A Study of Tarceva (Erlotinib) in Combination With Avastin (Bevacizumab) in Patients With Advanced Non-Small Cell Lung Cancer.: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for A Study of Tarceva (Erlotinib) in Combination With Avastin (Bevacizumab) in Patients With Advanced Non-Small Cell Lung Cancer. (NCT00531960), a randomized phase 2 study. The primary outcome is Percentage of Participants With Disease Progression or Death.

2. Clinical Question

The study compared two treatment strategies that both included bevacizumab: adding erlotinib or adding platinum-based chemotherapy. Random allocation supports a between-group comparison, but the open-label design means participants and study personnel were not masked to treatment. This is a Phase 2 study, and the registry record describes an efficacy and safety comparison. The posted primary results show similar event percentages at the level of the reported summaries, with uncertainty that spans values favoring either group. They do not establish a definitive treatment difference.

3. Trial Design

This was a randomized, open-label, parallel-group Phase 2 study sponsored by Hoffmann-La Roche. It enrolled 124 participants with advanced non-small cell lung cancer. The brief summary describes first-line treatment and a target sample size of 100-500 individuals. The registry lists two groups: bevacizumab plus erlotinib and bevacizumab plus chemotherapy. Masking was none, so the treatment assignments were not blinded.

In the erlotinib group, participants received bevacizumab 15 mg/kg intravenously on Day 1 of Cycles 1 through 7 and erlotinib 150 mg orally daily. In the comparator group, bevacizumab was given on the same cycle schedule with 4 to 6 cycles of a standard platinum-containing chemotherapy regimen. The registry descriptions state that treatment continued until disease progression, unacceptable toxicity, death, or withdrawal. The particular chemotherapy regimen details are not reported in the registry record.

The primary outcome was the percentage of participants with disease progression or death, with progression defined using RECIST version 1.0. The primary population is identified as the FAS. The assessment schedule included screening and evaluations at the end of every second cycle through Cycle 8, then every 12 weeks thereafter until study end and final analysis, up to a maximum of 42 months. The registry also lists outcomes for death, overall survival, confirmed complete or partial response, and disease control.

Enrollment and treatment assignment are distinct from the denominators used for each result. For the primary percentage, the registry lists 61 participants in the chemotherapy group and 63 in the erlotinib group. The registry’s original statistical design assumptions, including any formal sample-size justification, are not reported in the registry record. The planning illustrations accompanying the results are based on observed effects and are not evidence of the original design rationale.

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: FAS

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: Full analysis set (FAS): all randomized participants.

6. Primary Endpoint

The registry lists the primary outcome as Percentage of Participants With Disease Progression or Death, measured over Screening, end of every 2nd cycle through Cycle 8 (21-day cycles), and every 12 weeks thereafter until the end of study and final analysis up to a maximum of 42 months..

The primary outcome is titled “Percentage of Participants With Disease Progression or Death.” The accompanying definition describes progression-free survival as the time from randomization to disease progression or death from any cause. Progression was assessed according to RECIST version 1.0. For target lesions, progression required at least a 20% increase in the sum of the longest diameters, using the smallest recorded sum since treatment began as the reference. For non-target lesions, progression meant unequivocal worsening of existing lesions.

The registry reports the primary outcome as a percentage rather than as a median time or a hazard ratio. That summary communicates how many participants in each group had a recorded progression-or-death event, but does not convey when the events occurred. The time-to-event definition remains relevant because it specifies the event and the follow-up rules: participants were censored at the last post-baseline tumor assessment documenting no progression, or at randomization if no post-baseline assessment was available.

Progression-free survival is used in trials as a way to describe the time until worsening disease or death, combining these clinically important events into one endpoint. General statistical context: time-to-event methods can account for different follow-up durations and censoring, while a simple percentage does not show the event timing. The posted primary result is therefore best interpreted as the registry’s event-percentage comparison, alongside the endpoint definition and assessment schedule, rather than as a complete account of progression-free survival over time.

7. Results Posted in the Registry

• Bevacizumab + Erlotinib: 76.2% (48/63) versus 72.1% (44/61) on control; risk difference 4.1 percentage points (95% CI -11.3 to 19.5), risk ratio 1.06 (0.86–1.30); unadjusted two-proportion z-test p = 0.606.

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 Who Died

• 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:
  • Bevacizumab + Erlotinib: 52.4% (33/63) versus 45.9% (28/61) on control; risk difference 6.5 percentage points (95% CI -11.1 to 24.0), risk ratio 1.14 (0.80–1.64); unadjusted two-proportion z-test p = 0.471.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Bevacizumab + Erlotinib and the control. The interval gives the range of effects that remain compatible with the data.

