Comparison of Combination Chemotherapy Regimens With or Without Cetuximab in Treating Patients Who Have Undergone Surgery For Stage III Colon Cancer: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Comparison of Combination Chemotherapy Regimens With or Without Cetuximab in Treating Patients Who Have Undergone Surgery For Stage III Colon Cancer (NCT00079274), a randomized phase 3 study. The primary outcome is Disease-free Survival (Arms A and D: Wild-type KRAS Patients).

2. Clinical Question

The study record describes a randomized phase III trial after surgery for stage III colon cancer. Its initial design addressed multiple chemotherapy combinations; later changes narrowed the treatment schedules relevant to the primary endpoint. The primary question is therefore specific to the wild-type KRAS subgroup and the Arm A and Arm D schedules, rather than a comparison across every treatment group in the overall trial. The registry record does not report a primary-endpoint sponsor analysis estimate or p-value in the posted results. The estimates and interval limits are reported for each arm, and an unadjusted comparison is also listed.

3. Trial Design

The registry identifies NCT00079274 as a completed, randomized phase III study with a parallel-group design and no masking. Enrollment is listed as 3397. The study concerned patients with stage III colon cancer and adenocarcinoma of the colon after surgery. The sponsor is the National Cancer Institute. The registry lists a start date of 2004-02 and primary completion in 2012-11.

The brief summary says the study was initially designed to compare three chemotherapy regimens. As of September 1, 2004, it was expanded to six arms: the original three schedules and three corresponding schedules that added cetuximab. The arm descriptions distinguish oxaliplatin-based treatment, irinotecan-containing schedules, and cetuximab combinations. The primary endpoint does not compare all of these schedules: it concerns wild-type KRAS patients treated according to Arm A or Arm D, and the endpoint description excludes patients who received irinotecan.

The registry summary reports that, as of 6/1/2005, patients no longer received irinotecan and arms B, C, E, and F were discontinued. Patients on arms B and C crossed to arm A, and patients on arms E and F crossed to arm D. The summary also describes the handling of patients in arms C and F who had not received irinotecan. It then begins to describe a change to KRAS prescreening as of 8/18/2008, but the text in the registry record is incomplete. Further details of that change are not reported in the registry record.

The posted primary endpoint groups contain 909.0 patients in wild-type KRAS Arm A and 954.0 in wild-type KRAS Arm D. These endpoint counts should not be conflated with the overall enrollment or adverse-event groups, which describe different populations. The registry record does not report the trial’s original sample-size justification, full protocol, or the reasons for every design change.

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.

Only the first 6 of 7 groups are shown.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: The analysis was performed using intention to treat principles. All patients that were wild-type KRAS and received randomized treatment according to the Arm A or Arm D intervention schedule were evaluated for this endpoint. Any patient receiving irinotecan was not included in the evaluation of this endpoint.

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.

Only the first 6 of 8 groups are shown.

6. Primary Endpoint

The registry lists the primary outcome as Disease-free Survival (Arms A and D: Wild-type KRAS Patients), measured over At 3 years.

The primary endpoint is disease-free survival at 3 years among wild-type KRAS patients treated according to the Arm A or Arm D schedule. The registry defines disease-free survival as the time from randomization until tumor recurrence or death, whichever occurs first. It states that estimates were made using the Kaplan–Meier method and that the result is expressed as a percentage of participants with the endpoint measured at 3 years.

Disease-free survival is useful in a study after curative resection because it captures whether recurrence or death has occurred during follow-up, rather than considering only one type of event. As a time-to-event endpoint, it also uses the timing of events and accommodates participants whose event status is not observed through the full follow-up period by accounting for censoring under the method’s assumptions. This is a general explanation of the method, not a claim about the trial’s unreported censoring patterns.

The registry identifies this endpoint as the study’s primary endpoint for the specified KRAS subgroup and treatment schedules. It reports 3-year estimates of 74.6% for Arm A and 71.5% for Arm D, with respective 95% confidence intervals of 71.1% to 78.3% and 67.8% to 75.4%. The endpoint description says the analysis followed intention-to-treat principles for the eligible randomized schedules, while also stating that patients receiving irinotecan were not included. The registry record does not report additional endpoint-handling details, such as the number censored or the follow-up distribution.

