Individualized Adaptive De-escalated Radiotherapy for HPV-related Oropharynx Cancer: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Individualized Adaptive De-escalated Radiotherapy for HPV-related Oropharynx Cancer (NCT03416153), a non-randomized phase 2 study. The primary outcome is The Percentage of Patients With Local Regional Recurrence (LRR) of Disease.

2. Clinical Question

Statistically, the central question is descriptive: what proportion of participants treated under this PET-guided strategy experienced local-regional recurrence during the stated timeframe? The study was designed to estimate risk in a treatment setting where some participants continued standard therapy while others received de-escalated therapy according to parameters. That is different from asking whether one strategy causes better outcomes than another. A direct treatment-effect comparison would require an analysis that addresses how treatment groups were formed and how their baseline characteristics differ; the registry record does not report such an analysis.

The design is prospective and phase II, but the primary outcome data needed to assess the recurrence estimate are absent from the record. The record also does not provide a reported result for either treatment group separately. Consequently, the study question can be described, but the observed frequency of recurrence and the precision of any estimate cannot be assessed from these registry entries.

3. Trial Design

This is a completed, prospective, multicenter phase II study sponsored by the University of Michigan Rogel Cancer Center. It is registered as NCT03416153 and has a parallel, non-randomized design with no masking. The record lists enrollment of 91 participants, a start date of 2018-05-21, and primary completion on 2023-05-05. Completion status is listed as completed.

The study concerns individualized adaptive de-escalated radiotherapy for HPV-related oropharynx cancer. Its stated aim is to use PET/CT before treatment and during treatment to guide radiation de-escalation. The listed interventions are carboplatin, radiation therapy, and paclitaxel. In the standard-treatment group, the description specifies 70 Gy in 35 fractions, once daily, 5 days a week, together with weekly carboplatin and paclitaxel. The de-escalation group initially follows that same described regimen. If certain parameters are met, radiation to the high-risk Planned Target Volume is described as being reduced to 54Gy. The remaining text of that arm description, including the criteria and any additional dose details, is not reported in the registry record.

Non-randomized allocation means participants were not assigned by random chance to the two treatment groups, according to the registry. The record does not describe how treatment assignment was decided. As a result, the group labels should not be read as if they were randomized arms. Clinical or imaging features that influenced the treatment strategy could also be related to recurrence risk. Without the assignment rules and relevant participant characteristics, the extent of this issue cannot be evaluated.

The primary outcome description states that all patients would be analyzed together to estimate local-regional recurrence risk under the overall strategy, and that results would also be estimated separately by treatment received. However, the primary outcome’s actual result and separate group estimates are not reported in the registry record. The record lists secondary outcomes covering patterns of failure, overall survival, progression-free survival, toxicity, and detectable circulating tumor DNA. Results for those outcomes are not included in the registry entries. A sample-size justification, interim analysis plan, and more complete statistical analysis plan 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: 6 subjects were removed from the trial before being assigned a cohort.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: 1 subject did not proceed with study treatment in the De-escalation treatment

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 The Percentage of Patients With Local Regional Recurrence (LRR) of Disease, measured over 1 Year.

The primary endpoint is the percentage of participants with local-regional recurrence of disease at 1 Year. Local-regional recurrence refers to the return or persistence of cancer in the local or regional disease area, as assessed in this study using RECIST criteria. The registry describes the endpoint as a number of participants expressed as a percentage and specifies a 95% confidence interval as the dispersion measure.

This endpoint is used to quantify recurrence risk in the population treated under the study strategy. The study’s stated purpose is not limited to comparing two treatment groups: its outcome description says the overall group is to be used to estimate risk when some participants receive standard therapy and others receive de-escalated therapy. Separate estimates by treatment group are also described. Such estimates could help characterize outcomes within each treatment pathway, but non-randomized allocation means that the groups should not be treated as directly comparable randomized arms.

