Erlotinib in Women With Squamous Cell Carcinoma of the Vulvar: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Erlotinib in Women With Squamous Cell Carcinoma of the Vulvar (NCT00476476), a na phase 2 study. The primary outcome is Response Rate.

2. Clinical Question

3. Trial Design

This was a phase II, single-group study of erlotinib in women with squamous cell carcinoma of the vulva. The registry record lists Dana-Farber Cancer Institute as the sponsor, enrollment of 41, study start in 2006-12, primary completion in 2012-10 and status as completed. These details describe the registry record; they do not establish how many participants contributed to the primary endpoint analysis.

The study included two clinically distinct cohorts. Cohort 1 comprised women with locally advanced measurable vulvar cancer, either primary or recurrent, who were candidates for definitive surgery or chemoradiation. They received erlotinib for at least 28 days and no more than 42 days before that definitive treatment. Cohort 2 comprised women with radiographically measurable distant metastatic cancer, at presentation or recurrence. They continued erlotinib in 28-day cycles until disease progression, unacceptable toxicity or withdrawal. The number enrolled in each cohort is not reported in the registry record.

Because this was a single-group study, the design does not support a like-for-like comparison with an untreated or alternative-treatment group. Allocation is listed as not applicable, and masking as none. The record does not report random assignment, blinding, or a control arm. Differences between the two cohorts also matter: their disease settings, treatment paths and response assessment schedules were not identical.

The stated primary endpoint was response rate, assessed before definitive surgery or chemoradiation in cohort 1 and after 2 cycles in cohort 2. The registry record does not report the cohort-specific response results, the total number included in the response analysis, or whether any participants were excluded from that analysis. It also does not report the study protocol’s statistical plan or sample-size justification.

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: The analysis dataset is comprised of response evaluable patients.

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: The analysis population is comprised of all enrolled patients.

6. Primary Endpoint

The registry lists the primary outcome as Response Rate, measured over Assessed prior to definitive surgery or chemoradiation therapy (cohort 1 pts) or after 2 cycles of therapy (cohort 2 pts)..

The primary endpoint, response rate, measures the proportion of response-evaluable participants who achieved either a complete response or a partial response. A complete response required resolution of all identified vulvar tumor masses in both cohorts, or disappearance of all target and non-target lesions with no new lesions, documented in two disease assessments at least 4 weeks apart. Partial response used different criteria across cohorts. In cohort 1 it was defined as a 30% reduction in the product of all diameters of the vulvar tumor or tumors compared with baseline. In cohort 2 it was defined using RECIST 1.1 as at least a 30% decrease in the sum of the longest diameters of target measurable lesions.

Response rate is a binary endpoint: each response-evaluable participant is classified as having met the response definition or not. That classification can be summarized as a proportion, with a confidence interval expressing uncertainty around the estimate. It is useful for describing whether measurable tumor reduction or disappearance was observed during the specified assessment period. It does not, on its own, measure how long a response lasts, whether disease control persists, or how outcomes compare with another treatment.

The assessment schedule differed by cohort, reflecting the distinct treatment paths. The response count, proportion and 90% confidence interval are not reported in the registry record. The number of participants who were response-evaluable is also not reported.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

What does a response proportion represent?

For a binary endpoint such as response, participants are classified according to whether they meet the defined outcome. The response proportion summarizes the number classified as responders relative to the response-evaluable analysis population. It is a description of the observed participants under the endpoint rules, not a measure of the size or duration of every individual response. The endpoint definition is particularly important because classifications depend on the imaging or clinical criteria and assessment schedule.

In this study, complete response and partial response were combined for the primary response-rate endpoint. Complete response involved resolution or disappearance of disease, with the applicable definitions requiring absence of new lesions and, for the stated lesion assessment, confirmation on two assessments at least 4 weeks apart. Partial response was defined differently across cohorts: a reduction in the product of tumor diameters for cohort 1, and a reduction in the sum of target-lesion longest diameters under RECIST 1.1 for cohort 2. These rules mean that a response estimate would reflect both participant outcomes and the operational definitions used to classify them.

