Reproducibility of Intratumor Distribution of 18F-EF5 in Head and Neck Cancer: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Reproducibility of Intratumor Distribution of 18F-EF5 in Head and Neck Cancer (NCT01774760), a na phase 2 study. The primary outcome is Change From Baseline in 18F-EF5 Standardized Uptake Values (SUV).

2. Clinical Question

3. Trial Design

This was a Phase 2, prospective, non-randomized study with a single group and enrollment of 10. The registry lists patients with stage III-IV head and neck cancer and gives head and neck squamous cell carcinoma as the population for the primary outcome. The sponsor was Turku University Hospital. The study began in 2013-03, had primary completion in 2016-09, and is marked completed.

The study was designed around repeated imaging rather than treatment assignment. Participants were to undergo pretreatment PET/CT scans with 18F-EF5 performed two times. The registry describes the second assessment as 7 days after baseline. That within-participant structure makes the repeated scan the comparator for the initial scan; it does not create a randomized or parallel control group. Allocation is listed as not applicable, and masking is listed as none.

The main design aim was to obtain information on the reproducibility of 18F-EF5 intratumor distribution. The record describes a continuous primary outcome: change from baseline in 18F-EF5 SUV, measured in SUVmax. It says the difference between the baseline and second SUV would be calculated and that Bland-Altman plots would be generated to compare group means. The record does not give the exact acquisition or image-processing procedures, the analysis population, the number of participants with paired scans, or the completed analysis details.

Although the study status is completed, that status alone does not indicate that outcome estimates are available in the registry. No SUV outcome values or analyses are reported. The registry record also does not report a sample-size justification or a detailed statistical analysis plan. Consequently, the design and intended assessment can be described, but the study's empirical reproducibility findings cannot be characterized from this 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: 1 patient withdrawn because of a negative FDG PET scan

Overall Study

Analysis population for the primary outcome, as defined in the registry: Head and neck squamous cell carcinoma

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 Change From Baseline in 18F-EF5 Standardized Uptake Values (SUV), measured over Baseline and 7 days (ie time between the two scans).

The primary endpoint is the change from baseline in 18F-EF5 standardized uptake values, reported in SUVmax. SUV is a measure derived from PET images that expresses tracer uptake in a standardized way. The registry specifies baseline and a second scan 7 days later, with the difference between the two SUV measurements calculated. It identifies the outcome as a mean of a continuous measure with standard deviation as the dispersion measure.

For a reproducibility study, repeated measurements are used to assess whether the same imaging approach yields similar results when repeated under the study conditions. Small scan-to-scan variation would be consistent with greater repeatability, while larger variation would indicate that measured uptake changes between scans. That is a general interpretation; the registry does not report the observed change or its variability for this study.

The registry also describes generating Bland-Altman plots to compare group means. In general, such plots examine the relationship between paired measurements by considering their differences and their average. They can help show whether the discrepancy between repeated measures varies across the measurement range. This differs from relying only on a group-average change: an average near zero can coexist with substantial participant-level variation. The record does not provide the plotted values, limits of agreement, or any other numerical result, so the endpoint's observed precision and repeatability are not reported in the registry record.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

What does a repeated-scan comparison estimate?

In general, a repeated-scan design assesses how measurements vary when the measurement procedure is repeated under specified conditions. Here, the registry describes pretreatment PET/CT scans performed two times, with baseline and a second scan 7 days later. The primary outcome is the change in 18F-EF5 SUV. Because the scans are repeated within the same study participants, the natural unit of comparison is the paired difference for each participant. A group summary of those differences can describe an average shift, but it does not by itself describe how much participants' measurements differ from one scan to the next.

Paired analysis matters because two observations from the same participant are generally related. A method that treats the baseline and repeat scans as though they came from unrelated participants would ignore that pairing. The registry does not give enough analysis detail to confirm what statistical procedure was ultimately used; it describes calculating the difference between baseline and the second SUV.

What do the mean and standard deviation convey?

For a continuous outcome, the mean describes the average value or average change in the group. The standard deviation describes the spread of participant-level values around that average. In a repeatability setting, a mean change near zero would indicate little average directional shift, but would not by itself establish close agreement for individual participants. Differences in opposing directions can cancel in the average while individual scan pairs remain far apart. The standard deviation, along with the actual paired values or agreement analysis, helps describe that variation.

The registry identifies the endpoint as a mean of a continuous measure and names standard deviation as the dispersion measure. It does not provide the mean, standard deviation, or participant-level results. No numerical interpretation of the endpoint's center or spread is therefore possible from the registry record.

Why use Bland-Altman methods?

In general, Bland-Altman analysis evaluates paired measurements by examining the difference between each pair in relation to the pair's average. It is intended to assess agreement, rather than simply test whether group averages differ. The pattern can help identify a systematic shift between repeated measures and whether discrepancies become larger or smaller across the measurement range. Limits of agreement are commonly used to describe the range in which paired differences lie, subject to the method's assumptions. These are general statistical concepts, not reported findings from this study.

