A Phase 2 Study of Sitravatinib in Combination With PD-(L)1 Checkpoint Inhibitor Regimens in Patients With Advanced or Metastatic Urothelial Carcinoma: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for A Phase 2 Study of Sitravatinib in Combination With PD-(L)1 Checkpoint Inhibitor Regimens in Patients With Advanced or Metastatic Urothelial Carcinoma (NCT03606174), a non-randomized phase 2 study. The primary outcome is Objective Response Rate (ORR).

2. Clinical Question

3. Trial Design

This was a Phase 2, parallel, non-randomized, unmasked study sponsored by Mirati Therapeutics Inc. The registry lists an enrollment of 260 and a terminated status. The recorded study start was 2018-09-11, and the primary completion date was 2022-08-03. Those dates and status describe the registry entry; they do not by themselves explain why the study was terminated or how termination affected follow-up.

The study evaluated sitravatinib in combination with PD-(L)1 checkpoint inhibitor regimens in advanced or metastatic urothelial carcinoma. Cohorts 1 to 8 paired sitravatinib with nivolumab, with cohort eligibility defined by prior treatment and, in some groups, platinum-based chemotherapy eligibility. Cohort 9 was described as having a lead-in dose-escalation portion and a dose-expansion portion and involved sitravatinib, pembrolizumab and enfortumab vedotin. Further details of the cohort 9 dosing description are not reported in the registry record.

The non-randomized allocation matters for interpretation. The cohorts were not assigned by random allocation, and their entry criteria reflected different histories of prior treatment. Differences in observed outcomes across cohorts, if reported, could reflect those population differences as well as treatment; they would not constitute an unbiased head-to-head comparison by virtue of being displayed side by side. There is no like-for-like randomized control comparison in the registry record.

The listed primary endpoint was confirmed investigator-assessed ORR, measured in the Full Analysis Set. The endpoint timeframe was up to approximately 3 years. The record also lists 11 outcome measures and names secondary outcomes including adverse events, duration of response, clinical benefit rate, progression-free survival, 1-Year Survival Rate and overall survival. The detailed statistical analysis plan, the reasons for termination, and whether termination altered planned endpoint assessments are not reported in the registry record.

4. Participant Flow and Accounting

The registry reports how participants moved through each stage of the study. All counts below are exactly as posted; missing counts are not inferred from other stages.

Before assignment: Screening tests and procedures were performed within the 28 days preceding administration of the first dose.

Only the first 6 of 11 groups are shown.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Measured in the Full Analysis Set, which included participants who received ≥ 1 dose of each study treatment drug (ie, ≥ 1 dose of both sitravatinib and nivolumab for Cohorts 1 to 8, or ≥ 1 dose each of sitravatinib, pembrolizumab, and enfortumab vedotin for Cohort 9). Per the statistical analysis plan (SAP), Cohort 9 arms are grouped as pre-specified as there was never intent to compare efficacy endpoints across Cohort 9 dose levels.

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: Measured in the safety population, which included all participants who received ≥ 1 dose of any study treatment (sitravatinib, nivolumab, pembrolizumab, or enfortumab vedotin).

Only the first 6 of 12 groups are shown.

6. Primary Endpoint

The registry lists the primary outcome as Objective Response Rate (ORR), measured over Up to approximately 3 years.

Objective response rate (ORR) was defined as the number of participants with a confirmed investigator-assessed complete response (CR) or partial response (PR), using RECIST version 1.1. A CR required complete disappearance of target lesions, except for nodal disease, with target nodes reduced to normal size (short axis < 10 mm). A PR required a decrease of at least 30% from baseline in the sum of diameters of target measurable lesions.

ORR is used in general as a measure of observed tumor shrinkage meeting prespecified response criteria. Counting confirmed responses applies a defined assessment rule rather than treating every change in tumor measurements as a response. It does not, on its own, describe how long a response lasts, whether disease remains stable, or how survival compares with another treatment.

The registry identifies the Full Analysis Set as participants who received at least one dose of each study treatment drug: sitravatinib and nivolumab in cohorts 1 to 8, or sitravatinib, pembrolizumab and enfortumab vedotin in cohort 9. The endpoint timeframe was up to approximately 3 years. The actual ORR count and any response proportion or confidence interval are not reported in the registry record. The response criteria and analysis population are therefore described, but the primary endpoint result cannot be interpreted from the available registry results.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The primary endpoint is binary at the participant level: each participant either meets the confirmed CR or PR criteria or does not. The registry describes the endpoint as a count of participants, rather than posting an ORR count or response proportion. The definitions specify the tumor-measurement criteria used to classify response, but the record does not include the assessment results needed to reproduce an endpoint estimate.

