A Translational Study of Bevacizumab in Participants With Metastatic Colorectal Cancer: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for A Translational Study of Bevacizumab in Participants With Metastatic Colorectal Cancer (NCT01588990), a na phase 4 study. The primary outcome is Association Between Neutrophil to Lymphocyte Ratio (NLR) [NLR ≤5 Versus NLR >5] and Progression-Free Survival (PFS) as Assessed by Hazard Ratio.

2. Clinical Question

The question concerns a marker’s association with the time to progression or death during the treatment strategy, rather than a comparison of bevacizumab with another treatment. NLR was calculated as the ratio of neutrophils to lymphocytes measured in laboratory values. The primary analysis therefore asks whether the two baseline NLR categories had different PFS patterns in this study population. An association could be useful for describing prognosis, but this single-group design cannot establish that NLR caused a difference in PFS or that changing NLR would change outcomes.

3. Trial Design

This was a phase 4, open-label, prospective, multicenter, single-group study of bevacizumab-based treatment for participants with previously untreated metastatic colorectal cancer. The registry lists enrollment of 128 and a completed status. The study start date was 2012-06-26, and the primary completion date was 2016-09-15. The anticipated time on study treatment was 4 years.

Treatment was planned in two phases. In Phase A, participants received bevacizumab with either oral capecitabine plus infusional oxaliplatin (XELOX), or modified infusional 5-fluorouracil, leucovorin and oxaliplatin (mFOLFOX6). Phase A treatment continued until first disease progression. Participants then continued to Phase B, receiving infusional 5-fluorouracil, leucovorin and irinotecan (FOLFIRI) plus bevacizumab until second disease progression. This describes the treatment strategy in the registry summary; further details of how the chemotherapy option was selected are not reported in the registry record.

Allocation is listed as not applicable, and masking as none. There was no randomized control group. The primary analysis compares outcomes across baseline NLR categories within this treatment cohort, not outcomes across assigned treatments. Consequently, differences between the NLR categories may reflect characteristics associated with prognosis as well as any relationship with the marker itself.

The registry lists 22 outcome measures and 5 analyses. It names several secondary outcomes, including PFS through first and second progression, survival measures, duration of disease control, time to failure of strategy, and response in Phase A. The names of these outcomes do not provide their numerical results. A sample-size justification, interim-analysis plan and detailed 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.

Phase A

Reasons for not completing this period:

Phase B

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: FAS population. "Overall Number of Participants Analyzed" = participants who were evaluable for this outcome.

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: Full Analysis Set (FAS) included all participants who received at least 1 dose of bevacizumab.

6. Primary Endpoint

The registry lists the primary outcome as Association Between Neutrophil to Lymphocyte Ratio (NLR) [NLR ≤5 Versus NLR >5] and Progression-Free Survival (PFS) as Assessed by Hazard Ratio, measured over Baseline up to disease progression, death or end of study (up to 4 years).

The primary endpoint is the association between baseline neutrophil-to-lymphocyte ratio (NLR) and progression-free survival (PFS), reported as a hazard ratio comparing NLR of 5 or lower with NLR above 5. NLR is the ratio of neutrophils to lymphocytes calculated from laboratory values. PFS starts at initial treatment and ends at the first documented disease progression or death from any cause, whichever occurs first. The observation period is baseline through progression, death or study end, up to 4 years.

Progression was assessed by the investigator according to routine clinical practice. The registry describes progression as requiring unequivocal evidence, such as a clinically meaningful increase in known tumors, a new lesion, disease-related death without earlier objective documentation, elevated CEA with other radiological or clinical evidence, or symptomatic deterioration. These criteria make the endpoint a combined measure of disease progression and death, rather than a measure of tumor size alone.

General statistical context: a hazard ratio compares the rate at which events occur over follow-up between groups, taking account of event timing. It is used for time-to-event outcomes such as PFS. The value’s direction and precision depend on the group ordering and the estimate and confidence interval. The registry record identifies the comparison and intended measure, but does not report the HR estimate or its 95% confidence interval. The association’s magnitude and precision therefore cannot be described from the registry record.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

What does the hazard ratio describe?

General statistical context: time-to-event analysis uses both whether an event occurred and when it occurred during follow-up. For this study, the event in the primary endpoint was first disease progression or death, whichever came first. A hazard ratio compares the event rates over time between the NLR categories. It is not the same as a difference in median PFS, and it does not directly state how long an individual participant will remain free of progression.

The registry specifies an HR for NLR of 5 or lower versus NLR above 5. A complete interpretation would require the numerical estimate, its confidence interval, and clear identification of the reference category and numerator. Those numerical results are not reported in the registry record. It is therefore not possible to state which category had a higher estimated event rate, how large any association was, or how compatible the data were with no association.