The registry says that Bland-Altman plots would be generated to compare group means. It does not report a plot, limits of agreement, a confidence interval, or the method used to calculate any such quantity. A stated plan to generate a plot is not evidence of a particular level of agreement. Without the plotted results, no assessment can be made of bias, variation, or patterns across the range of SUV measurements.

How should precision be assessed?

A point estimate describes the estimated amount of change or agreement, while a confidence interval describes uncertainty around an estimate under the statistical model used. A narrow interval generally indicates greater precision than a wide interval, though precision does not establish that a measurement is clinically useful. The registry does not report confidence intervals or another uncertainty measure for the primary endpoint. With enrollment of 10, estimates could be sensitive to the observed values, but the actual degree of uncertainty cannot be determined without the results and analysis details.

General considerations for continuous outcomes include the distribution of paired differences, unusual observations, and whether variability changes with the average measurement. If variability grows as SUV increases, a single summary of spread may not describe agreement equally well across the full range. The registry record does not report diagnostic checks or transformations, so no claim can be made about these features in this study.

What do the planning assumptions mean?

The registry's sample-size assumptions describe a planning illustration based on the observed effect, with a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation; the effect to detect is a stated fraction of the observed effect. Such an illustration is not the design actually described for this single-group repeated-scan study, which has no randomized arms. Any sample-size values associated with those assumptions are planning illustrations, not observed power and not evidence that the enrolled study had a particular probability of detecting an effect. The registry does not report a sample-size justification for the actual study design.

In general, sample-size calculations depend on the assumed effect, variability, significance level, allocation, and analysis method. Changing those assumptions changes the planning result. A planning calculation based on an observed effect is especially dependent on that effect estimate and should not be read as an independent measure of evidential strength. The record does not report enough information to evaluate a study-specific calculation for paired reproducibility.

What remains unknown?

The registry describes the endpoint and intended comparison, but does not report the completed statistical analysis. It does not specify the analysis population, treatment of missing or unusable scans, handling of outliers, or whether other outcomes were analyzed. It also does not report a prespecified decision threshold for acceptable reproducibility. These omissions limit assessment of how the planned methods were implemented and how the findings, if any, should be interpreted. The outcome results and analysis details are not reported in the registry record.

10. Confidence Intervals for Each Arm

No two-arm comparison is shown for this trial, but each arm's mean can be given with a t-based 95% confidence interval computed from the posted mean, spread and number of participants.

R code and output

arm <- c("Patients With Stage III-IV Head and Neck Cancer")
n   <- c(10)
m   <- c(1.49)
sp  <- c(0.16)
sd  <- sp      # posted standard deviations
for (i in seq_along(n)) {
  se <- sd[i] / sqrt(n[i]); h <- qt(0.975, n[i] - 1) * se
  cat(sprintf("CARM|%s|%d|%.3f|%.3f|%.3f|%.3f\n", arm[i], n[i], m[i], sd[i], m[i] - h, m[i] + h))
}

CARM  Patients With Stage III-IV Head and Neck Cancer  10  1.490  0.160  1.376  1.604

11. Multiplicity, Missing Data and Analysis Populations

The registry lists one primary outcome measure: change from baseline in 18F-EF5 SUV. It records no analyses. That information does not establish that no additional statistical tests were performed; it means that analysis results and details are not reported in the registry record.

Multiplicity arises when a study makes several statistical comparisons or evaluates multiple outcomes. In general, the more inferential questions considered, the greater the chance that at least one apparently unusual result could arise by chance. Methods for addressing multiplicity depend on the questions and analysis plan. The record does not report whether any adjustment was planned or used, whether there were additional exploratory analyses, or how the primary outcome was prioritized beyond its designation as primary.

The registry does not report a missing-data plan or the number of participants with usable paired scans. It therefore does not say whether incomplete scan pairs were excluded, handled by another method, or absent. Nor does it describe an analysis population, such as whether analyses were intended to include everyone enrolled or only participants with measurements from both scans. These details matter because reproducibility estimates based only on complete pairs may describe a subset if some participants lack usable repeat measurements.

The analysis plan, missing-data approach, and analysis population are not reported in the registry record. The endpoint description alone is not enough to determine whether the final analysis followed a prespecified plan or how uncertainty from incomplete data was handled.

12. Safety Results

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

Other (non-serious) adverse events

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

The registry lists 10 participants at risk for serious adverse events and 0 affected. For other adverse events, it also lists 10 at risk and 0 affected. No serious adverse-event terms are listed. These entries indicate that the registry reports no affected participants in either category for the group, but they should be read as registry entries rather than as a complete account of every possible event or safety assessment.

The deaths field is not reported in the registry record. That is distinct from a reported count of zero deaths. The record also does not give event timing, severity breakdowns beyond the serious-versus-other categories, or additional details about collection and reporting procedures. The absence of listed event terms does not establish that all forms of harm were systematically assessed or that no event occurred outside the reported categories.

These safety figures concern this study group and do not provide a comparison with a control group. The design has no separate comparator, so the figures cannot estimate a relative safety difference between groups. They also do not answer the primary reproducibility question, which concerns repeated SUV measurements. The registry's safety entries should therefore be interpreted narrowly: 0 affected participants are listed for serious and other adverse events, with deaths not reported in the registry record.