A percentage alone would not convey the precision of the estimate. A confidence interval describes the uncertainty around an estimated value; a wider interval generally indicates less precision, while a narrower interval indicates more precision. This is a general statistical explanation. The registry record specifies the intended confidence interval but does not report the primary endpoint percentage or its interval. It therefore does not allow the recurrence risk or its precision to be judged.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The primary endpoint is a binary event outcome: each participant either has a recorded local-regional recurrence within the specified assessment framework or does not. The registry describes the result as a percentage of participants. In general, a percentage for a binary outcome is an estimate of the share of the analyzed population with the event. The actual percentage and event count are not reported in the registry record, so no observed estimate can be interpreted.

What does a confidence interval communicate?

In general, a confidence interval expresses the statistical uncertainty around an estimate. It is not a guarantee that the true value lies within the interval, nor does it describe differences between individual participants. Its width depends on factors such as the number of observations and the amount of information in the data. The primary endpoint description specifies a 95% confidence interval, but the interval itself is not reported in the registry record. Without it, the precision of the planned recurrence estimate cannot be assessed.

Why is the analysis population important?

An endpoint estimate depends on which participants are included and how missing outcome information is handled. The record reports enrollment of 91 and notes that 1 subject did not proceed with study treatment in the de-escalation group. It does not state how many participants had evaluable outcome information, whether anyone was excluded from the primary analysis, or how participants without a documented assessment were treated analytically. In general, those choices can affect both the estimated percentage and its uncertainty. It would be inappropriate to infer the analyzed sample from enrollment alone.

What does non-randomized allocation mean for comparisons?

Random assignment is a general design method intended to make treatment groups comparable, on average, with respect to both measured and unmeasured characteristics. In this study, allocation is listed as non-randomized. The registry record does not describe the criteria used to assign participants to standard or de-escalation treatment. If clinical or imaging features influenced that decision, those features could also be associated with recurrence risk. A difference between group percentages could therefore reflect differences in participant characteristics, treatment selection, or other factors, as well as any treatment effect. The record identifies the arms as not being a like-for-like control comparison.

The outcome description calls for an overall estimate and separate estimates by treatment received. Those are descriptive aims. An adjusted comparison could account for measured differences between groups, but the record does not report whether any adjustment was planned or performed, or which variables it might have included. No unadjusted or adjusted primary outcome estimates are reported either.

What can be inferred from the stated timeframe?

The primary endpoint timeframe is 1 Year. In general, a fixed-time percentage requires a clear definition of how recurrence status is established for every participant at that point. Follow-up can vary, and some participants may have incomplete assessments. Time-to-event methods can handle event timing and censoring differently from a simple proportion, but the registry record does not say which method was used for the primary endpoint. It also does not report follow-up completeness or censoring details. The endpoint description identifies a timeframe and a confidence interval, but those details do not establish the statistical method.

Why do analysis plans and sample-size assumptions matter?

A statistical analysis plan can specify the outcome definition, analysis population, handling of missing observations, and method used to calculate estimates and confidence intervals. The sample-size justification can explain how the planned enrollment relates to the study’s precision goals or ability to detect an effect. Neither a full analysis plan nor a sample-size justification is reported in the registry record. The record also does not report interim analyses. Their absence from the record does not establish whether such work occurred; it means the relevant details are not reported in the registry record.

General statistical interpretation also distinguishes a study’s planned ability to detect an effect from the precision of an observed estimate. A planning calculation is not evidence about what the study actually observed, and it is not observed power. No trial-specific sample-size justification or primary result is available in the registry record to support a numerical interpretation of precision or power.

How should the safety counts be understood?

The safety listing gives numbers of participants at risk and affected in each group, as well as deaths. These are counts rather than an analysis of time to event, severity over time, or causal attribution. In general, the meaning of an adverse-event count depends on the event definition, the period of observation, whether participants can contribute more than one event, and how events were coded. Those details are not reported in the registry record. The listed serious adverse-event terms and percentages describe the registry’s safety entries; they do not provide evidence about the primary recurrence endpoint.