What does a confidence interval add?

General statistical explanation: an estimate from a study is subject to sampling variability. A confidence interval communicates uncertainty around an estimate under the chosen statistical method and its assumptions. It is not a guarantee that the true value lies within the interval, nor does it address every source of bias. For this trial, the registry record does not report an ORR estimate or confidence interval, so neither the endpoint's magnitude nor its statistical precision can be evaluated from the posted results.

Confidence intervals are especially useful when a study has multiple cohorts with different eligibility criteria, because each cohort's estimate may be based on a different number of participants. The registry's cohort-level safety groups have different at-risk counts, but these safety counts are not ORR estimates and cannot be used to infer response precision. No cohort-specific response estimates or uncertainty intervals are reported in the registry record.

Why is non-randomized allocation important?

General statistical explanation: random allocation aims to balance measured and unmeasured prognostic factors between treatment groups, on average. In a non-randomized study, treatment cohorts may differ before treatment begins. Here, the registry explicitly separates cohorts by prior therapy and eligibility for platinum-based chemotherapy, among other treatment histories. Thus, even if numerical outcomes for each cohort were available, a simple contrast could mix treatment effects with differences in who entered each cohort. The record identifies no randomized control arm, and the listed design does not support a direct randomized comparison.

The study is also unmasked. In general, knowledge of treatment can potentially influence subjective parts of outcome assessment or reporting. RECIST criteria provide structured response definitions, and the registry specifies investigator assessment, but the record does not report details of independent review, adjudication, or any masking of outcome assessors. Those procedures are not reported in the registry record.

What can planning calculations tell a reader?

General statistical explanation: sample-size planning evaluates how many participants might be needed to detect a chosen effect under stated assumptions. The registry's planning description refers to a two-sided significance level of 0.05, equal allocation between two arms, a normal approximation, and detection of a stated fraction of the observed effect. It characterizes those calculations as planning illustrations from the observed effect, not the trial's own design. The resulting numerical planning estimates are not reported in the registry record.

These assumptions should not be mistaken for evidence that the trial had a particular probability of detecting an effect. In particular, a sample-size calculation is not observed power. A post hoc calculation based on an observed effect does not make an unreported primary result available, and it cannot correct for the non-randomized cohort structure. The registry record does not report the original sample-size justification, the target effect used for original design, or whether the stated planning illustration reflects the study's actual enrollment plan.

For time-to-event endpoints, general methods account for event timing and censoring rather than treating every participant as having the same complete follow-up. The registry lists progression-free survival, duration of response and overall survival, but it does not provide their estimates, event counts or analysis details in the posted results. The planning description notes that time-to-event illustrations use a total number of events; numerical illustrations are not reported in the registry record.

How should safety counts be read?

General statistical explanation: a count of participants with an event is a descriptive summary. Its interpretation depends on the definition of the event, the group at risk, observation time, and completeness of reporting. Counts alone do not account for differences in duration of exposure or follow-up. A group with more participants may have more recorded events even if individual risk is not greater. Comparisons require appropriate denominators and methods, and in this non-randomized study they remain subject to differences between cohorts.

The safety information is presented as counts by cohort, including participants at risk and participants with serious adverse events, other adverse events and deaths. The registry record does not provide the event timing or exposure-adjusted rates needed to characterize incidence over time. It also does not report a formal comparative safety analysis. The counts should therefore be read as descriptive cohort summaries rather than as estimates of causal treatment differences.

