A Study of Atezolizumab in Participants With Locally Advanced or Metastatic Urothelial Bladder Cancer (Cohort 1): Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for A Study of Atezolizumab in Participants With Locally Advanced or Metastatic Urothelial Bladder Cancer (Cohort 1) (NCT02951767), a na phase 2 study. The primary outcome is Percentage of Participants With a Confirmed Objective Response of Complete Response (CR) or Partial Response (PR) as Assessed by the Independent Review Facility (IRF) According to RECIST v1.1.

2. Clinical Question

3. Trial Design

This was a Phase II, multicenter, single-arm study of atezolizumab in locally advanced or metastatic urothelial bladder cancer. Cohort 1 included participants who were treatment-naïve for advanced disease and ineligible for cisplatin-containing chemotherapy. The recorded enrollment was 119. The study status is completed; the registry record lists a start date of 2014-05-31 and a primary completion date of 2016-07-04.

Participants received atezolizumab 1200 mg intravenously on Day 1 of each 21-day cycle. In Cohort 1, treatment was planned to continue until progression by RECIST v1.1 or unmanageable toxicity. The primary endpoint was the percentage with a confirmed objective response, defined as complete response or partial response, assessed by an Independent Review Facility. Response assessment was scheduled every 9 weeks for the first 12 months and every 12 weeks thereafter, through the data cutoff date of 04 July 2016; the listed maximum follow-up was 23.52 months.

There was no masking and no allocation to comparison arms. Because every participant in this cohort received the study intervention, the study can describe observed outcomes in the enrolled group but cannot by itself distinguish treatment effects from other explanations for those outcomes. The registry record does not report the primary response result, an analysis plan, or the rationale used to determine the enrollment target.

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.

Before assignment: The analysis included data up to cutoff date 28 February 2023.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Cohort 1 objective response-evaluable population included intent-to-treat (ITT) participants who had measurable disease per RECIST v1.1 at baseline. Cohort 1 ITT population included all participants from Cohort 1 who received any amount of study drug.

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: Cohort 1 Safety Evaluable Population included all participants who received any amount of study drug.

6. Primary Endpoint

The registry lists the primary outcome as Percentage of Participants With a Confirmed Objective Response of Complete Response (CR) or Partial Response (PR) as Assessed by the Independent Review Facility (IRF) According to RECIST v1.1, measured over Baseline until confirmed disease progression or death, whichever occurred first (assessed at every 9 weeks for the first 12 months, thereafter every 12 weeks until data cutoff date 04 July 2016, up to maximum length of follow-up of 23.52 months).

The primary endpoint was the percentage of participants with a confirmed objective response: complete response or partial response, assessed by an Independent Review Facility according to RECIST v1.1. Complete response required disappearance of target and non-target lesions and, when applicable, normalization of tumor marker levels. Any pathological lymph nodes had to decrease in short axis to less than 10 mm. Partial response required at least a 30% decrease in the sum of the longest diameters of target lesions relative to baseline. A response had to be confirmed at least 4 weeks after the initial assessment.

A response percentage is a binary-outcome summary: each participant either meets the defined response criteria or does not. It offers a direct description of the proportion meeting a prespecified tumor-response definition. Independent review provides an assessment separate from the treating investigator, while standardized RECIST criteria specify how changes in measurable disease are classified. These features support consistent response assessment, but do not make the endpoint a comparison with an untreated or alternative-treatment group.

The registry record states that an exact 95% confidence interval was calculated using the Clopper-Pearson method. It does not report the observed response percentage or interval, so the response estimate and its precision cannot be evaluated from the registry record.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The primary measure is a binary outcome at the participant level: a participant either had a confirmed complete or partial response according to the stated criteria, or did not. The response percentage expresses the frequency of participants meeting that definition in the analysis population. The registry record defines the endpoint and its population, but does not post the resulting percentage or the number of responders. Therefore, no numerical estimate can be interpreted from the record.

What does a confidence interval communicate?

General: a confidence interval describes the statistical uncertainty around an estimated quantity under a specified method and its assumptions. For a response proportion, an interval gives a range of values compatible with the observed data under that method. A wider interval indicates less precision than a narrower interval. The interval does not guarantee that the true value lies within its limits, and it does not account automatically for systematic errors such as differences in assessment or selection into the study.

The registry record specifies an exact 95% confidence interval using the Clopper-Pearson method. General: this is an interval procedure for a binomial proportion. Its use does not supply the missing numerator or denominator, and the interval itself cannot be reconstructed from the registry record because the response result is not posted. No independent recalculation is therefore presented.

Why does the assessment definition matter?