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("Standard Treatment", "De-escalation Treatment")
n      <- c(49, 35)
events <- c(1, 1)
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  Standard Treatment  49  1  2.0  0.1  10.9  0.4  10.7  0.5  10.6
ARM  De-escalation Treatment  35  1  2.9  0.1  14.9  0.5  14.5  0.7  14.5

11. Multiplicity, Missing Data and Analysis Populations

The registry names several secondary outcomes in addition to the primary local-regional recurrence endpoint: patterns of failure, overall survival, progression-free survival, incidence of toxicity, and detectable circulating tumor DNA. It reports 8 outcome measures, but does not provide results or a complete analysis plan for them. The record does not explain how multiple outcomes or multiple analyses were prioritized, whether any correction for multiple testing was planned, or how confirmatory claims would be distinguished from exploratory findings. These details are not reported in the registry record.

Multiplicity matters in general because testing or estimating many outcomes can increase the chance that some findings appear notable by chance. The consequences depend on the purpose of the analyses and the methods used; a correction is not automatically appropriate for every descriptive endpoint. Since outcome-specific analysis methods and results are absent, the approach to multiplicity cannot be assessed.

Missing-data handling is also not reported. The registry notes that 1 subject did not proceed with study treatment in the de-escalation group, but does not say whether this person was included in any analysis, whether other participants had incomplete follow-up, or how missing outcome information was handled. It does not identify an intention-to-treat, per-protocol, or other analysis population. No assumption about those choices should be made from enrollment alone.

The primary endpoint description says that all patients would be analyzed together and that separate estimates would also be reported for participants receiving standard or de-escalated treatment. However, the actual analysis populations, event counts, estimates, confidence intervals, and missing-data procedures are not reported in the registry record. These omissions limit interpretation of both the overall endpoint and the group-specific estimates.

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: cardiac arrest (4.1%), dysphagia (2.8%), hyperglycemia (2.8%), thromboembolic event (2.8%), metabolism and nutrition disorders - other, specify (2.8%).

The registry lists serious adverse-event information separately for the two treatment groups. For Standard Treatment, it reports 49 participants at risk and 5 affected by serious adverse events; for De-escalation Treatment, it reports 36 at risk and 3 affected. It also lists deaths as 2 in the standard group and 1 in the de-escalation group. The record does not state whether the deaths were caused by treatment or disease, or provide further context for their timing.

For other adverse events, the registry reports 49 at risk and 49 affected in the standard group, and 36 at risk and 36 affected in the de-escalation group. These are the registry’s group-level counts; the event definitions and observation period are not reported in the registry record. The counts should not be interpreted as a randomized comparison because treatment allocation was non-randomized.

The listed top serious-event terms include Cardiac arrest at 4.1%, Dysphagia at 2.8%, Hyperglycemia at 2.8%, Thromboembolic event at 2.8%, Metabolism and nutrition disorders - Other, specify at 2.8%, and Anorexia at 2.0%. The record does not provide additional details about event severity, duration, recurrence, attribution, or whether a participant could contribute to more than one term. These figures describe reported safety entries, not the overall balance of benefit and harm.

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

The record establishes that a prospective, completed phase II study evaluated a PET/CT-guided approach to radiotherapy in HPV-related oropharynx cancer. It defines a primary question about local-regional recurrence at 1 Year and describes plans to estimate risk for the overall study population and separately by treatment group. These design details identify what the trial set out to study.

The registry record does not report the primary endpoint result. It therefore does not establish the observed percentage of participants with local-regional recurrence, the precision of that percentage, or how the result compares with any expected benchmark. It also does not report primary outcome estimates separately for the standard and de-escalation groups. No conclusion about the observed recurrence outcome can be drawn from the design description alone.

The study is non-randomized, and the registry explicitly cautions that the arms are not a like-for-like control comparison. Even if group-specific outcome estimates were available, a difference would not, by itself, establish that de-escalation caused a change in recurrence risk. Treatment selection may be associated with participant characteristics that also relate to outcomes, and the record does not describe the assignment criteria or an adjustment strategy.

The safety entries provide some counts and selected serious-event terms, but they do not settle questions about comparative safety. The record does not provide complete event definitions, timing, attribution, or detailed follow-up information. The listed counts should be read as registry entries rather than as a full causal assessment of the treatment strategies.