7. Results Posted in the Registry

• Wild-type KRAS Arm D: 71.5% (682/954) versus 74.6% (678/909) on control; risk difference -3.1 percentage points (95% CI -7.1 to 0.9), risk ratio 0.96 (0.91–1.01); unadjusted two-proportion z-test p = 0.132.

8. Secondary Efficacy Endpoints

The registry also posts results for secondary outcomes. For each outcome below, the endpoint type, the test applied to the posted summary statistics, the result and its interpretation are given in turn. Secondary endpoints are usually not adjusted for multiple testing, so they support rather than confirm the primary finding.

Disease-free Survival (Arms A and D: Mutant KRAS Patients)

• Endpoint type: Binary: the percentage of participants with the event (counts rebuilt from the posted percentages and group sizes).
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • Mutant KRAS Arm D: 65.0% (223/343) versus 67.1% (251/374) on control; risk difference -2.1 percentage points (95% CI -9.0 to 4.8), risk ratio 0.97 (0.87–1.08); unadjusted two-proportion z-test p = 0.553.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Mutant KRAS Arm D and the control. The interval gives the range of effects that remain compatible with the data.

Overall Survival as Measured by the 3-year Event-free Rate (Arms A and D: Wild-type KRAS Patients)

• Endpoint type: Binary: the percentage of participants with the event (counts rebuilt from the posted percentages and group sizes).
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • Wild-type KRAS Arm D: 85.6% (817/954) versus 87.3% (794/909) on control; risk difference -1.7 percentage points (95% CI -4.8 to 1.4), risk ratio 0.98 (0.95–1.02); unadjusted two-proportion z-test p = 0.281.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Wild-type KRAS Arm D and the control. The interval gives the range of effects that remain compatible with the data.

Overall Survival as Measured by the 3-year Event-free Rate (Arms A and D: Mutant KRAS Patients)

• Endpoint type: Binary: the percentage of participants with the event (counts rebuilt from the posted percentages and group sizes).
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • Mutant KRAS Arm D: 82.8% (284/343) versus 88.0% (329/374) on control; risk difference -5.2 percentage points (95% CI -10.3 to 0.0), risk ratio 0.94 (0.89–1.00); unadjusted two-proportion z-test p = 0.050.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Mutant KRAS Arm D and the control. The interval gives the range of effects that remain compatible with the data.

The secondary outcomes cover disease-free survival and overall survival for both wild-type and mutant KRAS groups, along with toxicity outcomes. Across the posted survival estimates, the Arm D estimates are lower than the corresponding Arm A estimates, but the secondary analyses do not all provide the same kind of comparison. The registry includes sponsor-reported Cox results for the two overall-survival outcomes, while the posted disease-free-survival contrasts are listed as unadjusted calculations. These different methods and reporting statuses should not be treated as interchangeable evidence.

The secondary results offer context for the primary disease-free-survival comparison, but they do not replace it. The primary result concerns 3-year disease-free survival in wild-type KRAS patients on the Arm A and Arm D schedules. Secondary outcomes use different endpoints or a different KRAS subgroup, and the overall-survival measures are described as 3-year event-free rates. A pattern across endpoints can inform the breadth of the evidence, but it does not by itself establish consistency or resolve uncertainty in the primary comparison.

Unadjusted secondary results generally carry less inferential weight than a clearly specified, prespecified analysis that accounts for the study’s testing plan. Multiple endpoints and subgroup analyses create more opportunities for chance findings. If several hypotheses are examined without an appropriate multiplicity strategy, nominal p-values can overstate the strength of evidence. The registry record does not report a multiplicity plan, so the degree to which the secondary comparisons were intended as confirmatory cannot be determined from the posted record.

The registry lists no reported analyses for the mutant KRAS disease-free-survival endpoint, while it does list sponsor Cox analyses for the overall-survival outcomes. This distinction should be kept visible when interpreting the collection of findings. Secondary outcomes are useful for describing what was measured and for identifying the scope of posted results; without a reported analysis plan and multiplicity handling, they should not be promoted to independent proof of the primary result.

