A Study of Avastin (Bevacizumab) in Combination With Chemotherapy in Patients With Breast Cancer Progressing After First-Line Therapy With Avastin and Chemotherapy (TANIA): Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for A Study of Avastin (Bevacizumab) in Combination With Chemotherapy in Patients With Breast Cancer Progressing After First-Line Therapy With Avastin and Chemotherapy (TANIA) (NCT01250379), a randomized phase 3 study. The primary outcome is Percentage of Participants With Second-Line Progression-Free Survival (PFS) According to Response Evaluation Criteria in Solid Tumors (RECIST) Version 1.1 (v1.1).

2. Clinical Question

The trial addressed whether continuing or reintroducing bevacizumab with chemotherapy was associated with a different second-line disease-progression experience than chemotherapy alone after progression during first-line bevacizumab and chemotherapy. The primary outcome is presented as the percentage of participants with second-line progression-free survival. The registry defines the underlying event process using time from randomization to progressive disease or death from any cause, whichever occurred first.

This distinction matters for interpretation: the endpoint description includes follow-up over time and censoring, while the posted primary result is a percentage. The registry record does not explain in the posted primary result how that percentage was derived from the time-to-event data. The comparison therefore describes the posted proportions and the associated unadjusted risk-difference calculation, rather than establishing a full account of every time-to-event analysis.

3. Trial Design

TANIA was a Phase III, randomized, open-label, parallel-group study sponsored by Hoffmann-La Roche. It enrolled 494 participants with breast cancer. The study status is completed; the registry lists a start in 2011-02 and primary completion in 2013-12. The record describes the intended comparison as bevacizumab plus chemotherapy versus chemotherapy alone for second- and third-line treatment after progression on first-line bevacizumab and chemotherapy.

Participants were randomized between the two treatment strategies. The posted primary outcome includes 247 participants in each group. The chemotherapy-alone group is labelled CT Arm, and the bevacizumab-plus-chemotherapy group is labelled CT+BV Arm. The study was open-label, so masking was not used. The record describes a parallel-group design, meaning participants were assigned to one treatment group rather than crossing between the groups as part of the design.

The registry’s brief summary says treatment was anticipated to continue until third-line disease progression or unacceptable toxicity. Bevacizumab dosing was described as 15 mg/kg every 3 weeks or 10 mg/kg every 2 weeks intravenously, with standard chemotherapy. Specific chemotherapy regimens and detailed arm descriptions are not reported in the registry record.

The primary outcome was assessed in the intent-to-treat population, with tumor assessments described as occurring at baseline, no more than 28 days after randomization, and every 8-9 weeks thereafter according to the treatment site’s standard of care, until approximately 3 years. Several secondary outcomes address response, duration of response, and progression-free survival in second- or third-line treatment. The registry lists 27 outcome measures and 34 analyses. The original sample-size justification and the trial’s own statistical design assumptions are not reported in the registry record.

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

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 Percentage of Participants With Second-Line Progression-Free Survival (PFS) According to Response Evaluation Criteria in Solid Tumors (RECIST) Version 1.1 (v1.1), measured over Baseline (less than or equal to [≤] 28 days after randomization), every 8-9 weeks thereafter according to the standard of care of the treatment site until approximately 3 years.

The primary endpoint is the percentage of participants with second-line progression-free survival according to RECIST v1.1. The registry defines progression-free survival as the time from randomization to progressive disease or death from any cause during second-line treatment with bevacizumab and/or chemotherapy, whichever occurred first. For target lesions, progression was defined as at least a 20 percent increase in the sum of the largest diameters, relative to the smallest recorded sum, or the appearance of one or more new lesions. For non-target lesions, progression meant one or more new lesions or unequivocal progression of existing lesions.

Participants without documented progression or death were censored at their last tumor assessment documenting non-progression. Censoring is a standard way to retain follow-up information when the event of interest has not been observed by a participant’s last assessment; it does not mean that the participant had no later event.

Progression-free survival is used in general because it captures disease progression and death within a defined treatment period, rather than relying only on tumor response. However, the primary result is posted as a percentage of participants, not as a median time or a hazard ratio. The registry record does not report a further explanation of the calculation that connects this percentage to the time-to-event definition. The percentage should therefore be read in conjunction with the endpoint definition and the registry’s stated assessment schedule.

7. Results Posted in the Registry

• CT+BV Arm: 93.9% (232/247) versus 88.7% (219/247) on control; risk difference 5.3 percentage points (95% CI 0.3 to 10.2), risk ratio 1.06 (1.00–1.12); unadjusted two-proportion z-test p = 0.038.