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", "Cohort 2", "Cohort 3", "Cohort 4", "Cohort 5", "Cohort 6")
n      <- c(49, 23, 18, 9, 53, 27)
events <- c(6, 5, 4, 0, 17, 9)
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  49  6  12.2  4.6  24.8  5.7  24.2  5.8  24.3
ARM  Cohort 2  23  5  21.7  7.5  43.7  9.7  41.9  9.8  42.2
ARM  Cohort 3  18  4  22.2  6.4  47.6  9.0  45.2  9.1  45.6
ARM  Cohort 4  9  0  0.0  0.0  33.6  -0.0  29.9  0.3  30.8
ARM  Cohort 5  53  17  32.1  19.9  46.3  21.1  45.5  21.1  45.6
ARM  Cohort 6  27  9  33.3  16.5  54.0  18.6  52.2  18.6  52.4

11. Multiplicity, Missing Data and Analysis Populations

The registry lists 11 outcome measures, including ORR as the primary endpoint and several secondary outcomes. Testing or estimating many outcomes can raise multiplicity concerns. General statistical explanation: when multiple hypotheses are tested, the chance of at least one apparently notable result can increase if no adjustment or hierarchical testing strategy is used. The registry record does not report whether multiplicity was controlled, how primary and secondary outcomes were ordered for inference, or whether any adjustment was planned.

The record lists 0 analyses. It does not report a statistical analysis plan, the number of planned comparisons, analysis time points, interim analyses, or any correction for multiple testing. These details are not reported in the registry record, so the intended inferential status of secondary outcomes cannot be determined from the posted information.

Missing-data handling is also not reported in the registry record. For ORR, relevant details would include how participants without an evaluable post-baseline assessment were handled, how missing or delayed assessments were treated, and whether confirmation requirements were applied consistently. The endpoint population is described as the Full Analysis Set, requiring at least one dose of each study treatment drug, but that definition does not explain the handling of missing outcome assessments after treatment began.

For time-to-event outcomes, the registry record does not report censoring rules, follow-up completeness, or how loss to follow-up was handled. Nor does it describe whether safety analyses used all treated participants or another analysis population. These omissions limit the ability to assess potential bias from missing outcomes or differing follow-up. They do not establish that missing data occurred or that any particular method was used.

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: infusion site extravasation (25.0%), sepsis (25.0%), urinary tract infection (25.0%), dizziness (25.0%), syncope (25.0%).

The registry reports serious adverse event counts by cohort, with a count at risk and a count affected. For cohorts 1 to 9 dose level 3, respectively, the recorded at-risk and affected counts were 49 and 35; 23 and 15; 18 and 10; 9 and 4; 53 and 26; 27 and 14; 56 and 23; 9 and 3; 8 and 5; 4 and 1; and 4 and 3. These are participant counts, not event rates or estimates adjusted for time receiving treatment.

The record also lists deaths by group: 37 in cohort 1, 15 in cohort 2, 15 in cohort 3, 7 in cohort 4, 35 in cohort 5, 12 in cohort 6, 36 in cohort 7, 8 in cohort 8, 5 in cohort 9 dose level 1, 3 in cohort 9 dose level 2 and 0 in cohort 9 dose level 3. These figures are registry safety-group counts; the record does not report causes, timing, attribution to study treatment or comparative analysis for these deaths.

General statistical explanation: serious adverse event counts describe how many participants were recorded as affected, while an incidence rate would also account for exposure time. Cohorts differ in size and treatment history, and the registry does not report exposure-adjusted rates. The listed serious terms each have a recorded value of 25.0%: infusion site extravasation, sepsis, urinary tract infection, dizziness, syncope and haematuria. The denominator and detailed event-level context for those terms are not reported in the registry record, so they should not be read as population-wide rates.

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

The record establishes that a Phase 2, non-randomized study evaluated sitravatinib-containing checkpoint inhibitor regimens in advanced or metastatic urothelial carcinoma and defined confirmed investigator-assessed ORR as its primary endpoint. It also describes the analysis population and provides cohort-level adverse event counts. These details help identify what the investigators intended to measure and how the study groups were organized.

The registry record does not report the ORR result. Without the number of confirmed responses, an estimate of response frequency, or an interval describing precision, the primary efficacy endpoint cannot be judged from the posted results. The absence of an ORR result in the record is not evidence that participants did or did not respond; it is a limit on what can be concluded from this record.

The safety figures are descriptive counts. They do not by themselves establish whether an event was caused by a study drug, how risk changed with time, or whether one regimen was safer than another. Cohorts were not randomized and differed by prior treatment and eligibility, so direct comparisons among them would be vulnerable to confounding. The available counts also do not supply an exposure-adjusted comparison.

General statistical explanation: efficacy and safety conclusions depend not only on numerical estimates but also on study design, precision, completeness of follow-up and analysis methods. A confidence interval, if reported, would help characterize uncertainty but would not eliminate non-randomized design limitations. The registry does not report confidence intervals, p-values, a full analysis plan, or handling of missing outcomes for the primary endpoint.

