Atezolizumab and Bevacizumab in Treating Patients With Recurrent, Persistent, or Metastatic Cervical Cancer: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Atezolizumab and Bevacizumab in Treating Patients With Recurrent, Persistent, or Metastatic Cervical Cancer (NCT02921269), a na phase 2 study. The primary outcome is Number of Participants With Objective Response Rate (ORR, Either Partial or Complete Response) Defined by Response Evaluation Criteria in Solid Tumors (RECIST) Version 1.1 Criteria.

2. Clinical Question

3. Trial Design

This was a phase 2, single-group study sponsored by the National Cancer Institute. It is listed as completed, with an enrollment of 11. The study began on 2017-03-10, and its primary completion date is recorded as 2020-10-01. No masking was used. Allocation is listed as not applicable, consistent with a design that did not assign participants to competing treatment arms.

Participants received atezolizumab and bevacizumab together. Both were administered intravenously on day 1, and courses repeated every 21 days while there was no disease progression or unacceptable toxicity. The registry record lists laboratory biomarker analysis as an additional intervention. The listed secondary outcomes include progression-free survival, overall survival, adverse events, tumor and immune-cell PD-L1 expression, and intratumoral and peripheral T-cell receptor clonality.

A single-group design can describe what happened among participants receiving the regimen, but it cannot by itself separate treatment effects from the natural course of disease, participant characteristics, or other influences. Because there is no control group, any response or safety counts should be read as observations in this study population, not as a comparative estimate against another treatment. The record does not report a protocol-level explanation for the choice of design, a sample-size justification, or details of an analysis plan.

The enrollment count is small, and the posted record does not give the primary response count or an inferential analysis. Consequently, the design and endpoint can be described, but the size and precision of the primary result cannot be assessed from the recorded primary-outcome information. The record does not report whether all enrolled participants contributed to each outcome or how participants who stopped treatment were handled in each analysis.

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:

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.

6. Primary Endpoint

The registry lists the primary outcome as Number of Participants With Objective Response Rate (ORR, Either Partial or Complete Response) Defined by Response Evaluation Criteria in Solid Tumors (RECIST) Version 1.1 Criteria, measured over Up to 2 years.

The primary endpoint counts participants with an objective response: either a complete response or a partial response under RECIST version 1.1. A complete response is defined as disappearance of all lesions. A partial response requires at least a 30% decrease in the sum of diameters of target lesions, using the baseline sum as the reference. The timeframe is up to 2 years.

Objective response is used to describe tumor change according to predefined measurement criteria. The criteria make the response classification more standardized than an informal judgment of whether disease has improved. Combining complete and partial responses into an overall response measure gives a count of participants meeting either response category during the specified assessment period.

The endpoint is a participant count, not a direct measure of how long a response lasts, whether disease later progresses, or how participants feel or function. Progression-free survival and overall survival are listed as secondary outcomes, rather than the primary endpoint. The primary result would need the number of participants meeting the response criteria and a clearly defined analysis population to support interpretation. Although the registry record defines the endpoint, it does not report the response count, so the observed objective-response result cannot be described from the primary outcome entry.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

Statistical methods should match the design and the type of endpoint. This study is recorded as single-group, and its primary endpoint is a count of participants with an event, where the event is achieving an objective response. The registry record lists no analyses, so the method used to summarize or test the primary endpoint is not reported in the registry record. The concepts below are general explanations, not a description of unreported trial procedures.

What does a participant count mean?

For a binary endpoint, each participant is classified according to whether the defined event occurred. For this endpoint, the event is an objective response under RECIST version 1.1. A response count alone needs a denominator to be interpretable as a proportion. That denominator might be all enrolled participants or a defined analysis population, but the choice must be specified. The registry record gives enrollment as 11, but does not report the response count, the number assessed, or the analysis population. It is therefore not possible to state an observed response proportion.

General statistical explanation: a confidence interval around a binary response estimate communicates uncertainty arising from the number of participants and the number with the event. With a small study population, uncertainty can remain substantial even when a point estimate is available. A confidence interval is not itself a probability that a particular treatment effect is true; it is a way of expressing the precision of an estimate under the method’s assumptions. No primary confidence interval is reported in this registry record.

What can a single-group design establish?

A single-group study can record outcomes among participants receiving the study regimen. It does not include an internal control group that would show what happened under a different treatment or under a control condition. General statistical explanation: a comparison is most informative when groups are sufficiently comparable and differences in outcomes can be related to the assigned treatment. Without a comparator, an observed outcome cannot isolate the effect of treatment from other explanations, including the underlying disease course and differences among participants.