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.

Percentage of Participants With a Second-Line Best Overall Response (BOR) of Complete Response (CR) or Partial Response (PR) According to RECIST v1.1 (Data Cutoff 20 December 2013)

• 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:
  • CT+BV Arm: 20.9% (38/182) versus 16.8% (31/185) on control; risk difference 4.1 percentage points (95% CI -3.9 to 12.1), risk ratio 1.25 (0.81–1.91); unadjusted two-proportion z-test p = 0.312.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between CT+BV Arm and the control. The interval gives the range of effects that remain compatible with the data.

Duration of Second-Line Objective Response (Data Cutoff 20 December 2013)

• Endpoint type: Time to event: compared as a hazard ratio.
• Test: Hazard ratio as reported by the sponsor (Log Rank); it was not recalculated.
• Result:
  • CT Arm vs CT+BV Arm: hazard ratio 1.01 (95% CI 0.51–1.99), reported p = 0.983 (as reported in the registry).
  • CT Arm vs CT+BV Arm: hazard ratio 1.31 (95% CI 0.73–2.34), reported p = 0.360 (as reported in the registry).
• Interpretation:
  • The 95% confidence interval includes no difference (a ratio of 1), so these data do not show a statistically significant difference between CT Arm vs CT+BV Arm and the control. The interval gives the range of effects that remain compatible with the data.
  • The 95% confidence interval includes no difference (a ratio of 1), so these data do not show a statistically significant difference between CT Arm vs CT+BV Arm and the control. The interval gives the range of effects that remain compatible with the data.

Percentage of Participants With Third-Line PFS According to RECIST v1.1

• 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:
  • CT+BV Arm: 96.1% (124/129) versus 94.3% (99/105) on control; risk difference 1.8 percentage points (95% CI -3.7 to 7.4), risk ratio 1.02 (0.96–1.08); unadjusted two-proportion z-test p = 0.509.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between CT+BV Arm and the control. The interval gives the range of effects that remain compatible with the data.

The secondary outcomes cover tumor response, duration of response, and progression-free survival in second- and third-line treatment. Taken together, they add descriptive context about disease response and later treatment, but they do not provide a single uniform pattern that independently settles the primary question. Some secondary estimates lean toward a higher outcome in the combination group, while other comparisons are uncertain or have wide intervals.

These outcomes relate to the primary result in different ways. Response is a measure of tumor shrinkage or disappearance under RECIST criteria, whereas progression-free survival includes progression or death over time. Duration of response applies only to participants who achieved a response, and third-line outcomes concern a later treatment stage. Because these endpoints have different definitions and analysis populations, agreement or disagreement among them should not be treated as a direct replication of the primary comparison.

The response and duration analyses use fewer participants than the randomized groups, and the third-line analyses are also based on subsets entering that treatment line. In general, restricting analysis to a post-randomization subset can affect the comparability achieved through randomization. The registry record does not report a complete explanation of the subset-selection process or adjustment methods for these analyses.

The registry lists many outcome measures and analyses. When numerous hypotheses are examined, some nominally notable findings can occur by chance even if none of the corresponding effects is real. The registry record does not report a multiplicity-control strategy, so unadjusted secondary p-values should be interpreted cautiously and given less weight than a clearly prespecified primary comparison. Secondary results remain useful for understanding what was measured, but they are not equivalent to a confirmed independent treatment effect.

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

• Second-Line PFS by Baseline Risk Factor (Data Cutoff 20 December 2013)
• Percentage of Participants With a Second-Line Best Overall Response (BOR) of Complete Response (CR) or Partial Response (PR) According to RECIST v1.1 (Data Cutoff 20 December 2013)
• Percentage of Participants With a Second-Line CR, PR, Stable Disease, and PD According to RECIST v1.1 (Data Cutoff 20 December 2013)
• Duration of Second-Line Objective Response (Data Cutoff 20 December 2013)
• Percentage of Participants With a Second-Line Documented CR or PR According to RECIST v1.1 Estimated to be Alive and Free of Disease Progression at Months 3, 6, and 9 (Data Cutoff 20 December 2013)
• Percentage of Participants With Third-Line PFS According to RECIST v1.1
• Third-Line PFS
• Percentage of Participants With Second- and Third-Line PFS According to RECIST v1.1

9. Statistical Interpretation

10. Statistical Methods Explained

The trial’s randomized parallel-group design supports a direct comparison of outcomes between the assigned groups. Randomization is intended, in general, to make treatment groups comparable on average at the point of assignment. The primary outcome is specified for the intent-to-treat population, which includes all randomized participants, preserving the comparison based on initial assignment even if treatment later changes or is discontinued. The record does not provide a detailed account of how every post-randomization event or missing assessment was handled in each analysis.