OS

• 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:
  • Bevacizumab + Chemotherapy vs Bevacizumab + Erlotinib: hazard ratio 1.24 (95% CI 0.75–2.05), reported p = 0.406 (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 Bevacizumab + Chemotherapy vs Bevacizumab + Erlotinib and the control. The interval gives the range of effects that remain compatible with the data.

Percentage of Participants With a Best Overall Response (BOR) of Confirmed Complete Response (CR) or Partial Response (PR) According to RECIST V 1.0

• 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:
  • Bevacizumab + Erlotinib: 27.0% (17/63) versus 44.3% (27/61) on control; risk difference -17.3 percentage points (95% CI -33.9 to -0.7), risk ratio 0.61 (0.37–1.00); unadjusted two-proportion z-test p = 0.044.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the proportion with the event is statistically lower for Bevacizumab + Erlotinib than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

The secondary outcomes provide additional views of the comparison, but they do not all measure the same thing. Death as a percentage is an event proportion, overall survival is reported as a time-to-event analysis, and confirmed response is based on tumor response categories. Their estimates therefore answer related but distinct questions. The primary outcome concerns progression or death, while overall survival focuses on death over follow-up and response describes tumor shrinkage meeting specified criteria.

Across the posted results, the secondary outcomes do not provide a single, consistent account that settles the primary comparison. The overall-survival comparison is uncertain, and the response comparison has a different observed pattern from the primary event-percentage result. The registry lists disease control as an outcome, but its result is not reported in the registry record. These differences are not inherently contradictory: outcome definitions, follow-up, and event timing vary, so results can differ without one measure invalidating another.

General statistical context: secondary outcomes are often useful for understanding a trial’s broader pattern, but they usually carry less confirmatory weight than a prespecified primary outcome. When several outcomes are tested, an apparently notable result can arise by chance, particularly when no multiplicity procedure or testing hierarchy is reported. The registry record does not report such an adjustment or hierarchy. The secondary p-values should therefore be treated as unadjusted comparisons, interpreted alongside their estimates and intervals rather than as standalone proof.

The proper reading is cumulative but cautious: consider the primary result first, then use the secondary findings to describe other observed aspects of the trial. A secondary endpoint may add context, but it does not erase uncertainty in the primary comparison. Nor does a single secondary result establish an overall benefit or harm across outcomes. The registry’s incomplete reporting of disease control and the missing chemotherapy median for overall survival further limit how completely the secondary outcome pattern can be characterized.

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

• Percentage of Participants Who Died
• OS
• Percentage of Participants With a Best Overall Response (BOR) of Confirmed Complete Response (CR) or Partial Response (PR) According to RECIST V 1.0
• Percentage of Participants With Disease Control According to RECIST V 1.0

9. Statistical Interpretation

10. Statistical Methods Explained

The study used randomized allocation to assign participants to one of two parallel treatment groups. Randomization is intended, in general, to reduce systematic differences between groups at the start of a trial. Open-label treatment means assignments were not masked. That feature can matter for outcomes involving clinical assessments or decisions, although progression was defined using RECIST version 1.0 and the registry gives rules for censoring.

What does the primary percentage compare?

The primary outcome is a binary event summary: whether progression or death was recorded for a participant. The posted group summaries are percentages, and the comparison is given as a risk difference. A risk difference describes the difference between the event proportions in the two groups, with the direction determined by the group ordering. The posted estimate is 0.041. Its interval, from -0.113 to 0.195, crosses zero, indicating that the data do not rule out no difference under the interval’s assumptions. The unadjusted p-value is 0.606. A p-value describes how compatible the observed data are with a specified no-difference model; it is not the probability that either treatment is better.

How does the endpoint definition affect interpretation?

The endpoint description defines progression-free survival as time from randomization until progression or death, but the primary results are expressed as event percentages. These are not interchangeable summaries. A percentage omits the timing of events, whereas a time-to-event analysis uses event times and can accommodate censoring. The registry’s progression rules specify at least a 20% increase in the sum of longest diameters for target lesions, or unequivocal progression of existing non-target lesions. Participants without documented progression were censored at the last post-baseline tumor assessment documenting no progression; if there was no post-baseline assessment, censoring was at randomization.

