Stereotactic Body Radiation Therapy in Treating Patients With Stage I or Stage II Non-Small Cell Lung Cancer That Can Be Removed By Surgery: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Stereotactic Body Radiation Therapy in Treating Patients With Stage I or Stage II Non-Small Cell Lung Cancer That Can Be Removed By Surgery (NCT00551369), a na phase 2 study. The primary outcome is Primary Tumor Control at 2 Years.

2. Clinical Question

3. Trial Design

The registry describes a phase II trial with a single-group model and no masking. SBRT was the listed intervention. No randomized allocation or comparator arm is listed, so the design does not provide a like-for-like control comparison. The study addressed patients with operable stage I or stage II non-small cell lung cancer; the registry record’s primary endpoint population is all eligible patients who started study treatment.

The registry lists enrollment of 33 and reports the study status as completed. The recorded start was 2007-12, and primary completion was 2012-05. These are registry-level study details; they do not establish how many participants contributed information to a particular outcome. For example, the adverse-event record has its own at-risk count, while the primary endpoint population is described separately.

The primary endpoint was primary tumor control at 2 years from the start of treatment. The record defines failure as in-field failure or marginal failure within 1.0 cm of the treatment field. Secondary outcomes included treatment-related grade 3 or 4 toxicity, other grade 3–5 adverse events, several failure and survival outcomes at 2 years, comorbidity burden, and predictive blood-marker assessments. The registry record does not report detailed methods for those secondary outcomes or their results.

A single-group phase II study can describe outcomes after the intervention and can assess them against a prespecified benchmark. It cannot, without a concurrent comparator, isolate the intervention’s effect from other explanations for observed outcomes. General interpretation therefore needs to distinguish the study’s benchmark-based primary test from a comparative treatment-effect estimate.

4. Participant Flow and Accounting

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

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: All eligible patients who started study 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.

Population: All eligible patients who started study treatment

6. Primary Endpoint

The registry lists the primary outcome as Primary Tumor Control at 2 Years, measured over From start of treatment to 2 years..

The primary endpoint was primary tumor control at 2 years, expressed as a percentage of patients. The timeframe ran from the start of treatment to 2 years. Control meant no primary tumor failure by that time. The registry definition counted either failure within the SBRT treatment fields or failure within 1.0 cm of the treatment field as primary tumor failure.

This endpoint focuses on whether the treated primary tumor remains free of the specified local failure events over a defined follow-up period. A percentage gives a clinically interpretable summary of how many patients in the endpoint population met the control definition, while the stated confidence interval would convey uncertainty around that estimate. In general, the interval is important because an observed percentage is an estimate from a finite group rather than a perfectly precise population value.

The registry record identifies all eligible patients who started study treatment as the endpoint population. It also describes a planned test using a hazard-based benchmark. No observed control percentage, confidence interval, hazard estimate, or test result is reported in the registry record. Consequently, the endpoint’s definition and planned testing approach can be described, but the study’s observed tumor-control result cannot be interpreted from this record.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The registry describes a one-sided Z-test comparing the logarithm of the observed hazard rate with the logarithm of the hypothesized hazard rate of 0.015. It specifies a one-sided type 1 error of 0.05 and statistical power of 90%. These values describe the planned inferential framework in the endpoint description; they are not observed study results.

What does the benchmark test ask?

The endpoint description calls a 90% tumor-control rate acceptable and a 70% rate unacceptable. In general, a benchmark test asks whether the observed data are sufficiently inconsistent with a stated reference value under a statistical model. Such a test is different from comparing two randomized groups. Here, the record describes a hazard-based test against the monthly hazard corresponding to the stated unacceptable benchmark. The exact statistical hypotheses and any decision rule beyond the description are not reported in the registry record.

A hazard is a rate of event occurrence over time among those who have not yet experienced the event, rather than a simple count of events divided by the total follow-up time. The endpoint is presented as a percentage of patients with control, but the planned test is described using a monthly hazard. That distinction matters: a time-to-event analysis can use information about when failures occur and may account for differing follow-up, whereas a simple percentage does not itself express event timing. The registry record does not provide the observed hazard or further details on how censoring and follow-up were handled.

Why use the logarithm of a hazard?

In general, hazards are positive quantities, and taking a logarithm can place comparisons on a scale that is convenient for statistical modeling. A difference between log hazards corresponds to a relative contrast on the original hazard scale. The registry specifically says the test compares the logarithm of the observed hazard with the logarithm of the hypothesized hazard. It does not report the observed value, the resulting test statistic, or a p-value, so the test’s result cannot be reconstructed from the record.

A one-sided test evaluates evidence in a prespecified direction rather than treating departures in both directions as equally relevant. Its interpretation depends on which direction is considered evidence against the null hypothesis. The endpoint text identifies the unacceptable benchmark and test specification, but the registry record does not add details about the alternative hypothesis or any deviations from the stated plan. A stated type 1 error of 0.05 is a design threshold governing the long-run false-positive rate under the test assumptions; it is not the probability that the observed conclusion is wrong.