Allocation is listed as not applicable, masking as none, and the model as single group. Thus, the record does not describe random assignment or blinding. No randomized comparison should be inferred from this design. The listed secondary outcomes include time-to-event measures, but the registry record does not provide analyses for them. The record also does not specify the methods used to estimate survival outcomes, handle censoring, or define analysis populations for those outcomes.

How should response criteria be understood?

RECIST version 1.1 provides the response definitions stated in the registry record. Complete response means disappearance of all lesions. Partial response means at least a 30% decrease in the sum of diameters of target lesions, relative to the baseline sum. The primary endpoint combines these categories as overall response. General statistical explanation: predefined criteria can reduce variation in how outcomes are classified, but interpretation still depends on how often assessments occurred, how missing or unevaluable assessments were treated, and which participants were included. Those operational details are not reported in the registry record.

The outcome timeframe is up to 2 years. That timeframe indicates the stated period for the endpoint, but does not by itself explain the assessment schedule or the rules for assigning response. The registry record does not report those details. It also does not provide the number of assessments, the response count, or the timing of any observed responses.

What do planning calculations represent?

The record describes planning illustrations based on the observed effect, with a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation. It states that the effect to detect is a stated fraction of the observed effect. For time-to-event outcomes, the illustration concerns a total number of events; for other outcomes, it concerns participants per arm. These are planning assumptions, not a description of this trial’s design: the study itself is single-group. The registry record does not report the trial’s own sample-size justification. Planning illustrations are not observed power, and they do not show that the study had a particular probability of detecting an effect. The effect estimates and planning values are not reported in the record, so no numerical sample-size implication can be drawn.

General statistical explanation: a normal approximation is a method for estimating sampling behavior using a normal distribution. Its adequacy depends on the setting and the available data. For count outcomes, small numbers can make approximations less reliable; methods that directly model the event probability may be considered, depending on the analysis objective. No method choice or diagnostic assessment is reported for this trial. Over-dispersion, in general, refers to variability greater than a statistical model expects; it matters when a model’s assumed variability understates uncertainty. The registry record does not identify a model for which over-dispersion was assessed.

No interim analyses, multiplicity procedure, missing-data method, or analysis population is reported in the registry record. Without these details and without the primary response result, there is no basis for reconstructing an inferential analysis. General statistical principles can clarify what such methods would address, but they cannot fill in the trial’s unreported procedures or 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("Treatment (Atezolizumab, Bevacizumab)")
n      <- c(11)
events <- c(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  Treatment (Atezolizumab, Bevacizumab)  11  0  0.0  0.0  28.5  -0.0  25.9  0.2  26.5

11. Multiplicity, Missing Data and Analysis Populations

The registry record lists a primary objective-response endpoint and several secondary outcomes, including progression-free survival, overall survival, adverse events, PD-L1 expression, and T-cell receptor clonality. It reports 7 outcome measures and no analyses. The listed secondary outcomes create several possible questions, but the record does not report a multiplicity strategy that would explain how the chance of false-positive findings across multiple statistical tests was addressed. No adjustment method or prespecified testing sequence is reported in the registry record.

The primary endpoint is defined, but its participant count is absent. The record does not state the number of participants with an evaluable response, how unevaluable assessments were handled, or whether missing outcome data were excluded or classified in another way. The analysis population is also not reported in the registry record. These are important details because the response count and its denominator determine the endpoint estimate.

The registry record does not report interim analyses, a missing-data strategy, or details of analysis populations for the secondary outcomes. It also does not provide a statistical analysis plan or describe whether any analyses were adjusted for participant characteristics. These omissions mean that the posted information does not support an assessment of how missing data, multiple outcomes, or differences in analytic populations might affect the findings.

These are reporting limitations, not evidence that a particular method was or was not used outside the registry record. Without the methods and results, it is not possible to assess the impact of multiplicity or missingness on the study’s primary conclusion.

12. Safety Results

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

Deaths

Other (non-serious) adverse events

Participants with at least one non-serious event at or above the registry's 5% reporting threshold.

Most frequent non-serious adverse event terms

The registry safety group lists 11 participants at risk for serious adverse events and 0 affected. It lists 11 participants at risk for other adverse events and 11 affected. The group also records 11 deaths. These are counts in the registry record, and should be read as separate entries rather than as evidence about cause or treatment attribution.

The record does not list serious adverse-event terms for this group. It does not explain the definitions or observation periods underlying the safety counts, nor does it specify whether the deaths were due to disease, treatment, another cause, or an unknown cause. The fact that deaths are recorded does not establish that they were treatment-related. Likewise, the count for other adverse events does not describe their type, severity, timing, or relationship to study treatment.

Safety counts are most interpretable when accompanied by event definitions, exposure time, attribution rules, and detail on the events observed. Those details are not reported in the registry record. The figures therefore describe what the registry lists, but do not support a causal safety conclusion or a comparison with another treatment group.