How should a confidence interval be read?

A confidence interval describes uncertainty in an estimated quantity under a specified statistical procedure. For a response proportion, it gives a range around the observed estimate rather than a guarantee that the underlying value lies within that particular interval. The width depends on factors including the amount of information in the analysis and the method used to construct the interval. A relatively imprecise estimate has a wider interval; a more precise estimate has a narrower one. The primary endpoint lists a 90% confidence interval, but its limits and the estimate itself are not reported in the registry record.

A confidence interval is not a substitute for the underlying denominator. The same proportion can have different precision depending on the number of evaluable participants. The registry record states that the analysis dataset comprised response-evaluable patients, but does not report that population’s size. Enrollment of 41 is a study-level figure and cannot be assumed to be the endpoint denominator.

Why does the single-group design matter?

A single-group study describes outcomes among participants receiving the study intervention. Without a concurrent comparator, it cannot directly estimate the difference between treatment and control outcomes. Changes over time or responses observed during treatment may be relevant descriptions, but the design does not isolate treatment effects from other possible influences. That limitation is especially important when interpreting an outcome as evidence of comparative benefit. No randomized comparison or like-for-like control arm is reported in the registry record.

The cohorts also differ in disease setting and assessment timing. Cohort 1 received a short course before definitive surgery or chemoradiation; cohort 2 continued treatment in cycles and had the stated response assessment after 2 cycles. A pooled estimate could mask differences between these settings. Cohort-specific response numbers and any method for combining them are not reported in the registry record.

What do the planning illustrations mean?

Any sample-size illustrations associated with the posted analysis are planning calculations, not observed study outcomes. The stated illustration assumptions use a two-sided significance level of 0.05, equal allocation between two arms and a normal approximation, with the effect to detect defined as a stated fraction of the observed effect. Such calculations address the hypothetical number of participants per arm under those assumptions. They do not describe the original single-group study design and must not be interpreted as observed power or as the study’s own sample-size justification.

For a planning calculation based on an observed effect, the result depends on that effect estimate and the assumptions used. The registry record does not report the primary response estimate, and it does not report the original protocol’s sample-size justification or statistical plan. Planning values therefore cannot fill the gap in the missing endpoint results or establish the study’s precision.

What can be inferred from the posted statistical information?

The record gives the endpoint type, its assessment schedule, its definition and the stated confidence-interval level. It does not give the primary estimate, interval limits, analysis denominator or number of analyses. No inferential result or treatment comparison is reported. Accordingly, statistical interpretation is limited to explaining how the endpoint was intended to be defined and what information would be needed to assess its magnitude and uncertainty. General statistical concepts describe how such estimates are read; they do not supply missing trial results.

10. Exact Estimates for Each Arm

No two-arm comparison is shown for this trial, but each arm's result can still be summarized with exact interval estimates computed from the posted counts. The Clopper-Pearson interval is exact and conservative; the Wilson interval is a common closer-to-nominal alternative; the Bayesian credible interval uses a uniform Beta(1, 1) prior. For small groups the three can differ noticeably.

R code and output

arm    <- c("Erlotinib-Cohort 1", "Erlotinib-Cohort 2")
n      <- c(17, 23)
events <- c(0, 0)
for (i in seq_along(n)) {
  ci <- binom.test(events[i], n[i])$conf.int          # Clopper-Pearson exact interval
  cr <- qbeta(c(0.025, 0.975), 1 + events[i], 1 + n[i] - events[i])   # Bayesian, uniform Beta(1,1) prior
  wi <- (function(x, n) { p <- x / n; z <- qnorm(0.975); d <- 1 + z^2 / n
    c((p + z^2 / (2 * n) - z * sqrt(p * (1 - p) / n + z^2 / (4 * n^2))) / d, (p + z^2 / (2 * n) + z * sqrt(p * (1 - p) / n + z^2 / (4 * n^2))) / d) })(events[i], n[i])
  cat(sprintf("ARM|%s|%d|%d|%.1f|%.1f|%.1f|%.1f|%.1f|%.1f|%.1f\n", arm[i], n[i], events[i], 100 * events[i] / n[i],
              100 * ci[1], 100 * ci[2], 100 * wi[1], 100 * wi[2], 100 * cr[1], 100 * cr[2]))
}