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

This study addresses a measurement question: whether repeated pretreatment PET/CT scans using 18F-EF5 produce reproducible information about intratumor distribution in head and neck cancer. A repeated-scan design can help characterize scan-to-scan variation under the conditions studied. The registry identifies the intended comparison, the continuous SUV endpoint, and a 7-day interval between baseline and the second scan.

However, the registry record does not report the endpoint results. It gives no observed SUV values, mean change, standard deviation, confidence interval, or Bland-Altman results. Therefore, it is not possible to determine from the posted record how similar the measurements were, how much variability occurred, or how precisely reproducibility was estimated. A completed study status does not supply those missing findings.

Even if numerical agreement results were available, reproducibility would be a measurement property, not by itself evidence that the tracer improves diagnosis, changes clinical decisions, or benefits patients. Those questions are different from whether repeated measurements agree. The registry describes no randomized comparison and no separate control group, so it cannot support a comparative treatment-effect interpretation.

The study's population is described as head and neck squamous cell carcinoma, with patients with stage III-IV head and neck cancer listed in the group description. Enrollment was 10. The registry does not report enough participant-level detail to assess how broadly any result might apply, and no outcome estimates are available to judge even the findings within this group.

The appropriate conclusion from the registry record is limited: a completed Phase 2, single-group study was intended to assess repeatability of 18F-EF5 PET/CT using repeated scans and SUV measurements. Its actual reproducibility results and their precision are not reported in the registry record. The posted safety entries list 0 affected participants for serious and other adverse events among 10 at risk in each category, while deaths are not reported.

14. Why This Trial Matters Statistically

Imaging measurements are useful for research only if their variation is understood. When an imaging measure is repeated, some observed change may reflect measurement variability rather than a real change in the underlying target. A study designed to examine repeat scans can therefore address an important statistical question before a measure is used to compare groups or assess change over time. This is a general rationale for reproducibility research.

The design described in the registry focuses on repeated pretreatment 18F-EF5 PET/CT scans in a single group, with SUV as the continuous endpoint. That makes the central comparison a within-participant measurement question rather than a treatment-effect question. The planned use of Bland-Altman plots is relevant in general because agreement is not captured fully by comparing group averages alone; the distribution of paired differences also matters.

The record also illustrates why a completed status is not equivalent to an interpretable result. Without reported estimates and their uncertainty, readers cannot assess the size of scan-to-scan variation or decide whether the measurement is sufficiently stable for a particular research purpose. Enrollment of 10 describes the study group size, but it does not reveal the estimate's precision without the outcome data.

Finally, the distinction between planning and evidence matters. Hypothetical sample-size illustrations do not replace an analysis of paired measurements and are not observed power. For reproducibility work, clear reporting of paired results, variation, uncertainty, missing scans, and analysis methods is central to judging what the study establishes. Those result and analysis details 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 most important limitation for interpreting this trial record is that it does not report the primary outcome results. Although the endpoint description states that baseline and second-scan SUV differences would be calculated and Bland-Altman plots generated, no values or plot findings are given. The magnitude and precision of scan-to-scan reproducibility therefore cannot be evaluated.

Enrollment was 10, but the record does not say how many participants completed both scans or contributed to the endpoint analysis. It does not report missing or unusable images, participant withdrawals, or the handling of incomplete scan pairs. Those details are especially relevant to a repeated-measure design because a paired comparison requires measurements from both scans for each included participant.

The registry record does not provide detailed image acquisition or analysis procedures, the statistical analysis plan, the analysis population, or the uncertainty estimates. It also does not state a criterion for what level of agreement would count as adequate for a specific use. Without these details, the intended methods can be described only at a high level, and the analysis cannot be independently assessed from the registry record.

The single-group design has no separate control group and is not randomized. It is directed at repeatability, not a comparative treatment effect. The record describes patients with stage III-IV head and neck cancer and head and neck squamous cell carcinoma, but does not provide further participant characteristics or outcome results with which to assess generalizability. The study's completed status does not resolve these reporting limitations.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Reproducibility of Intratumor Distribution of 18F-EF5 in Head and Neck Cancer — NCT01774760.
• Registry record last updated 2019-10-04; status last verified 2019-09.
• Results first posted 2019-10-04.
• 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

The registry describes a completed Phase 2, prospective, non-randomized, single-group study sponsored by Turku University Hospital. Enrollment was 10. The study began in 2013-03 and had primary completion in 2016-09.

The study concerned reproducibility of intratumor distribution of 18F-EF5 in head and neck cancer. Participants were to undergo pretreatment PET/CT scans two times, with the second scan 7 days after baseline. The primary outcome was change from baseline in 18F-EF5 SUV, measured in SUVmax. The record describes calculating scan-to-scan differences and generating Bland-Altman plots.

No primary outcome values or analysis results are reported in the registry record. Serious and other adverse-event entries each list 10 at risk and 0 affected; deaths are not reported.