What does the primary risk difference describe?

The primary comparison is expressed as a risk difference: the estimated proportion in the combination group minus the estimated proportion in the chemotherapy group, on the probability scale used for the endpoint. The recalculated estimate is 0.053. Its 95% confidence interval is 0.003 to 0.102. A risk difference is an absolute contrast, unlike a risk ratio or odds ratio, which express a relative comparison. In general, an absolute contrast is interpreted in the context of the outcome definition and the population assessed.

The reported arm percentages are 88.7% and 93.9%. The recalculated comparison uses the posted summary counts and group sizes, and is unadjusted. It is not described as an adjusted model estimate. The unadjusted p-value is 0.038. In general, an unadjusted test does not account for other covariates or for the testing of multiple outcomes. The registry record does not report covariate adjustment for the primary comparison.

How should a p-value be read?

In general, a p-value describes the compatibility of observed data with a specified null hypothesis and statistical model. A smaller p-value can indicate greater incompatibility with that null model, but it is not a measure of effect size, clinical importance, or the probability that a treatment effect is real. Its interpretation depends on the analysis having been selected and conducted appropriately. The registry record does not report whether interim analyses were planned or performed, or give the trial’s complete statistical analysis plan.

What role does RECIST v1.1 play?

RECIST v1.1 provides standardized criteria for classifying tumor response and progression. The registry specifies separate progression rules for target and non-target lesions and includes new lesions in those rules. Standardized criteria can improve consistency in outcome assessment, although the study was open-label. The registry record does not report whether tumor assessments were independently reviewed or blinded to treatment assignment.

The endpoint definition is time-to-event: it begins at randomization and ends at progression or death, whichever occurs first. Participants who have not experienced either event by their last documented non-progressing assessment are censored at that time. In general, time-to-event methods use information about both whether an event occurred and how long participants were followed. The primary posted measure, however, is a percentage. The registry does not report enough detail alongside that measure to reconstruct the precise handling of follow-up time in the primary result.

Why can secondary analyses be harder to interpret?

Secondary outcomes may involve different populations, event definitions, and follow-up windows. For example, response analyses are based on participants with response assessments, while third-line results concern participants reaching third-line treatment. Such analyses can be useful for describing the course of treatment, but a subset defined after randomization may no longer retain all the comparability created by the original assignment. General interpretation should distinguish these descriptive or later-line comparisons from the primary randomized contrast.

The registry reports a chi-squared analysis for the second-line complete or partial response outcome, with a mean difference and confidence interval, and log-rank analyses for duration of response. A chi-squared test is generally used to compare categorical outcome distributions. The log-rank test is generally used to compare time-to-event patterns between groups. The record does not report the full analysis specifications, including any stratification, covariate adjustment, or detailed censoring rules for these secondary analyses.

In general, over-dispersion means that observed variability exceeds what a simple model assumes. When present, it can make standard errors too small and confidence intervals too narrow if not addressed. The registry results do not report an over-dispersion model or indicate that over-dispersion was assessed, so this concept is relevant as a general statistical consideration rather than a claim about this trial.

11. Confidence Intervals and Precision

A confidence interval gives a range of values compatible with the observed data under the assumptions of the analysis. It reflects statistical precision, not the probability that the true value lies within the particular interval. In general, a narrower interval indicates greater precision, while a wider interval leaves more uncertainty about the effect’s size and direction.

For the primary risk difference, the recalculated 95% confidence interval runs from 0.003 to 0.102 around an estimate of 0.053. The interval is entirely above zero, but its lower end is close to zero. Thus, the interval allows for a small positive difference as well as larger positive differences; it does not imply that the point estimate is known precisely. The unadjusted p-value of 0.038 and the confidence interval address related aspects of uncertainty under the same general comparison, but neither conveys clinical importance by itself.

Precision is more limited for several secondary outcomes. For the second-line response comparison, the registry-reported 95% confidence interval for the mean difference extends from -4.2 to 12.4. The reported duration-of-response hazard-ratio intervals extend from 0.51 to 1.99 and from 0.73 to 2.34. Those intervals cover a broad range of possible relative event patterns. The third-line progression-free survival risk-difference interval extends from -0.037 to 0.074, including zero. These results should be read as uncertain estimates, not as proof of no difference.