Secondary outcomes registered for this trial, for reference (only those above were recalculated):

• Disease-free Survival (Arms A and D: Mutant KRAS Patients)
• Overall Survival as Measured by the 3-year Event-free Rate (Arms A and D: Wild-type KRAS Patients)
• Overall Survival as Measured by the 3-year Event-free Rate (Arms A and D: Mutant KRAS Patients)
• Toxicity, Assessed Using National Cancer Institute (NCI) Common Terminology Criteria for Adverse Events (CTCAE) Version 3.0 (v3) (Arms A and D: Wild-type KRAS Patients)
• Toxicity, Assessed Using National Cancer Institute (NCI) Common Terminology Criteria for Adverse Events (CTCAE) Version 3.0 (v3) (Arms A and D: Mutant KRAS Patients)

9. Statistical Interpretation

10. Statistical Methods Explained

Random allocation and the comparison

Randomization is intended to make treatment groups comparable, on average, with respect to factors that could influence outcome. This is a general property of randomized designs, not a guarantee that every measured or unmeasured factor is balanced in a particular trial. This registry identifies the study as randomized, parallel-group, and unmasked. The primary comparison is between wild-type KRAS patients following the Arm A and Arm D schedules, not between all of the trial’s original arms.

Kaplan–Meier estimation

The registry says disease-free survival was estimated by the Kaplan–Meier method. In general, this method estimates the probability of remaining free of the defined event over time while accounting for participants whose event time is not observed during follow-up. Such participants may contribute information until the point their follow-up ends; they are not automatically treated as having experienced the event. The method relies on assumptions about censoring and follow-up. The registry record does not report the numbers censored, the reasons for censoring, or detailed checks of those assumptions.

The primary result is reported as a percentage at 3 years, with a 95% confidence interval for each arm. Disease-free survival is defined as the time to recurrence or death, whichever comes first. The primary endpoint description identifies the analysis population as wild-type KRAS patients receiving randomized treatment according to Arm A or Arm D and says the analysis followed intention-to-treat principles. It also says anyone receiving irinotecan was not included in the evaluation. These population details matter when interpreting the arm-specific estimates.

Confidence intervals

A confidence interval communicates uncertainty around an estimate under the method used to construct it. In general, a wider interval indicates less precision, while a narrower interval indicates greater precision; neither interval guarantees that the underlying value lies within its limits. The primary arm estimates have 95% confidence intervals of 71.1% to 78.3% for Arm A and 67.8% to 75.4% for Arm D. The interval for the unadjusted difference, -0.071 to 0.009, includes zero. That means the comparison does not exclude a zero difference at the interval’s stated confidence level, without proving that the regimens are equivalent.

Unadjusted comparison

The registry lists an unadjusted risk-difference estimate for Arm D relative to Arm A, with a confidence interval and p-value. An unadjusted comparison uses the posted summary information without adjustment for additional covariates. It can be helpful for a simple descriptive contrast, but it is not automatically the trial’s planned or sponsor-reported inferential analysis. The primary endpoint has no sponsor analysis estimate listed in the registry record. The comparison therefore should be described as a recalculated unadjusted result, not as a reported adjusted treatment effect.

A p-value describes how incompatible the observed data are with a specified null model, under the assumptions of the statistical procedure. It is not the probability that the null hypothesis is true, nor does it measure the magnitude or clinical importance of an effect. The listed primary unadjusted p-value is 0.132. Its interpretation is limited by the unadjusted nature of the calculation and by the absence of a reported sponsor analysis for this endpoint.

Hazard ratios in the secondary outcomes

The registry reports Cox regression analyses for the two secondary overall-survival outcomes. A Cox model compares event rates over time through a hazard ratio and generally relies on a proportional-hazards assumption. The sponsor-reported hazard ratios are 1.25 for the wild-type KRAS analysis and 1.27 for the mutant KRAS analysis, with their reported confidence intervals and p-values. These should be distinguished from the posted 3-year event-free-rate estimates and from the unadjusted risk-difference calculations. The registry record does not report diagnostics for the Cox model assumptions or additional model details.