General statistical context: censoring is a way to retain information from participants whose event status is not observed throughout the full study period. Time-to-event methods rely on assumptions about censoring and follow-up. The registry’s stated censoring rules clarify how observations were handled, but the primary result does not include a posted time-to-event estimate. Consequently, readers should not treat the primary percentages as a full description of progression-free survival.

What do the secondary comparisons add?

The registry reports an overall-survival hazard ratio from a Log Rank analysis, with a 95% confidence interval and p-value. A hazard ratio compares event rates over follow-up under the time-to-event model; it is not the same as a ratio of percentages or a difference between medians. The reported estimate is 1.24, with a 95% confidence interval from 0.75 to 2.05 and p = 0.406. The interval includes 1, the no-relative-difference value for a hazard ratio. The chemotherapy-group median is not reported in the registry record, limiting a direct comparison of the posted medians.

The response endpoint is reported as percentages with 95% confidence intervals, and the sponsor’s reported comparison uses a Chi-squared method for the difference in response rates. The result gives a difference of -17.28, with a 95% confidence interval from -34.8 to 0.3 and p = 0.044. The registry also includes an unadjusted calculated risk-difference comparison. Such an unadjusted result should be distinguished from the stated sponsor analysis, which is the primary source for that reported comparison.

What does an interval communicate?

A confidence interval describes a range of effect values compatible with the data and the method’s assumptions. Its width reflects statistical precision: wider intervals leave more uncertainty about the size and direction of an effect. An interval that contains the no-difference value does not prove that the groups are identical. It means the data do not exclude that value at the interval’s confidence level. Conversely, an interval that does not contain that value does not by itself address the importance, reliability, or clinical meaning of the observed effect.

General statistical context: analyses of several outcomes raise multiplicity concerns. When many outcomes are examined, the chance of at least one seemingly notable result can increase, especially if results are interpreted without adjustment or a prespecified hierarchy. The registry record does not report a multiplicity adjustment or a hierarchy governing the secondary outcomes. Therefore, secondary p-values should be read as unadjusted evidence for their individual comparisons, not as a complete family-wise assessment.

The registry identifies the primary analysis population as FAS. Details of the FAS definition, the analysis population for each secondary outcome, missing-outcome handling beyond the stated censoring rule, and the statistical plan are not reported in the registry record. Those omissions matter because analysis populations and missing-data methods can affect estimates and their interpretation. The results are best read as the posted analyses, with the population and methods details that are actually stated kept distinct from unreported protocol information.

11. Confidence Intervals and Precision

The primary comparison has a risk-difference estimate of 0.041 and an interval from -0.113 to 0.195. The interval spans zero and includes effects in either direction. This makes the size and direction of the primary contrast uncertain: the observed estimate is not a precise basis for concluding that one group had a meaningfully different event percentage. The registry does not state the confidence level for this particular interval, so it should not be assigned a confidence level that is not given in the record.

The overall-survival hazard ratio is 1.24, with a 95% confidence interval from 0.75 to 2.05. That interval includes 1 and covers a broad range of relative effects. The reported p-value is 0.406. The chemotherapy-group median is not reported in the registry record, while the erlotinib-group median is 16.4 months with a lower confidence limit of 11.0 months. For the chemotherapy group, a lower confidence limit of 12.7 months is listed without a median value. These entries do not support a complete comparison of the two medians.

For confirmed response, group-level 95% confidence intervals are 31.5 to 57.6 for chemotherapy and 16.6 to 39.7 for erlotinib. The reported difference in response rates has a 95% confidence interval from -34.8 to 0.3. Such intervals make the uncertainty around the estimates visible; they are not guarantees that the true effect lies within the displayed range. General statistical context: greater precision usually requires more information, such as more participants or events, while the achieved precision also depends on outcome frequency and variability. The posted intervals should be considered alongside the trial’s group sizes and the fact that several outcomes were examined.

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 are based on the observed effects, not on the trial’s original design. For the binary outcomes, the entries are participants per arm; for the time-to-event outcome, they are total events. The stated assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. Each illustration considers an effect equal to the observed effect, 75% of the observed effect, or 50% of the observed effect. The trial’s original sample-size assumptions and justification are not reported in the registry record.

The 80% power column describes the number of participants per arm, or total events for the time-to-event outcome, that the planning calculation estimates would be needed to have an 80% chance of detecting the specified effect under its assumptions. The 90% power column uses the same idea but targets a 90% chance. These are planning targets, not probabilities that the trial’s findings are correct. This is not observed power.