ARM  Erlotinib-Cohort 1  17  0  0.0  0.0  19.5  0.0  18.4  0.1  18.5
ARM  Erlotinib-Cohort 2  23  0  0.0  0.0  14.8  0.0  14.3  0.1  14.2

11. Multiplicity, Missing Data and Analysis Populations

The registry record lists one outcome measure and no analyses. That information does not establish whether the study examined additional outcomes, conducted interim analyses or made multiple statistical comparisons; those details are not reported in the registry record. Without a reported statistical plan, it is not possible to assess whether multiplicity adjustments were planned or used, or how any decisions about analyses were made.

The primary analysis population is described as response-evaluable patients. The record does not report how response-evaluable status was determined, the number of participants meeting that definition, or how participants without an assessment were handled. It does not report whether missing endpoint data were excluded, imputed or otherwise addressed. Those choices can affect a response proportion because the analysis denominator and outcome classifications depend on the handling of missing assessments.

Enrollment was 41, but that figure does not establish the number analyzed for response. The number in each cohort and the number assessed at the stated time points are not reported in the registry record. Nor does the record state whether the endpoint analysis followed an intention-to-treat approach or another analysis-population rule. Such details should not be inferred from the term “response-evaluable.”

The registry record also does not report a statistical analysis plan, prespecified subgroup analyses, or procedures for handling multiple endpoints or comparisons. These gaps limit assessment of the risk that analytical choices or missing observations influenced any result. Because the response estimate and its denominator are absent, the effect of those choices on the primary endpoint cannot be evaluated from the record.

12. 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: diarrhea w/o prior colostomy (17.1%), nausea (7.3%), dehydration (7.3%), hypokalemia (7.3%), renal failure (7.3%).

The registry record lists 41 participants at risk for both serious and other adverse events. It reports 19 participants affected by serious adverse events and 40 affected by other adverse events. These are separate reported counts; they should not be added, since the record does not state whether the groups overlap. The record does not provide a death count in the adverse-event summary.

The listed serious-event terms include diarrhea without prior colostomy at 17.1%, nausea at 7.3%, dehydration at 7.3%, hypokalemia at 7.3%, renal failure at 7.3%, and death due to disease progression, not otherwise specified, at 4.9%. These percentages describe the registry’s reported terms. They should not be interpreted as treatment-attributable risks: the record does not report whether each event was judged related to erlotinib, how event terms were coded, or whether participants could appear under multiple terms.

Serious adverse events and other adverse events are different reporting categories. The counts do not, by themselves, describe severity across all events, duration, timing, or clinical course. The registry record does not report a comparator group or an adjusted safety analysis. These figures therefore describe reported events among the study population and cannot establish comparative safety or causation.

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

The study’s primary question was whether participants in two defined cohorts met the complete- or partial-response criteria while receiving erlotinib. The endpoint and assessment schedule are described in the registry record, but the response count, response proportion and 90% confidence interval are not. Consequently, the registry entries do not establish the observed frequency of response or how precisely it was estimated.

Even if a response result were available, its meaning would be bounded by this study’s design. A single-group study can describe outcomes in its treated population but cannot, without a comparator, quantify how those outcomes differ from another treatment or from no study treatment. The design also does not separate treatment effects from other influences on observed outcomes. No randomized or concurrent control comparison is reported.

The two cohorts represent different disease settings and treatment paths. Their response assessments occurred at different points, and the partial-response criteria were not identical. A combined result would require careful interpretation in light of those distinctions. Cohort-specific estimates and the method for any combined analysis are not reported in the registry record.

The study is listed as completed, with enrollment of 41, but neither fact supplies the primary result or the endpoint analysis denominator. It is not possible to infer how many participants were response-evaluable or how missing assessments affected the result. The safety figures offer descriptive event information, not proof that the intervention caused the events or a comparison with another treatment.