Over-dispersion and binary summaries

The primary display includes a percentage estimate and confidence interval rather than a conventional event-count model. In general, over-dispersion means that variation in observed data is greater than a model expects. It can make standard errors too small if ignored. The registry record does not describe an over-dispersion adjustment for these time-to-event estimates, and no such adjustment should be inferred. The general concept is relevant when considering count or proportion models, but it does not establish a problem with this trial’s reported analysis.

Planning calculations

The sample-size illustrations use the stated assumptions: two-sided significance level 0.05, equal allocation between the two arms, and a normal approximation. The effect to detect is set to the stated fraction of the observed effect. For time-to-event tables the values refer to total events; for other tables they refer to participants per arm. The trial’s original sample-size design assumptions are not reported in the registry record. Planning illustrations based on observed effects are not the same as the trial’s actual design calculation.

11. Confidence Intervals and Precision

The primary arm estimates are 74.6% for Arm A and 71.5% for Arm D. Their 95% confidence intervals, 71.1% to 78.3% and 67.8% to 75.4%, express uncertainty around each Kaplan–Meier estimate. The intervals overlap, but overlap alone is not a formal test of the difference between arms. The listed unadjusted comparison gives a more direct contrast: a risk-difference estimate of -0.031 with a 95% confidence interval from -0.071 to 0.009.

Because the difference interval includes zero, the posted unadjusted result does not rule out no difference under that interval procedure. It also allows values in either direction. The interval therefore indicates uncertainty about the direction and size of the contrast; it does not show that the regimens are equivalent. The listed p-value is 0.132, and should be read with the same caution about its unadjusted status.

Confidence intervals are about statistical precision, not clinical importance. A comparatively broad range allows a wider set of underlying values than a narrow range. The interval’s endpoints are not bounds on every possible value, and a 95% interval does not mean that 95% of patients have outcomes within it. It describes uncertainty in an estimated group-level quantity under the analysis method.

Precision also depends on the endpoint and analyzed population. The primary endpoint uses a time-to-event estimate for the wild-type KRAS subgroup and the Arm A and Arm D schedules. It is not interchangeable with the secondary event-free-rate summaries or with the adverse-event groups. The registry record does not report details such as censoring counts or follow-up completeness that would help assess precision beyond the posted intervals. The intervals should therefore be interpreted as the uncertainty quantified by the reported analysis, not as a complete account of every source of uncertainty.

12. Planning a Similar Trial: Sample Size

This is a planning illustration, not a re-evaluation of the original trial. It asks: if a new study wanted to detect a difference like the one observed, or a smaller one, how large would it need to be? It is deliberately not “observed power” of the finished trial, which is not a meaningful quantity.

The sample-size illustrations are planning calculations, not the enrollment target or the number actually analyzed in this trial. For the primary binary percentage endpoint, the listed figures are participants per arm, not a combined total. The sample-size assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. The effect to detect is set to the stated fraction of the observed effect. The original trial-design assumptions and sample-size justification are not reported in the registry record.

The 80% power column describes the planned number of participants per arm under which the calculation would have an 80% chance of detecting the specified effect if that effect and the stated assumptions held. The 90% power column describes the corresponding calculation for a 90% chance. These are prospective planning concepts, not statements about the probability that the trial’s result is correct. This is not observed power.

The rows that plan for 100% of the observed effect, 75% of the observed effect, and 50% of the observed effect represent progressively smaller target effects. A smaller effect is harder to distinguish from random variation, so the calculation requires more participants per arm. The numbers are illustrations based on the observed effect rather than a reconstruction of the study’s original design process. They should not be read as evidence that the trial was or was not adequately sized, because the registry record does not report the assumptions used for its original sample-size planning.

For these non-time-to-event calculations, the stated unit is participants per arm. The general sample-size description notes that time-to-event illustrations use total events instead. That distinction is important: participants enrolled and events observed are different quantities, and an event-based planning figure should not be interpreted as a per-arm participant count.