Confidence intervals do not account automatically for multiplicity, bias, missing data, or limitations in the outcome definition. The registry record does not report a multiplicity adjustment. In general, repeated testing can make nominal confidence levels and p-values less informative across a collection of outcomes. The reported intervals describe uncertainty for their individual analyses, subject to their statistical assumptions.

12. Planning a Similar Trial: Sample Size

This is a planning illustration, not a re-evaluation of the original trial. It asks: if a new study wanted to detect a difference like the one observed, or a smaller one, how large would it need to be? It is deliberately not “observed power” of the finished trial, which is not a meaningful quantity.

The planning illustrations translate an observed effect into hypothetical sample-size requirements. For binary outcomes, the figures are participants per arm. For time-to-event outcomes, the figures are total events across groups, not participants. The second-line response and third-line progression-free survival illustrations therefore use participants per arm, while the duration-of-response illustrations use total events.

The stated planning assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. The effect used for each illustration is the stated fraction of the observed effect. The illustrations show planning for 100%, 75%, and 50% of the observed effect, with columns for participants per arm for 80% power and participants per arm for 90% power, or total events for the time-to-event calculations.

In general, power is the probability that a statistical test will reject a specified null hypothesis if a particular effect is true, under the assumptions used in planning. The 80% column represents a design targeting an 80% chance of detecting the stated effect, and the 90% column represents a higher target chance for that same effect. These are planning calculations based on the observed effect, not the trial’s original design justification; the trial’s own sample-size assumptions are not reported in the registry record. This is not observed power.

Smaller effects generally require more information to detect with the same power and significance level. The illustrations reflect that principle: the rows planning for smaller fractions of the observed effect require larger sample sizes or event totals. They should be read as hypothetical planning aids, not as evidence that the trial was or was not adequately designed.

13. How the Sample Size Is Derived

Power is the probability that a study of a given size rejects the null hypothesis when a specific true effect exists. It depends on four things: the effect to detect, the variability of the outcome, the significance level (here a two-sided α of 0.05, so z_1−α/2 = 1.960) and the desired power (z_1−β = 0.842 for 80% and 1.282 for 90%).

Formula

n is the number of participants in each arm, p₀ and p₁ are the control and treatment proportions, and p̄ is their average.

Worked calculation for this trial

1. Control proportion p₀ = 0.887; treatment proportion p₁ = 0.939; average p̄ = 0.913.
2. First term: z_1−α/2 × √[2p̄(1−p̄)] = 1.96 × 0.3987 = 0.7814.
3. Second term: z_1−β × √[p₀(1−p₀) + p₁(1−p₁)] = 0.842 × 0.3969 = 0.3341.
4. n = (0.7814 + 0.3341)² / (0.0526)² = 449.2, rounded up to 450 participants per arm for 80% power.

The smaller-effect rows in the table above repeat the same calculation with the effect multiplied by 0.75 and 0.5 (for ratios, applied to the logarithm of the ratio).

R code

alpha <- 0.05
z_a   <- qnorm(1 - alpha / 2)          # two-sided
p0 <- 0.88663967611336                          # control proportion
p1 <- 0.939271255060729                          # treatment proportion
n_per_arm <- function(power, frac = 1) {
  q1 <- p0 + (p1 - p0) * frac               # smaller effect = fraction of the observed difference
  pbar <- (p0 + q1) / 2
  z_b <- qnorm(power)
  ceiling(round((z_a * sqrt(2 * pbar * (1 - pbar)) + z_b * sqrt(p0 * (1 - p0) + q1 * (1 - q1)))^2 / (q1 - p0)^2, 6))
}
for (f in c(1, 0.75, 0.5))
  cat(sprintf("%3.0f%% of observed effect: n per arm = %s (80%% power), %s (90%% power)\n", 100 * f,
              format(n_per_arm(0.80, f), big.mark = ","), format(n_per_arm(0.90, f), big.mark = ",")))

# Check against R's built-in function (unrounded n per arm)
cat("power.prop.test:", round(power.prop.test(p1 = p0, p2 = p1, power = 0.80)$n, 1), "\n")

# Simulation: how often does the z-test reject at the 80%-power sample size?
set.seed(1250379)
n    <- n_per_arm(0.80)
reps <- 20000
x0 <- rbinom(reps, n, p0); x1 <- rbinom(reps, n, p1)
ph <- (x0 + x1) / (2 * n)
z  <- ((x1 - x0) / n) / sqrt(ph * (1 - ph) * 2 / n)
cat(sprintf("Simulated power at n = %s per arm: %.3f (%d replicates)\n", format(n, big.mark = ","), mean(abs(z) > z_a, na.rm = TRUE), reps))