More broadly, the record is useful for understanding the study’s aim, treatment descriptions, design, and endpoint definition. It does not provide the outcome results or methodological detail needed to judge the trial’s efficacy estimates, their precision, missing-data sensitivity, or comparative treatment effects. The absence of those details is a limit of what can be interpreted from the registry record, not evidence that an outcome did or did not occur.

14. Why This Trial Matters Statistically

This trial raises a statistically important question about adapting treatment using information collected before and during therapy. Its primary endpoint is framed as an estimate of recurrence risk in a population treated under a strategy that includes both standard and de-escalated treatment. That framing distinguishes estimating outcomes under a treatment strategy from testing a randomized contrast between two fixed regimens.

The study’s non-randomized design makes the distinction especially important. Participants were not randomly assigned, and the record does not report the criteria used to determine who received de-escalated treatment. Statistical interpretation must therefore account for the possibility that the groups differed before treatment. A group comparison without sufficient information about those differences could be misleading, even if its arithmetic were straightforward.

The record also illustrates why an endpoint definition is not the same as an outcome result. The primary endpoint, timeframe, assessment framework, and intended confidence interval are described, but the value and interval are not reported. Without those, readers cannot evaluate either the estimated recurrence risk or how precisely it was measured. Enrollment alone cannot replace event counts or an endpoint estimate.

Finally, the registry lists several secondary outcomes and safety entries, but does not report their analyses in the available record. Understanding the primary outcome alongside follow-up, missingness, analysis populations, and prespecified methods would be important for a complete statistical assessment. The study is therefore statistically notable for the questions its design raises about adaptive treatment selection and descriptive risk estimation, while its registry entries do not permit evaluation of its primary result.

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 main trial-specific limitation in interpreting the registry record is the absence of the primary local-regional recurrence result. No event count, percentage, confidence interval, or separate estimate for the two treatment groups is reported. Although the outcome is set at 1 Year and RECIST is named for evaluating response and recurrence, follow-up completeness and the method for handling participants without an assessment are not reported.

The treatment allocation is non-randomized. The record does not explain how participants entered the standard or de-escalation group, what parameters were used to trigger de-escalation, or whether analyses adjusted for factors associated with treatment selection. The de-escalation arm description is incomplete in the registry record: it states that radiation to the high-risk Planned Target Volume would be reduced to 54Gy if certain parameters were met, but the criteria and remaining treatment details are not reported.

The record notes that 1 subject did not proceed with study treatment in the de-escalation group, but does not identify the primary analysis population or explain whether this participant was included. Other missing-data details, the analysis method, sample-size justification, and any interim analysis plan are not reported in the registry record. These gaps prevent assessment of how the endpoint estimate was constructed and how robust it might be to analysis choices.

Safety interpretation is also limited by the absence of details about event definitions, observation periods, attribution, and repeated events. The registry lists selected serious-event terms and group counts, but those entries do not provide a complete comparative safety analysis. These limitations concern interpretation of the registry information and do not determine what the study investigators may have reported elsewhere.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Individualized Adaptive De-escalated Radiotherapy for HPV-related Oropharynx Cancer — NCT03416153.
• Registry record last updated 2026-08-07; status last verified 2026-07.
• Results first posted 2025-05-22.
• 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

NCT03416153 is a completed, prospective, multicenter phase II study sponsored by the University of Michigan Rogel Cancer Center. It is listed as a parallel, non-randomized study with no masking and enrollment of 91 participants. The study evaluates PET/CT-guided adaptive de-escalated radiotherapy for HPV-related oropharynx cancer, using radiation therapy with weekly carboplatin and paclitaxel.

The primary endpoint is the percentage of participants with local-regional recurrence at 1 Year, assessed using RECIST criteria, with a 95% confidence interval specified. The registry describes an overall estimate and separate estimates by treatment group. The primary result, confidence interval, and group-specific outcome estimates are not reported in the registry record.

The record lists secondary outcomes and adverse-event information, but does not provide a complete statistical analysis plan, sample-size justification, missing-data approach, or analysis-population details. Because treatment allocation was non-randomized, the groups are not a like-for-like control comparison.