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

This record describes a phase 2, single-group study of atezolizumab and bevacizumab in patients with recurrent, persistent, or metastatic cervical cancer. Its primary endpoint was the number of participants with an objective response by RECIST version 1.1, assessed up to 2 years. The registry record does not report the objective-response count, so it does not support a numerical statement about the primary outcome or its precision.

The enrollment of 11 provides context for the study’s scale, but does not reveal how many participants were evaluable or met the response criteria. There are no analyses listed. As a result, an estimate, confidence interval, or statistical test for objective response cannot be described. The absence of these primary results is not evidence that no responses occurred; it means the response result is not reported in the registry record.

The design also limits what could be concluded even if a response count were available. A single-group study has no concurrent comparator, so outcomes cannot be interpreted as a randomized treatment difference. The record does not establish what would have happened to similar participants under another treatment or control condition. The safety counts likewise do not identify causes or treatment relationships for recorded events or deaths.

The study therefore provides a defined clinical question and records some safety information, but the posted information is insufficient to evaluate the primary response finding. No conclusion that the regimen does or does not work follows from the missing response result. A fuller interpretation would require the primary outcome result, the analysis denominator, methods for handling missing assessments, and appropriate context for the safety events.

14. Why This Trial Matters Statistically

This trial is statistically notable for the gap between a clearly defined primary endpoint and the absence of its result in the registry record. Objective response is specified using RECIST version 1.1, and the assessment timeframe is stated as up to 2 years. Yet the registry does not report how many participants achieved a complete or partial response. Without that count and an analysis denominator, readers cannot assess even the basic magnitude of the primary outcome.

The single-group design is also central to interpretation. It can describe outcomes among participants receiving the regimen, but it does not provide a controlled estimate of a treatment difference. This distinction matters because a count observed in a treated group is not, on its own, evidence of what would have happened under an alternative. A comparator and a prespecified analytic approach would address different questions from the descriptive response count.

The enrollment of 11 underscores why the denominator and uncertainty are important. General statistical explanation: estimates based on small groups can be imprecise, and the outcome of individual participants may have an appreciable influence on the estimate. Here, the primary result and its uncertainty are not reported, so the degree of precision cannot be judged. The safety counts likewise need definitions and context before they can be interpreted as evidence about treatment-related risk.

Finally, the registry lists multiple outcomes but no analyses, and does not report a multiplicity plan or handling of missing outcomes. These omissions make it difficult to assess how robust a statistical conclusion would be. The trial’s statistical importance lies not in a result that can be inferred, but in the need to distinguish a documented study question from results and methods that are not reported.

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 primary limitation for interpretation is that the registry record defines objective response but does not report the number of participants who achieved it. There is also no reported response denominator, confidence interval, or analysis. The study enrolled 11 participants, but the record does not state whether every participant was evaluable for the primary endpoint or how any missing assessments were handled.

The single-group design has no concurrent control condition. It cannot provide a randomized comparison with another treatment, and no like-for-like control comparison should be inferred. No masking was used. The registry record does not provide a sample-size justification or the trial’s statistical analysis plan, and does not report interim analyses or multiplicity procedures.

The stated primary endpoint timeframe is up to 2 years, but assessment timing and response adjudication details are not reported in the registry record. The record lists secondary outcomes, but does not report analyses for them. The safety entries do not provide event terms for serious adverse events and do not explain the causes or treatment relationships of the recorded deaths. Their observation periods and event-attribution rules are not reported.

These limitations concern what can be established from the registry record. They should not be interpreted as proof that unreported methods or results do not exist elsewhere. However, without those details, the posted record does not permit a complete assessment of the primary endpoint, precision, missing data, or causal safety interpretation.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Atezolizumab and Bevacizumab in Treating Patients With Recurrent, Persistent, or Metastatic Cervical Cancer — NCT02921269.
• Registry record last updated 2023-02-21; status last verified 2023-01.
• Results first posted 2023-02-21.
• 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

NCT02921269 is a completed phase 2, single-group study sponsored by the National Cancer Institute. It enrolled 11 participants with recurrent, persistent, or metastatic cervical cancer. Participants received atezolizumab and bevacizumab intravenously on day 1, with courses repeating every 21 days in the absence of disease progression or unacceptable toxicity.

The primary endpoint was the number of participants with an objective response by RECIST version 1.1, assessed up to 2 years. The registry record describes the response criteria but does not report the primary response count or any analyses. Listed safety counts are 0 serious adverse events among 11 at risk, 11 other adverse events among 11 at risk, and 11 deaths; causes and treatment relationships are not reported. The single-group design does not provide a controlled comparison.