Thus, the record establishes the design, intervention, endpoint definitions and selected safety summaries. It does not provide enough primary outcome information to determine the magnitude or precision of response in the study. Nor does it support a comparative treatment conclusion. Those boundaries distinguish what the registry record describes from claims that would require additional outcome data and a suitable comparison.

14. Why This Trial Matters Statistically

This study illustrates why an endpoint definition and an endpoint result are both necessary for statistical interpretation. The record defines response using complete- and partial-response criteria, identifies distinct rules for the two cohorts and specifies when assessment was intended. Yet without the response count, analysis denominator and confidence interval, the central efficacy estimate cannot be evaluated. A well-described endpoint does not replace its observed result.

It also highlights the interpretive challenge of combining clinically distinct cohorts. Locally advanced disease treated before definitive surgery or chemoradiation differs from distant metastatic disease treated in cycles. Assessment timing and partial-response definitions differed as well. Any overall estimate would need context about the cohort mix and the analysis approach, details that are not reported in the registry record.

The single-group design makes the distinction between description and comparison especially important. Observed responses and adverse events can describe participants in the study, but without a control group they do not provide a direct estimate of how outcomes differ under another treatment. The design and the available outcome reporting must therefore be considered together.

Finally, the record shows why enrollment and analysis population should not be conflated. Enrollment was 41, while the primary analysis was defined in terms of response-evaluable patients. The analysis denominator is not reported. Without it, readers cannot judge how much information supports the endpoint estimate or how excluded and missing assessments may matter. These are general principles for reading clinical-trial statistics, made particularly visible by the information gaps in this record.

15. 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.

The most immediate limitation is that the registry record does not report the primary response result. Neither the number of complete or partial responses, the response proportion, nor the 90% confidence interval is available in the registry record for this analysis. The response-evaluable population is named, but its size is not reported. This prevents assessment of the endpoint’s magnitude, precision and denominator.

The study had a single-group design, with no reported control arm. Consequently, the record cannot support a direct comparison against another treatment or an estimate of a comparative treatment effect. Allocation is listed as not applicable and masking as none. The record does not describe randomization or blinding.

The two cohorts differed in disease setting, treatment course and assessment timing. Their partial-response definitions also differed. The number enrolled or analyzed in each cohort, cohort-specific results and the approach to any pooled estimate are not reported in the registry record. These omissions make it difficult to understand how an overall response summary would represent the distinct groups.

The record does not report how missing or unevaluable response assessments were handled, whether all enrolled participants contributed to the primary analysis, or the analysis-population rules beyond the phrase response-evaluable patients. It also does not report the protocol’s statistical plan, sample-size justification, interim analyses or multiplicity procedures.

The safety summary does not provide a comparator, event attribution, detailed timing or a death count. Its listed terms and percentages should therefore be read descriptively. Taken together, these trial-specific limitations restrict conclusions to the stated study design, endpoint definitions and reported safety summaries; the missing primary outcome prevents a numerical assessment of response.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Erlotinib in Women With Squamous Cell Carcinoma of the Vulvar — NCT00476476.
• Registry record last updated 2015-05-07; status last verified 2015-03.
• Results first posted 2015-03-19.
• 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.

19. Record Summary

NCT00476476 is listed as a completed phase II, single-group study sponsored by Dana-Farber Cancer Institute. It enrolled 41 women in a study of oral erlotinib at 150 mg/day for squamous cell carcinoma of the vulva. The record describes two cohorts: locally advanced measurable disease before definitive treatment, and radiographically measurable distant metastatic disease treated in cycles.

The primary endpoint was response rate among response-evaluable patients, with cohort-specific assessment schedules and partial-response definitions. The response count, proportion, confidence interval and analysis denominator are not reported in the registry record. The record lists serious adverse events affecting 19 participants and other adverse events affecting 40 participants, with selected serious-event terms and percentages. No control group is reported, so the design does not provide a direct treatment comparison.