13. How the Sample Size Is Derived

Power is the probability that a study of a given size rejects the null hypothesis when a specific true effect exists. It depends on four things: the effect to detect, the variability of the outcome, the significance level (here a two-sided α of 0.05, so z_1−α/2 = 1.960) and the desired power (z_1−β = 0.842 for 80% and 1.282 for 90%).

Formula

n is the number of participants in each arm, p₀ and p₁ are the control and treatment proportions, and p̄ is their average.

Worked calculation for this trial

1. Control proportion p₀ = 0.746; treatment proportion p₁ = 0.715; average p̄ = 0.730.
2. First term: z_1−α/2 × √[2p̄(1−p̄)] = 1.96 × 0.6276 = 1.2300.
3. Second term: z_1−β × √[p₀(1−p₀) + p₁(1−p₁)] = 0.842 × 0.6272 = 0.5279.
4. n = (1.2300 + 0.5279)² / (-0.0310)² = 3,217.7, rounded up to 3,218 participants per arm for 80% power.

The smaller-effect rows in the table above repeat the same calculation with the effect multiplied by 0.75 and 0.5 (for ratios, applied to the logarithm of the ratio).

R code

alpha <- 0.05
z_a   <- qnorm(1 - alpha / 2)          # two-sided
p0 <- 0.745874587458746                          # control proportion
p1 <- 0.714884696016772                          # treatment proportion
n_per_arm <- function(power, frac = 1) {
  q1 <- p0 + (p1 - p0) * frac               # smaller effect = fraction of the observed difference
  pbar <- (p0 + q1) / 2
  z_b <- qnorm(power)
  ceiling(round((z_a * sqrt(2 * pbar * (1 - pbar)) + z_b * sqrt(p0 * (1 - p0) + q1 * (1 - q1)))^2 / (q1 - p0)^2, 6))
}
for (f in c(1, 0.75, 0.5))
  cat(sprintf("%3.0f%% of observed effect: n per arm = %s (80%% power), %s (90%% power)\n", 100 * f,
              format(n_per_arm(0.80, f), big.mark = ","), format(n_per_arm(0.90, f), big.mark = ",")))

# Check against R's built-in function (unrounded n per arm)
cat("power.prop.test:", round(power.prop.test(p1 = p0, p2 = p1, power = 0.80)$n, 1), "\n")

# Simulation: how often does the z-test reject at the 80%-power sample size?
set.seed(79274)
n    <- n_per_arm(0.80)
reps <- 20000
x0 <- rbinom(reps, n, p0); x1 <- rbinom(reps, n, p1)
ph <- (x0 + x1) / (2 * n)
z  <- ((x1 - x0) / n) / sqrt(ph * (1 - ph) * 2 / n)
cat(sprintf("Simulated power at n = %s per arm: %.3f (%d replicates)\n", format(n, big.mark = ","), mean(abs(z) > z_a, na.rm = TRUE), reps))

Output from running the code

100% of observed effect: n per arm = 3,218 (80% power), 4,308 (90% power)
 75% of observed effect: n per arm = 5,669 (80% power), 7,589 (90% power)
 50% of observed effect: n per arm = 12,637 (80% power), 16,917 (90% power)
power.prop.test: 3217.7 
Simulated power at n = 3,218 per arm: 0.801 (20000 replicates)

The simulation draws many pairs of event counts at the planned sample size and counts how often the two-proportion z-test rejects the null. It should land close to the target power; at very small required sample sizes it can fall below it, because the normal approximation is rough there.

14. Multiplicity, Missing Data and Analysis Populations

The record lists several outcome measures: the primary disease-free-survival endpoint in wild-type KRAS patients, additional disease-free-survival and overall-survival outcomes in KRAS subgroups, and toxicity outcomes. Examining multiple endpoints and subgroups raises a multiplicity issue. In general, the more statistical comparisons made, the greater the opportunity for at least one apparently notable result to occur by chance. A prespecified testing hierarchy or other multiplicity strategy can clarify how evidence should be assessed across comparisons.

