event percentages in this trial’s summary.

That equality is not evidence that the strategies are equivalent. The confidence interval for the unadjusted primary contrast is wide and includes differences in either direction. The trial’s recorded enrollment and group sizes are limited, and the registry does not report a sponsor analysis for the primary result. A non-significant or null-looking comparison cannot establish that meaningful differences are absent.

The endpoint combines progression, a second primary cancer, and death, and the posted outcome is a percentage rather than an estimate of event timing. The record does not report a primary median time-to-event result or a fixed-time estimate. It also does not provide details on follow-up duration, missing tumour evaluations, or censoring beyond the stated rule. Thus, the percentage should not be interpreted as a complete account of when or why events occurred.

The secondary outcomes offer additional descriptive context but do not settle the primary uncertainty. Their unadjusted comparisons are not reported with a documented multiplicity plan. The safety summary likewise gives event figures with differing risk-set denominators, without enough information to explain those differences or assess causality.

General statistical context: a trial result is one estimate from a particular population, design, and period of observation. It can inform assessment of a research question, but its precision and reporting completeness matter. The registry record supports a description of the recorded outcomes; it does not by itself support a broad claim about treatment benefit, lack of benefit, or equivalence. No medical advice follows from these registry summaries.

15. Why This Trial Matters Statistically

Statistically, this trial illustrates the distinction between an observed point estimate and the precision of that estimate. The primary event percentages match, producing a risk difference of 0.0, but the recalculated confidence interval is wide. A reader who focuses only on the point estimate could mistake a lack of observed separation for strong evidence that the underlying groups are the same. The uncertainty interval makes clear why that interpretation is not warranted.

The endpoint also illustrates a common reporting challenge: progression-free survival is defined using event timing and censoring, but the registry’s primary summary is a percentage with an event. Those summaries are not interchangeable. Reporting a time-to-event estimate and a clear follow-up context can help readers understand when events occurred, whereas an event percentage alone does not capture that timing.

The planning illustrations show how sample-size requirements depend on the effect size chosen for planning. They are based on observed secondary effects and stated assumptions, rather than on the trial’s own design justification. Their scale also underscores that small observed effects can require much larger studies to estimate or detect reliably. These calculations are not observed power and do not explain why the trial enrolled the recorded number of participants.

Finally, this record demonstrates why analysis details matter. The registry identifies an ITT population and an endpoint censoring rule, but does not report a full statistical analysis plan, missing-data methods, or multiplicity handling. That information is important for evaluating which comparisons were intended to be confirmatory and how robust the results are. The posted findings are therefore most useful when read as limited, descriptive evidence with substantial uncertainty.

16. Limitations

• The posted summary statistics are the only data these results rest on; patient-level data, full protocols and statistical analysis plans are not available.
• Recalculated results are simple, unadjusted calculations from posted summaries. They do not replace the sponsor's prespecified analysis.
• Participant counts were reconstructed from posted percentages and arm sizes, so small rounding differences are possible.

Several limitations specific to this trial affect interpretation of the registry results. Recorded enrollment was 58, with 29 participants in each outcome arm. The brief summary describes a target sample size of 100-500 individuals, but a detailed sample-size target and justification are not reported in the registry record. The record does not explain why recorded enrollment differs from that broad target range.

The primary outcome is presented as a percentage with an event, although its definition is based on time from randomization and includes censoring at the last tumour evaluation. The registry does not report the distribution of event times, a time-specific progression-free survival estimate, or a median. Follow-up duration and the number and reasons for censored observations are also not reported. Consequently, readers cannot determine from the posted percentage how the event timing or follow-up pattern differed between groups.

Both chemotherapy schedule and dose were at investigator discretion, and trastuzumab was also given according to investigator discretion within the listed schedules. This makes the comparison a study of treatment strategies with clinician-selected regimens rather than a single uniform chemotherapy schedule. The record does not provide information about the specific regimens used or their distribution across arms.

The open-label design is stated, but endpoint assessment procedures are not described. The registry does not report allocation concealment, baseline characteristics, protocol deviations, missing-data handling, or a full analysis plan. It also lists no reported analyses for the primary outcome. Safety risk-set denominators differ from the outcome arm counts, with no explanation in the registry record. These limits constrain assessment of bias, precision, and the completeness of the posted comparisons.

17. Related Tutorials

18. Related Statistical Calculators

19. Sources

• ClinicalTrials.gov: THOR Study: A Study of Continued Herceptin (Trastuzumab) in Combination With Second Line Chemotherapy in Patients With HER2 Positive Metastatic Breast Cancer. — NCT00448279.
• Registry record last updated 2014-10-24; status last verified 2014-10.
• Results first posted 2014-10-24.
• Data retrieved from ClinicalTrials.gov on 2026-11-09.
• Group sizes, means, event counts and sponsor-reported p-values are reproduced as posted. Effect estimates, confidence intervals and p-values not attributed to the sponsor are independently calculated from the posted summary statistics.

20. Record Summary

THOR Study (NCT00448279) was a phase 3, randomized, open-label, parallel-group study sponsored by Hoffmann-La Roche. It compared chemotherapy alone with chemotherapy plus trastuzumab in HER2-positive metastatic breast cancer after progression on first-line chemotherapy plus trastuzumab. The registry records enrollment of 58 and lists the study as completed, with start in 2007-04 and primary completion in 2010-09.

The primary outcome was the percentage of participants with a progression-free survival event, defined using progression, a second primary cancer, or death. The registry reports 58.6 in each arm. The equal observed percentages do not establish equivalence; the recalculated comparison is imprecise, and a sponsor-reported analysis is not listed. Secondary outcome results and safety figures provide additional descriptive information, but the registry record does not report a full analysis plan, multiplicity procedure, or detailed missing-data methods.