What does the stated power mean?

In general, statistical power is the probability that a test rejects its null hypothesis under a specified alternative and assumptions. The endpoint description states 90% power for its planned test. This is not observed power, and it does not tell readers how much evidence the completed study produced. The registry record does not report a primary test result or the information needed to evaluate the achieved precision.

A confidence interval is a range calculated by a method designed to capture the underlying parameter at a stated rate over repeated applications, when the method’s assumptions hold. The endpoint dispersion is described as a 95% confidence interval. The registry record does not report the interval’s bounds. Without those bounds, readers cannot assess the precision of the estimated control percentage or whether the interval includes values consistent with the benchmark.

How should planning illustrations be understood?

General sample-size illustrations are planning exercises, not descriptions of what the trial observed and not measures of observed power. The planning assumptions described for illustrations include a two-sided significance level of 0.05, equal allocation between two arms, a normal approximation, and effects expressed as fractions of the observed effect. Time-to-event illustrations use a total event count; other illustrations are expressed as participants per arm. These are generic planning assumptions and do not change the actual single-group design of this trial.

The registry record does not report the observed effect needed to interpret illustrations based on fractions of that effect, nor does it identify those illustrations as the trial’s original design justification. The record also does not report a sample-size justification for the actual phase II study. Such illustrative planning values should not be read as observed enrollment requirements, as evidence of adequate precision, or as proof that the completed study had a particular probability of detecting an effect.

Finally, statistical tests and intervals rely on assumptions about outcome definition, follow-up, and the analysis population. The record states the primary population and test framework, but details on missing data, censoring, model diagnostics, and any alternative analyses are not reported in the registry record. Those omissions limit independent evaluation of the planned method and its application.

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("SBRT")
n      <- c(26)
events <- c(25)
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  SBRT  26  25  96.2  80.4  99.9  81.1  99.3  81.0  99.1

11. Multiplicity, Missing Data and Analysis Populations

The registry lists several secondary outcome families, including treatment-related grade 3 or 4 toxicity, other grade 3–5 adverse events, tumor failure categories, disease-free survival, overall survival, comorbidity burden, and predictive blood-marker assessments. Considering multiple outcomes can create a multiplicity issue: in general, when many statistical tests are conducted, the chance of at least one apparently unusual result can increase. Whether and how to control that issue depends on the analysis plan and the role of each outcome.

A multiplicity adjustment, hierarchical testing strategy, or designation of confirmatory versus exploratory analyses is not reported in the registry record. The record describes the primary endpoint’s one-sided test, but does not report a broader testing strategy for secondary outcomes. It also does not report outcome analysis results for these secondary measures.

The primary endpoint population is stated as all eligible patients who started study treatment. The registry record does not describe how missing endpoint assessments, incomplete follow-up, withdrawals, or censoring were handled. It does not report whether participants with unavailable outcome data were excluded, counted as failures, or addressed through another method. These details matter because missing information can affect both an estimated percentage and a time-to-event analysis.

Other analysis populations, such as a treated, evaluable, or per-protocol population, are not reported in the registry record. Nor are sensitivity analyses or procedures for handling missing data reported. These absences do not establish that such methods were not used; they mean the registry record does not document them. The stated analysis population and test framework can therefore be identified, but the full analysis process cannot be independently assessed from the record.

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.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: peripheral sensory neuropathy (3.8%).

The registry’s SBRT safety listing reports 1 participant with a serious adverse event among 26 at risk. It reports 19 participants with other adverse events among 26 at risk. The listing also identifies peripheral sensory neuropathy at 3.8% among the leading serious-event terms. These are separate posted safety figures; the record does not explain whether participants could appear in more than one category or how the term-specific percentage was calculated.

The counts describe events in the SBRT group, not a treatment comparison. There is no control group listed, so these figures cannot establish whether event frequency differs from another treatment or from background risk. The serious-event count should not be equated with the total number of event episodes, and the registry does not provide additional detail on severity, timing, attribution, duration, or resolution in this listing.

No death count is reported in the registry record. The absence of a reported count should not be interpreted as a count of zero. The record also does not explain the difference between the enrollment figure and the safety at-risk denominator, so those quantities should be kept distinct rather than reconciled by assumption.

In general, adverse-event summaries depend on definitions, observation time, and ascertainment procedures. Those details are not reported in the registry record for these figures. The posted counts therefore provide a limited descriptive account of safety events, not a complete estimate of comparative or long-term risk.

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

The record identifies a focused clinical question: primary tumor control through 2 years after SBRT among eligible patients with operable stage I or stage II non-small cell lung cancer who started treatment. It defines primary tumor failure using in-field failure and marginal failure within 1.0 cm of the treatment field. It also sets out acceptable and unacceptable tumor-control benchmarks and describes a planned one-sided hazard-based test.