A multiplicity adjustment or testing hierarchy is not reported in the registry record. The record also does not provide enough information to establish which secondary analyses were prespecified as confirmatory, how endpoints were prioritized, or whether any interim analyses were planned or performed. None of those details should be inferred from the listed outcomes or from the fact that the study was randomized.

The primary endpoint description says the analysis followed intention-to-treat principles for wild-type KRAS patients receiving randomized treatment according to Arm A or Arm D, and it says patients receiving irinotecan were excluded from this endpoint evaluation. These statements define important features of the reported primary analysis, but the registry record does not provide a complete participant flow, the number of missing endpoint observations, or detailed handling of missing data. It also does not report sensitivity analyses for missingness or censoring.

The study underwent treatment changes: the registry summary states that irinotecan was no longer used and arms B, C, E, and F were discontinued as of 6/1/2005, with specified crossovers. The summary also begins a description of KRAS prescreening changes as of 8/18/2008, but the text is incomplete. These changes make it important to distinguish the primary endpoint population and schedule from the broader trial enrollment. The original statistical analysis plan, sample-size rationale, interim-monitoring plan, and full details of the KRAS-related changes are not reported in the registry record.

15. 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 0% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: cardiac troponin i increased (4.8%), dyspnea (4.8%), cardiac pain (2.4%), left ventricular failure (2.4%), myocardial ischemia (2.4%).

The adverse-event record gives serious adverse-event counts and other adverse-event counts by treatment arm, with a group at risk shown for each. For Arm A, the record lists 93 serious events among 1378 at risk and 688 other events among 1378 at risk. For Arm D, it lists 150 serious events among 1326 at risk and 928 other events among 1326 at risk. These figures are counts, not percentages calculated for this interpretation. Arm B, C, E, and F figures are also listed in the registry, while Arm G has zero at risk and zero affected in these fields.

The record lists serious terms including cardiac troponin I increased and dyspnea at 4.8% each, and cardiac pain, left ventricular failure, myocardial ischemia, and confusion at 2.4% each. The registry does not clarify the denominator or the event-counting approach for these term percentages in this discussion available. They should be read as the listed serious-event terms, not as a complete account of all possible harms or as comparative rates between the primary endpoint groups.

Adverse-event groups are not identical to the wild-type KRAS populations used for the primary endpoint. The registry lists no death counts in these fields; deaths are not reported in the registry record. These summaries alone do not establish comparative safety or explain whether an event was caused by treatment.

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

The posted results describe group-level outcomes for a defined comparison: wild-type KRAS patients treated according to the Arm A or Arm D schedule after surgery for stage III colon cancer. The primary endpoint is disease-free survival at 3 years, defined by recurrence or death, whichever occurred first. The estimates and confidence intervals quantify what was recorded in those groups under the stated analysis approach.

The observed Arm A estimate is 74.6%, and the Arm D estimate is 71.5%. The listed unadjusted difference interval includes zero. Taken together, these figures show uncertainty around the size and direction of the difference; they do not establish equivalence, and they do not determine an outcome for an individual. The registry record does not report a sponsor analysis for the primary comparison, so the unadjusted comparison should not be treated as a substitute for an unavailable planned analysis.

The findings do not establish the effect in every patient with colon cancer, in other KRAS groups, or in treatment schedules outside the specified comparison. The trial record includes broader treatment changes and other arms, but the primary endpoint is not an all-arm comparison. Nor do the posted results alone establish why any difference might occur, whether the same pattern would appear in another study, or how results apply outside the trial’s population and follow-up.

Secondary findings add context, but they involve different outcomes and populations and include analyses with different reporting status. Because a multiplicity strategy is not reported, these results should not be treated as a set of independent confirmations. The safety summaries list event counts and terms, but they do not provide a complete basis for comparing risks or attributing events to treatment.

In general, clinical interpretation requires more than a point estimate or p-value: endpoint meaning, uncertainty, analysis population, study design, and limitations all matter. These results are statistical descriptions of the registry record and are not medical advice.