Output from running the code

100% of observed effect: n per arm = 450 (80% power), 601 (90% power)
 75% of observed effect: n per arm = 854 (80% power), 1,143 (90% power)
 50% of observed effect: n per arm = 2,043 (80% power), 2,734 (90% power)
power.prop.test: 449.2 
Simulated power at n = 450 per arm: 0.810 (20000 replicates)

The simulation draws many pairs of event counts at the planned sample size and counts how often the two-proportion z-test rejects the null. It should land close to the target power; at very small required sample sizes it can fall below it, because the normal approximation is rough there.

14. Multiplicity, Missing Data and Analysis Populations

Multiplicity arises when more than one statistical hypothesis is tested. This trial’s registry record lists 27 outcome measures and 34 analyses, including multiple response and progression-free survival outcomes. The record does not report a procedure for controlling the overall false-positive risk across these analyses, such as a prespecified hierarchy or adjustment method. Therefore, individual secondary p-values should be understood as unadjusted unless the registry explicitly reports otherwise. The record does not provide enough information to determine which analyses were planned as confirmatory and which were exploratory.

The primary population is identified as the intent-to-treat population, comprising all randomized participants. This is an important design principle because it retains participants in their assigned groups for the primary comparison. However, the registry record does not detail the handling of missing tumor assessments, treatment discontinuation, protocol deviations, or other analysis-population decisions beyond the stated censoring rule. For participants without documented progression or death, the primary endpoint description says censoring occurred at the last tumor assessment showing non-progression.

In general, censoring assumes that the available follow-up provides an appropriate basis for estimating event patterns; informative loss to follow-up can complicate interpretation. The registry record does not report the amount or pattern of missing follow-up, or analyses assessing the sensitivity of results to missing data. It also does not report interim-analysis plans, monitoring boundaries, or whether interim analyses took place. These details are not reported in the registry record, so they cannot be inferred from the final posted results.

Without a full statistical analysis plan, it is difficult to assess the extent of prespecification, handling of multiple testing, and robustness to missing data. This limits how confidently nominal statistical findings can be distinguished from exploratory patterns.

15. 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 5% 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 (4.9%), neutropenia (4.1%), pulmonary embolism (2.4%), thrombocytopenia (2.0%), pyrexia (1.6%).

The registry reports serious adverse events for 55 of 238 participants at risk in the CT Arm and 89 of 245 participants at risk in the CT+BV Arm. Other adverse events are reported for 181 of 238 participants and 218 of 245 participants, respectively. These are participant counts with corresponding at-risk denominators; the registry record does not provide deaths in these safety results.

The most frequently listed serious adverse event terms include febrile neutropenia at 4.9%, neutropenia at 4.1%, pulmonary embolism at 2.4%, thrombocytopenia at 2.0%, pyrexia at 1.6%, and diarrhoea at 1.6%. These terms describe events recorded in the trial and should not be read as proving that treatment caused an event. The record does not provide enough detail to assess event timing, duration, severity beyond the serious-event classification, or individual-level overlap among event categories.

Safety counts are descriptive and use denominators that differ from the randomized group sizes. They should not be directly treated as an adjusted comparison of risk. The registry record does not report a separate statistical analysis of these adverse-event differences in the registry record presented.

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

The registry results describe outcomes observed in a randomized comparison of chemotherapy alone with chemotherapy plus bevacizumab after progression on first-line bevacizumab and chemotherapy. For the primary posted percentage, the combination group had a higher reported proportion with second-line progression-free survival. The unadjusted risk-difference estimate is positive, and its confidence interval lies above zero, but the interval’s lower endpoint is close to zero and the posted percentage does not show the full time-to-event pattern.

That result is not, by itself, a complete assessment of treatment effect. The registry record does not report the primary analysis’s full model specification or an adjusted primary analysis, and the posted endpoint combines a time-to-event definition with a percentage summary. The secondary outcomes involve distinct endpoints and, in several cases, smaller subsets. Their unadjusted results are not a substitute for a prespecified primary analysis, particularly because multiplicity handling is not reported.