The study's terminated status is a recorded administrative status, not an explanation of why the study ended or a result about treatment. The reason for termination and its implications for planned analyses are not reported in the registry record. Accordingly, the appropriate reading is limited: the registry documents the study's design, endpoint definition and some safety counts, but it does not provide the primary efficacy finding needed to assess the endpoint itself.

14. Why This Trial Matters Statistically

This trial illustrates a common statistical challenge in oncology studies with multiple treatment-history cohorts. The cohorts address distinct clinical contexts, but non-randomized assignment and different eligibility criteria mean that between-cohort contrasts do not function as randomized comparisons. Keeping that design distinction explicit prevents descriptive differences from being mistaken for causal treatment effects.

The endpoint definition is also important. ORR depends on specified RECIST criteria, investigator assessment and confirmation of CR or PR. Readers need both the rule for classifying response and the observed result to understand what was measured and how often the criterion was met. The registry record supplies the rule but not the response count or estimate, leaving the primary endpoint unevaluable from the posted results.

Safety summaries show why denominators and follow-up matter. The cohort counts vary, and the record does not provide exposure-adjusted rates or the timing and attribution of events. General statistical explanation: raw counts can describe the recorded burden but cannot alone establish comparative risk, especially when groups differ in size, treatment history and time under observation.

The record also names a range of secondary outcomes while reporting 0 analyses and no statistical analysis plan. That makes it impossible to determine how multiplicity, missing assessments, interim analyses or outcome hierarchy were addressed. The stated planning assumptions are useful to distinguish planning illustrations from trial design, but numerical sample-size illustrations and the original sample-size justification are not reported in the registry record.

Statistically, the value of the record lies in clarifying its design and measurement framework while also making the limits of its posted results visible. A careful reading separates a defined endpoint from an observed endpoint result, descriptive safety counts from comparative evidence, and hypothetical planning calculations from observed study performance.

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 interpreting efficacy is that the registry record does not report the primary ORR count, response proportion, confidence interval or p-value. Although the endpoint definition and Full Analysis Set are described, the actual outcome cannot be assessed. The record also does not report numerical results for the listed secondary efficacy outcomes.

The non-randomized design does not provide a randomized control comparison. Cohorts were formed around different prior treatment histories and platinum-based chemotherapy eligibility. Any comparison across them would be subject to confounding, and the record does not describe an adjusted comparative analysis. The study was unmasked; procedures for independent or masked outcome assessment are not reported in the registry record.

The registry reports cohort-level safety counts, but does not provide exposure-adjusted event rates, event timing, treatment attribution, detailed follow-up, or a formal comparative safety analysis. The listed serious adverse event terms do not have a fully reported denominator or detailed event-level context in the available record. Death counts are presented without causes or attribution.

The statistical analysis plan, original sample-size justification, multiplicity strategy, missing-data procedures, censoring rules for time-to-event outcomes and interim analysis details are not reported in the registry record. The record's terminated status is stated, but the reason for termination and any effect on data collection or planned analysis are not reported. These limits prevent assessment of the completeness and inferential framework of the results beyond the information actually posted.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: A Phase 2 Study of Sitravatinib in Combination With PD-(L)1 Checkpoint Inhibitor Regimens in Patients With Advanced or Metastatic Urothelial Carcinoma — NCT03606174.
• Registry record last updated 2023-08-24; status last verified 2023-08.
• Results first posted 2023-08-24.
• 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

NCT03606174 is a Phase 2, parallel, non-randomized, unmasked study sponsored by Mirati Therapeutics Inc. It enrolled 260 participants and has a terminated status in the registry. The study evaluated sitravatinib with PD-(L)1 checkpoint inhibitor regimens in advanced or metastatic urothelial carcinoma across cohorts defined by prior treatment and eligibility.

The primary endpoint was the count of participants with confirmed investigator-assessed CR or PR under RECIST version 1.1, measured in the Full Analysis Set over up to approximately 3 years. Cohorts 1 to 8 received sitravatinib with nivolumab; cohort 9 involved sitravatinib, pembrolizumab and enfortumab vedotin.

The record lists 11 outcome measures and 0 analyses. It provides cohort-level adverse event and death counts, but does not report the primary ORR result, a statistical analysis plan, or the reason for termination. No randomized comparator is identified.