Those design details do not reveal the result. The registry record does not report the observed primary tumor-control percentage, its 95% confidence interval, the observed hazard, or the primary test result. It is therefore not possible to determine from this record whether the study’s observed data met the stated benchmark, how precise the estimate was, or how strong the statistical evidence was.

Even if a benchmark test result were available, its interpretation would be limited by the single-group design. In general, outcomes after treatment can be described in a single group, but a causal treatment effect requires a credible comparison that addresses what would have happened without the intervention or under an alternative. This registry record lists no comparator arm, so it does not support a randomized estimate of comparative benefit or harm.

The listed safety figures offer descriptive information for the SBRT group, but they do not establish comparative safety. Their denominators and event categories should be interpreted as recorded, without assuming overlap or converting them into a treatment contrast. The record does not report a death count or details needed to characterize event timing and follow-up.

The appropriate conclusion is narrow: the registry documents a completed phase II, single-group SBRT study and its intended primary endpoint and test framework. The outcome and its precision are not reported in the registry record. The design and posted safety counts alone cannot establish treatment effectiveness, rule out important harms, or show how the findings compare with another approach.

14. Why This Trial Matters Statistically

This trial illustrates the distinction between a benchmark-based question and a comparative treatment question. The primary endpoint description sets an acceptable and an unacceptable tumor-control rate and specifies a one-sided test of a monthly hazard. That framework can ask whether results are inconsistent with a particular benchmark under stated assumptions. It is not equivalent to estimating a difference between SBRT and a control treatment.

The distinction is especially important in a single-group study. Without a concurrent comparator, outcome descriptions do not separate effects of treatment from patient characteristics, disease course, other care, or differences in follow-up. General statistical methods can quantify uncertainty around an observed outcome, but they cannot by themselves create the missing comparison.

The record also demonstrates why endpoint definitions and analysis populations matter. Primary tumor control depends on a specific failure definition and a 2-year timeframe, while the population is limited to eligible patients who started treatment. Safety counts use a separately recorded at-risk group. Keeping these definitions distinct prevents enrollment, endpoint analysis, and safety denominators from being treated as interchangeable.

Finally, planned significance levels and power are not results. They describe an intended testing framework, while observed estimates and confidence intervals are needed to assess what the completed study found and how precisely. In this case those primary results are not reported in the registry record. Statistical interpretation must therefore remain centered on the design and the limits of the available record, rather than inferring an outcome from the planned benchmarks or the study’s completed status.

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 trial-specific limitation in the registry record is the absence of the primary endpoint result. Although the endpoint, timeframe, population, benchmarks, and planned test are described, the observed control percentage, confidence interval, hazard estimate, and test result are not reported in the registry record. The primary question therefore cannot be answered from the posted information.

The single-group design also limits interpretation. No randomized comparator is listed, so the study cannot provide a direct, like-for-like estimate of how outcomes differ under another treatment. A benchmark test, even when fully reported, would not remove that design limitation.

The registry gives enrollment of 33, while the safety listing reports 26 at risk. The endpoint analysis population is described separately as eligible patients who started treatment. The record does not explain how these figures relate, how many contributed to each outcome, or how exclusions and incomplete follow-up were handled. Missing-data and censoring methods are not reported in the registry record.

Further limitations in interpretation arise because secondary outcome results and a broader multiplicity strategy are not reported. Details of outcome ascertainment, follow-up completeness, adverse-event observation, and analysis methods beyond the primary test description are also not reported in the registry record. These gaps restrict assessment of precision, robustness, and the possibility of selective or incomplete outcome reporting.

These are limitations of what the registry record documents; they do not establish that additional analyses or procedures were absent from the trial. They do mean that readers cannot independently evaluate those matters using the posted record alone.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Stereotactic Body Radiation Therapy in Treating Patients With Stage I or Stage II Non-Small Cell Lung Cancer That Can Be Removed By Surgery — NCT00551369.
• Registry record last updated 2019-04-30; status last verified 2019-03.
• Results first posted 2016-10-27.
• 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

This is a completed phase II, single-group study of SBRT in patients with operable stage I or stage II non-small cell lung cancer. The registry lists enrollment of 33, no masking, and no randomized allocation. The recorded study start was 2007-12, with primary completion in 2012-05.

The primary endpoint was primary tumor control at 2 years from treatment start, defined by absence of in-field or marginal primary tumor failure. The record describes benchmarks of 90% as acceptable and 70% as unacceptable, with a one-sided hazard-based test, but does not report the observed endpoint result or its confidence interval.

The safety listing reports 1 with a serious adverse event and 19 with other adverse events among 26 at risk in the SBRT group. Peripheral sensory neuropathy is listed at 3.8% among serious-event terms. The record does not provide a comparator or enough analysis detail to support comparative conclusions.