RECIST v1.1 supplies explicit criteria for classifying tumor changes. The record describes complete response in terms of disappearance of lesions, with additional requirements for pathological lymph nodes and tumor markers when applicable. Partial response is based on a decrease in the summed longest diameters of target lesions relative to baseline. Requiring confirmation at least 4 weeks after an initial assessment is intended to ensure that a response classification is sustained on a later assessment rather than recorded from a single observation.

The primary endpoint was assessed by an Independent Review Facility. The registry also lists investigator-assessed response as a secondary outcome. These are different assessment sources, not interchangeable measures. Independent review can provide an assessment separate from the investigators responsible for care, while investigator assessment reflects the treating study team's evaluation. The registry record does not report either response result, so their values cannot be compared.

What can a single-arm analysis establish?

General: a single-arm study estimates outcomes in the group that received the intervention. It does not randomly assign participants to contrasting interventions, and it has no concurrent control group in this cohort. Thus, an observed response percentage would be descriptive of the enrolled population and study conditions. It would not, on its own, estimate what would have happened to the same participants under an alternative intervention.

The registry record identifies an intent-to-treat population as all Cohort 1 participants who received any amount of study drug. It separately defines the objective response-evaluable population as intent-to-treat participants with measurable disease at baseline. This distinction matters because an endpoint may be calculated in a subset rather than in everyone enrolled. The record does not post the population counts, the handling of missing response assessments, or the analysis details needed to understand exclusions from the primary endpoint calculation.

What analysis details are unavailable?

The registry record lists 14 outcome measures and secondary outcomes that include response assessed by the investigator, duration of response, progression-free survival, progression or death, and death. It does not report statistical analyses for these outcomes. The analysis plan, any multiplicity procedure, interim analyses, and sample-size justification are not reported in the registry record. These omissions prevent evaluation of how the different outcomes were prioritized or how missing observations and repeated assessments were handled.

General: with multiple endpoints, the chance of at least one apparently notable finding can increase if each is assessed separately without a plan for interpreting the collection. A prespecified multiplicity approach can address that issue, but no such approach should be assumed when it is not documented. Likewise, general sample-size calculations depend on design assumptions and a target effect; the registry record does not report the rationale for the 119 enrollment. That enrollment should not be treated as evidence of a particular power or precision.

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("Cohort 1: Treatment-naive Cisplatin Ineligible Par")
n      <- c(119)
events <- c(27)
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  Cohort 1: Treatment-naive Cisplatin Ineligible Par  119  27  22.7  15.5  31.3  16.1  31.0  16.1  31.0

11. Multiplicity, Missing Data and Analysis Populations

The registry record lists 14 outcome measures. The primary outcome is confirmed objective response assessed by an Independent Review Facility. Secondary outcomes include duration of response, progression-free survival, progression or death, response assessed by the investigator, and death. However, the registry record does not report a statistical analysis plan, a hierarchy among secondary outcomes, or a procedure for controlling multiplicity. It is therefore not possible to determine whether or how the set of outcomes was addressed as a family of statistical tests.

The primary endpoint population was defined as intent-to-treat participants with measurable disease at baseline. The intent-to-treat population included Cohort 1 participants who received any amount of study drug. These definitions clarify the intended populations but do not state how many participants were included in each analysis. The registry record does not report how missing assessments, withdrawals, or participants without an evaluable response were handled, or whether any sensitivity analyses were conducted.

The registry record also does not report interim analyses, a missing-data strategy, or the sample-size justification. No assumptions about these features can be inferred from the endpoint definition or enrollment count. The available information is sufficient to identify the intended primary outcome and its stated confidence-interval method, but not to assess the full analysis process or the effect of missing outcome data on the estimate.

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 serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: urinary tract infection (3.4%), acute kidney injury (3.4%), diarrhoea (2.5%), small intestinal obstruction (2.5%), sepsis (2.5%).

For the listed infusion group, 119 participants were recorded as at risk for serious adverse events, and 49 were recorded as affected. For other adverse events, 119 were recorded as at risk and 110 as affected. The registry record also lists 96 deaths. These are separate reported counts; they should not be added together, since participants and event categories may overlap. The record does not provide detail about timing, duration, event severity beyond the serious-event classification, or cause of death.

The most frequent serious adverse event terms listed are urinary tract infection and acute kidney injury, each at 3.4%. Diarrhoea, small intestinal obstruction, sepsis, and dehydration are each listed at 2.5%. These percentages describe the registry's event listings; they do not establish that an event was caused by atezolizumab. The record does not report a comparison group that could be used to contrast event frequencies.

General: safety counts describe events recorded during observation, while interpretation also depends on definitions, follow-up, attribution, and participant exposure. Because the record does not provide those details for these figures, the counts and listed percentages should be read as descriptive summaries, not as comparative risk estimates.