Reading across the effect assumptions shows why smaller effects require more information: the calculations call for larger group sizes or event totals when the target is 75% or 50% of the observed effect. For time-to-event outcomes, event totals are the relevant planning quantity rather than participants per arm. The calculations should not be interpreted as a retrospective verdict on the trial. They illustrate what a future study might require under the stated assumptions, which may differ from the original protocol’s assumptions and from the conditions of a future study.

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 = 0.721; treatment proportion p₁ = 0.762; average p̄ = 0.742.
2. First term: z_1−α/2 × √[2p̄(1−p̄)] = 1.96 × 0.6191 = 1.2134.
3. Second term: z_1−β × √[p₀(1−p₀) + p₁(1−p₁)] = 0.842 × 0.6184 = 0.5205.
4. n = (1.2134 + 0.5205)² / (0.0406)² = 1,824.3, rounded up to 1,825 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 <- 0.721311475409836                          # control proportion
p1 <- 0.761904761904762                          # 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(531960)
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 = 1,825 (80% power), 2,442 (90% power)
 75% of observed effect: n per arm = 3,286 (80% power), 4,398 (90% power)
 50% of observed effect: n per arm = 7,484 (80% power), 10,019 (90% power)
power.prop.test: 1824.3 
Simulated power at n = 1,825 per arm: 0.798 (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

The registry lists multiple outcomes, including the primary progression-or-death outcome and secondary outcomes for death, overall survival, response, and disease control. It reports several comparisons and p-values. The multiplicity strategy—such as an adjustment procedure, a testing hierarchy, or a rule for interpreting the family of secondary outcomes—is not reported in the registry record. As a result, the secondary p-values should be read as unadjusted evidence for their individual comparisons, not as a claim that the overall chance of a false-positive conclusion has been controlled across all outcomes.

The primary population is identified as FAS. The registry does not provide the full definition of that population or state, for every endpoint, exactly which participants were included in each analysis. The displayed denominators differ by outcome: the primary and response summaries use 61 and 63 participants, while the death percentages also use those group sizes; the overall-survival analysis reports event counts of 28 and 33. These are outcome-specific analysis details and should not be assumed to represent identical information for every endpoint.

The primary endpoint description gives a censoring rule for tumor assessments: participants were censored at the last post-baseline assessment documenting no progression, or at randomization when no post-baseline assessment was available. Other missing-data procedures, including any imputation approach or sensitivity analysis, are not reported in the registry record. The handling of missing response assessments and missing overall-survival information is also not reported in the registry record.

These reporting gaps do not demonstrate that methods were absent from the trial protocol; they mean only that those details cannot be established from the registry record. They limit evaluation of whether analyses followed a prespecified plan and how sensitive estimates might be to missing information or multiple testing. The registry reports the overall-survival method as Log Rank and the reported response comparison as Chi-squared, but the broader statistical analysis plan is not reported in the registry record.

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 0% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: pneumonia (5.0%), pulmonary embolism (5.0%), neutropenia (3.3%), anaemia (3.3%), febrile neutropenia (3.3%).

The serious adverse event summaries use groups of 60 participants at risk in the chemotherapy group and 63 in the erlotinib group. Serious adverse events affected 22 participants in the chemotherapy group and 17 in the erlotinib group. Other adverse events affected 58 and 60 participants, respectively, using the same at-risk group sizes. These are participant counts, not a comparison adjusted for follow-up time or differences in event duration.

The registry lists pneumonia and pulmonary embolism at 5.0% each among the top serious-event terms. Neutropenia, anaemia, febrile neutropenia, and thrombocytopenia are each listed at 3.3%. These entries describe reported serious-event terms; they do not provide a complete account of severity, timing, recurrence, or causal attribution. The registry does not report deaths in the adverse-event summary.

General statistical context: adverse-event counts are descriptive and should be read with the number at risk and the way events were collected. They do not, on their own, establish that one regimen caused more or fewer harms. A full safety comparison would also depend on event definitions, exposure duration, follow-up, and methods for attributing events, details that are not all reported in the registry record.

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

The results describe what was recorded in the randomized groups and how uncertain the between-group estimates remain. On the primary outcome, the observed progression-or-death percentages were 72.1% with bevacizumab plus chemotherapy and 76.2% with bevacizumab plus erlotinib. The risk-difference interval spans zero and a range of values in both directions. The comparison therefore does not establish a clear difference between the groups, nor does it prove that the treatments are equivalent.