17. Why This Trial Matters Statistically

This trial illustrates how a large, multi-arm study can yield a primary statistical question that is narrower than its overall enrollment. The registry describes an initial comparison of chemotherapy schedules, an expansion to include cetuximab combinations, later discontinuation of irinotecan-containing arms, and a primary analysis restricted to wild-type KRAS patients following the Arm A or Arm D schedule. Clear separation of the enrolled population, treatment history, and endpoint population is essential to understanding what the comparison estimates.

It also highlights why subgroup definitions and endpoint definitions matter. The primary endpoint is time to recurrence or death, estimated at 3 years using Kaplan–Meier methods. Secondary overall-survival results are reported as event-free rates and also have Cox regression analyses. Those are related clinical outcomes but different statistical quantities. Reporting them side by side should not blur their distinct definitions or imply that one analysis substitutes for another.

The record further illustrates the value of distinguishing sponsor-reported analyses from recalculated unadjusted contrasts. The primary arm-specific estimates and intervals are posted, but a sponsor analysis for that primary comparison is not. Secondary Cox analyses are reported. The different evidence status should be explicit rather than concealed behind a single summary of p-values.

Finally, a registry record can be incomplete for statistical appraisal. Here, the original sample-size rationale, analysis plan, multiplicity handling, interim-analysis details, and full description of the KRAS-related change are not reported. Naming those gaps helps readers distinguish what the estimates show from what the available record cannot establish.

18. 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.
• Participant counts were reconstructed from posted percentages and arm sizes, so small rounding differences are possible.

The primary endpoint applies to a specific subgroup and regimen comparison, not the full enrolled population or every arm. The registry reports that patients receiving irinotecan were excluded from the primary endpoint evaluation, and its summary describes treatment-arm changes and crossovers. The KRAS prescreening description is incomplete, limiting the detail available about how the subgroup was operationalized over the course of the study.

The primary outcome is reported through Kaplan–Meier estimates, but the registry does not report numbers censored, detailed follow-up information, or sensitivity analyses. Those omissions limit assessment of the completeness of follow-up and the assumptions relevant to time-to-event estimation. The registry record also does not list a sponsor-reported primary comparison, so the unadjusted risk-difference calculation cannot be assumed to represent the trial’s planned inferential analysis.

The record does not report the original sample-size justification, the statistical analysis plan, a multiplicity strategy, or interim-analysis details. As a result, the confirmatory status of the secondary endpoints and the intended control of false-positive findings cannot be established from the registry record. Secondary results also span different populations and endpoints, and include both sponsor Cox analyses and unadjusted calculations.

Safety reporting is presented in arm-level event counts and serious-event terms, but the record does not provide enough detail to assess event attribution, term-specific denominators, or a comprehensive comparative safety profile. These limitations are specific to the information available in the registry record and do not imply that such details were absent from other study documentation.

19. Related Tutorials

20. Related Statistical Calculators

21. Sources

• ClinicalTrials.gov: Comparison of Combination Chemotherapy Regimens With or Without Cetuximab in Treating Patients Who Have Undergone Surgery For Stage III Colon Cancer — NCT00079274.
• Registry record last updated 2026-06-23; status last verified 2026-05.
• Results first posted 2015-02-09.
• 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.

22. Record Summary

NCT00079274 is a completed randomized phase III, parallel-group, unmasked trial sponsored by the National Cancer Institute. The registry lists enrollment of 3397, a start date of 2004-02, and primary completion in 2012-11. The study concerned patients with stage III colon cancer after surgery and included chemotherapy regimens with or without cetuximab.

The primary endpoint was 3-year disease-free survival in wild-type KRAS patients treated according to Arm A or Arm D. The posted estimates are 74.6% for Arm A and 71.5% for Arm D, each with a 95% confidence interval. The registry lists an unadjusted comparison, but no sponsor analysis for the primary endpoint. The record also includes secondary survival and toxicity outcomes, and describes substantial treatment changes during the study.

The original sample-size rationale, multiplicity approach, full analysis plan, and complete KRAS prescreening details are not reported in the registry record. These omissions limit how fully the posted analyses can be placed in the context of the study’s planned statistical framework.