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

The registry record establishes that this Phase II study evaluated atezolizumab in a defined group of participants with locally advanced or metastatic urothelial bladder cancer who were treatment-naïve for advanced disease and ineligible for cisplatin-containing chemotherapy. It specifies a single-group design, the dosing schedule, the treatment stopping conditions, and a primary endpoint based on confirmed response by RECIST v1.1. It also states that an exact 95% confidence interval was to be calculated using the Clopper-Pearson method.

It does not report the primary response percentage or confidence interval. Consequently, the record does not establish the observed frequency of confirmed complete or partial response, nor how precisely that frequency was estimated. The enrollment count alone cannot answer those questions. Nor can the listed safety counts establish comparative safety in the absence of a control group and additional information on attribution and exposure.

Even a fully reported response estimate from this design would have a descriptive interpretation: it would summarize outcomes in the participants analyzed under the study's eligibility rules, treatment schedule, and assessment process. A single-arm study cannot isolate the effect of treatment from other possible influences by comparing randomized groups. It also cannot directly establish how outcomes would differ under another treatment.

General: statistical uncertainty and causal uncertainty are distinct. A confidence interval can quantify sampling uncertainty around an estimate under a specified model, but it cannot resolve the lack of a control group or correct for unreported analysis choices. The registry record is informative about the intended question and methods, but the missing primary result and analysis details limit what can be concluded from the posted information.

14. Why This Trial Matters Statistically

This trial is statistically informative as an example of a single-arm Phase II design centered on a defined binary response endpoint. The registry record provides operational detail about the target population, intervention, RECIST v1.1 response criteria, confirmation requirement, assessment schedule, and confidence-interval method. Those details matter because an outcome estimate is interpretable only in relation to who was assessed, how response was defined, and how uncertainty was to be summarized.

It also illustrates why the analysis population needs to be distinguished from overall enrollment. The record defines the intent-to-treat population and a narrower objective response-evaluable population, but does not give their counts or the endpoint result. Without those counts and the response estimate, readers cannot assess precision or determine how much the analyzed population differs from enrollment.

The design has an important inferential boundary: there is no concurrent comparison group. A response percentage, if posted, would describe the study group, but would not alone quantify a treatment contrast. This distinction is central when interpreting nonrandomized evidence and when considering comparisons with other studies.

Finally, the record lists several outcomes beyond the primary endpoint without reporting an analysis plan or results. That makes it useful to separate what is clearly specified from what remains unavailable: the endpoint and its intended statistical method are stated, while the estimate, analysis populations' sizes, missing-data handling, multiplicity approach, and sample-size rationale are not reported in the registry 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 principal limitation for statistical interpretation is that the registry record does not report the primary response percentage or its 95% confidence interval. It also does not report the number of participants included in the response-evaluable population or the number meeting the response definition. As a result, the primary endpoint cannot be quantitatively assessed from the record.

The single-arm design has no randomized or concurrent control group. Any response or safety summaries are therefore descriptive, rather than estimates of a contrast with another treatment. The registry record does not report a statistical analysis plan, sample-size justification, interim analyses, or a multiplicity procedure for the 14 outcome measures.

The record defines the relevant analysis populations but does not report their sizes or explain how missing or unevaluable assessments were handled. This leaves uncertainty about the influence of incomplete outcome data. The maximum listed follow-up of 23.52 months does not describe follow-up for each participant, and individual follow-up information is not reported.

For safety, the registry lists affected counts and selected serious adverse event terms, but the available record does not specify attribution, exposure-adjusted rates, or a comparative group. The listed deaths should not be interpreted as deaths caused by study treatment. These limitations restrict conclusions to the recorded design and descriptive safety summaries.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: A Study of Atezolizumab in Participants With Locally Advanced or Metastatic Urothelial Bladder Cancer (Cohort 1) — NCT02951767.
• Registry record last updated 2024-03-28; status last verified 2024-03.
• Results first posted 2017-04-21.
• 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

NCT02951767 is a completed Phase II, multicenter, single-arm study of atezolizumab in participants with locally advanced or metastatic urothelial bladder cancer. Cohort 1 included treatment-naïve, cisplatin-ineligible participants; recorded enrollment was 119. Treatment was 1200 mg intravenously on Day 1 of 21-day cycles, continuing until RECIST v1.1 disease progression or unmanageable toxicity.

The primary endpoint was the percentage with a confirmed complete or partial response assessed by an Independent Review Facility, with an exact 95% confidence interval specified using the Clopper-Pearson method. The primary response result, its interval, analysis population counts, and detailed analysis plan are not reported in the registry record. The design has no control arm, so it cannot provide a randomized treatment comparison.