The registry’s endpoint description defines progression-free survival as time to progression or death, but the posted primary result is a percentage. It does not provide a primary time-to-event estimate. The separate overall-survival analysis also has substantial uncertainty, with a 95% confidence interval for the hazard ratio from 0.75 to 2.05. The response outcome differs in its observed pattern, but it is secondary and the record does not report a multiplicity strategy. These findings should not be combined into a broad claim of overall superiority.

General statistical context: failure to establish a difference is not the same as evidence that the treatments have identical effects. Likewise, a result that appears notable on one secondary outcome does not by itself establish a reliable overall treatment effect. Confidence intervals help show the range of effects still compatible with the data, while endpoint definitions determine what each estimate can answer.

The findings also should not be generalized beyond the study’s population and design without additional evidence. This was a Phase 2, open-label study in advanced non-small cell lung cancer, with 124 participants. The record does not provide all protocol details, the original sample-size justification, or several analysis procedures. Accordingly, the posted estimates are best treated as trial-specific evidence with identifiable uncertainty and reporting limitations, not as a complete answer to every question about the treatment strategies.

17. Why This Trial Matters Statistically

This trial is statistically informative because it illustrates the distinction between an endpoint’s definition and the form of its posted result. The primary endpoint is described using a time-to-event concept—progression-free survival—yet the reported primary comparison is a percentage of participants with progression or death. Those summaries convey different information. A percentage reports event occurrence in the analyzed groups, while a time-to-event estimate would also characterize when events occurred and how censoring contributed.

The record also shows why estimates should be read with their uncertainty, not only by comparing point estimates or p-values. The primary risk-difference interval includes zero and spans effects in either direction. The overall-survival hazard-ratio interval is broad, and the chemotherapy median is not reported. These features make the limits of the evidence visible: a trial can produce numerical contrasts without resolving the size or direction of a treatment difference.

Finally, multiple endpoints require careful interpretation. The secondary outcomes describe distinct aspects of disease and survival, and one secondary comparison should not override the primary result without regard to its place in the analysis plan. The registry does not report a multiplicity procedure or a hierarchy. Together, endpoint form, uncertainty, and incomplete methodological detail provide a useful example of why statistical interpretation depends on more than a single estimate or threshold.

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 specific to the registry record shape interpretation. The primary endpoint is described as progression-free survival but reported as a percentage of participants with progression or death. A primary hazard ratio, median progression-free survival, or other time-to-event estimate is not reported in the registry record. The posted percentage therefore does not describe the timing of events, and the primary comparison’s confidence level is not stated alongside its interval.

The trial was open-label. Progression was defined with RECIST version 1.0, but the registry record does not report whether progression assessments were independently reviewed or blinded. The exact platinum-based chemotherapy regimen is not specified in the arm description. The overall-survival median for the chemotherapy group is not reported, although a lower confidence limit is listed. The disease-control outcome is named, but its result is not reported in the registry record.

The registry identifies the primary population as FAS but does not give its full definition. Missing-data methods beyond the primary censoring rules, the original sample-size justification, interim analyses, and the broader statistical analysis plan are not reported in the registry record. Nor does the record report a multiplicity adjustment or hierarchy for secondary outcomes. These gaps restrict assessment of prespecification, missing-data sensitivity, and the strength of secondary findings.

The adverse-event summaries are counts and listed terms; they do not provide a full account of exposure-adjusted risk, event duration, or attribution. These limitations do not invalidate the posted results, but they narrow the questions that the registry record can answer and call for restrained interpretation.

19. Related Tutorials

20. Related Statistical Calculators

21. Sources

• ClinicalTrials.gov: A Study of Tarceva (Erlotinib) in Combination With Avastin (Bevacizumab) in Patients With Advanced Non-Small Cell Lung Cancer. — NCT00531960.
• Registry record last updated 2014-11-11; status last verified 2014-11.
• Results first posted 2014-11-11.
• 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

NCT00531960 is a completed, randomized, open-label Phase 2, parallel-group study sponsored by Hoffmann-La Roche. It enrolled 124 participants with advanced non-small cell lung cancer to compare bevacizumab plus erlotinib with bevacizumab plus standard platinum-based chemotherapy. The primary outcome was the percentage of participants with disease progression or death; the registry defines progression-free survival using progression or death from randomization and RECIST version 1.0.

The posted primary event percentages were 72.1% for chemotherapy and 76.2% for erlotinib. The risk-difference interval crosses zero, leaving uncertainty about the direction and size of the contrast. Secondary results include death, overall survival, and confirmed response, but the disease-control result is not reported in the registry record. Important design and analysis details—including the original sample-size justification, multiplicity procedures, and some missing-data methods—are not reported in the registry record.