Statistical evidence and clinical importance are different questions. A confidence interval describes uncertainty around an estimate under analysis assumptions; it does not establish whether a difference matters to patients. The registry results also do not address all possible outcomes, treatment experiences, or longer-term effects beyond the described follow-up. Safety counts provide a record of adverse events but do not, on their own, establish causation or characterize every aspect of risk.

The appropriate conclusion is limited to what the registry reports: observed group percentages and statistical comparisons for this trial’s specified outcomes. The findings should not be interpreted as an individual treatment recommendation or as proof that either strategy is universally preferable. Broader conclusions require consideration of the full trial report, other evidence, clinical context, and the limitations of the registry’s posted analyses.

17. Why This Trial Matters Statistically

This trial is statistically informative because it compares two treatment strategies in a randomized, parallel-group Phase III study and defines a clinically relevant time-to-event outcome using RECIST v1.1 progression criteria. The intent-to-treat population gives the primary comparison a clear connection to randomized assignment. These design features help readers understand the intended contrast and the endpoint’s event definition.

The results also illustrate why an effect estimate should be considered alongside its precision. The primary unadjusted risk difference is positive, but its interval extends from near zero to a larger positive value. A p-value alone would hide that range of plausible estimates. Similarly, secondary estimates with wide intervals illustrate that sparse outcome information can leave substantial uncertainty even when a point estimate appears large or small.

The registry outcomes span response, response duration, and progression-free survival across second- and third-line treatment. That breadth can describe several dimensions of disease course, but it also introduces differing populations, follow-up structures, and statistical questions. Without a reported multiplicity strategy or full analysis plan, the distinction between a primary finding and exploratory secondary patterns is especially important.

Finally, the primary endpoint is defined as a time-to-event measure but posted as a percentage. This highlights the importance of understanding not only the endpoint label but also the estimand, analysis population, follow-up, censoring, and statistical method behind a result. The registry record supplies some of these elements, while other details are not reported. Reading the estimate with its design and reporting limits is essential to sound statistical interpretation.

18. 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 are specific to the registry information available for this trial. The primary endpoint is described as time from randomization to progression or death, with censoring at the last non-progressing tumor assessment, but the posted primary result is a percentage. The registry record does not explain how this percentage was derived from the time-to-event data or provide a primary hazard ratio or median progression-free survival result in the registry record presented.

The study was open-label, and the record does not report whether outcome assessment was blinded or independently reviewed. This leaves uncertainty about safeguards against assessment bias for outcomes based on tumor evaluation. Detailed chemotherapy regimens and arm descriptions are also not reported in the registry record, limiting interpretation of treatment consistency.

The registry lists multiple secondary endpoints and analyses but does not report multiplicity control, the full statistical analysis plan, interim analyses, or the original sample-size justification. Several secondary outcomes are based on subsets rather than all randomized participants; the record does not fully explain subset formation or provide missing-data sensitivity analyses. The duration-of-response entry contains two reported log-rank hazard-ratio estimates, and the registry record does not explain their distinction.

Safety results are descriptive, use at-risk denominators that differ between outcomes and from the randomized group sizes, and do not include reported deaths in the provided safety data. The record does not provide a detailed account of event timing or attribution in these results. These reporting limits do not negate the trial’s results, but they restrict independent assessment of analysis choices, precision, and generalizability.

19. Related Tutorials

20. Related Statistical Calculators

21. Sources

• ClinicalTrials.gov: A Study of Avastin (Bevacizumab) in Combination With Chemotherapy in Patients With Breast Cancer Progressing After First-Line Therapy With Avastin and Chemotherapy (TANIA) — NCT01250379.
• Registry record last updated 2016-02-11; status last verified 2016-01.
• Results first posted 2015-06-30.
• 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.

22. Record Summary

NCT01250379, known as TANIA, is a completed randomized, open-label, parallel-group Phase III study sponsored by Hoffmann-La Roche. It enrolled 494 participants with locally recurrent or metastatic breast cancer progressing after first-line bevacizumab and chemotherapy. The comparison was standard chemotherapy alone versus bevacizumab plus standard chemotherapy.

The primary outcome was the percentage of participants with second-line progression-free survival according to RECIST v1.1, assessed in the intent-to-treat population. The posted percentages were 88.7% in the CT Arm and 93.9% in the CT+BV Arm. The recalculated unadjusted risk difference was 0.053, with a 95% confidence interval from 0.003 to 0.102 and p=0.038.

Secondary outcomes covered response, response duration, and second- and third-line progression-free survival. The registry record does not report the trial’s original sample-size assumptions, full statistical analysis plan, multiplicity procedure, or interim-analysis details.