How should a confidence interval be read?

General statistical context: a confidence interval gives a range of values compatible with the data and the statistical model at the stated confidence level. The primary outcome specifies a 95% confidence interval. Its width communicates precision: a relatively broad interval would leave more uncertainty about the association’s size than a relatively narrow interval. A confidence interval is not a guarantee that the true value lies within its endpoints, nor does it establish clinical importance by itself. Because the interval endpoints are not reported in the registry record, its precision cannot be judged for this outcome.

What does the PFS definition imply?

PFS is a composite time-to-event endpoint. An event can be triggered by progression or death from any cause, so it incorporates more than radiological tumor change. Progression was assessed by the investigator in routine clinical practice, using radiological, biochemical or clinical factors. The record describes criteria such as new lesions, meaningful growth of known tumors, or symptomatic deterioration, and also includes disease-related death without earlier objective documentation. This approach makes the endpoint relevant to the treatment course described, but assessment in routine practice can depend on the timing and availability of evaluations. The registry does not report further details about assessment schedules or independent review.

General statistical context: time-to-event methods can include participants who have not experienced the event by the end of their observed follow-up, accounting for the time they were observed. Such participants are commonly described as censored. The registry names the follow-up endpoint but does not report the censoring rules, the statistical model, or how participants with incomplete follow-up were handled. Those details are necessary to assess the implementation of the analysis.

What can be learned from grouping NLR?

The primary comparison uses two categories, NLR of 5 or lower and NLR above 5. Categorization produces a straightforward group comparison, but it does not describe how PFS might vary across the full range of NLR values. It also makes results dependent on the chosen threshold. The registry record identifies the categories but does not report whether the threshold was selected in advance, whether another form of NLR analysis was planned, or how values near the threshold were handled.

General statistical context: a marker’s prognostic association is distinct from a treatment-predictive effect. A prognostic marker is associated with outcome regardless of whether it identifies a different response to one treatment versus another. Demonstrating treatment prediction generally requires a treatment comparison and an assessment of whether treatment effects vary by marker. This single-group study has no randomized treatment comparator, so the primary NLR association is not a randomized test of treatment-effect differences.

What do the planning assumptions mean?

The registry’s planning description refers to a sample-size illustration based on an 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 described as a stated fraction of the observed effect. For time-to-event planning, the description refers to the total number of events. For other outcomes, it refers to participants per arm. These are planning conventions, not observed trial results, and a planning illustration is not observed power. The numerical illustration values and a sample-size justification for the trial are not reported in the registry record.

General statistical context: a sample-size calculation depends on assumptions about the effect to detect, variability or event frequency, analysis method, and error rates. A calculation based on an observed effect is retrospective in the sense that the observed effect informs the planning scenario; it does not establish the probability that the completed study detected a true effect. The registry’s description is not enough to reconstruct a calculation, and no new sample size can be inferred from it.

What remains uncertain about the analysis?

The registry identifies the primary measure and analysis population label but does not report the primary numerical result or enough detail to reproduce the analysis. The model form, covariate adjustment, handling of missing data, censoring approach, assessment of model assumptions and exact number analyzed are not reported in the registry record. These omissions limit evaluation of both the estimate and its robustness. The definitions of outcomes and planning conventions can explain the intended analysis, but they cannot replace the missing estimates and methods.

10. Multiplicity, Missing Data and Analysis Populations

The registry lists 22 outcome measures and 5 analyses, including the primary NLR–PFS association and secondary measures of progression, survival, disease control and response. Multiple outcomes or analyses can create multiple opportunities to observe apparently noteworthy results. General statistical context: multiplicity procedures can control the overall chance of false-positive findings across a family of tests, but they can also affect the interpretation and precision of individual results. Whether such procedures were planned or used is not reported in the registry record. The number of outcomes and analyses does not itself show that multiplicity adjustment was or was not applied.

The primary outcome names the FAS population and defines the analyzed number as participants evaluable for that outcome. The record does not provide a numerical primary analysis denominator or define the FAS beyond that description. It also does not report how participants who lacked an evaluable NLR measurement, progression assessment or follow-up were handled. Missing-data rules, imputation methods, censoring procedures and sensitivity analyses are not reported in the registry record.

The registry record does not report a detailed statistical analysis plan, sample-size justification or interim-analysis plan. It is therefore not possible to determine from the record whether the analysis population, NLR categorization, model, or multiplicity approach was finalized before outcomes were examined. These are limits on interpretation, not evidence that a particular method was or was not used. Enrollment of 128 is stated, but enrollment alone does not reveal the number contributing to each outcome analysis.

11. 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: pulmonary embolism (10.2%), pyrexia (9.4%), vomiting (7.8%), diarrhoea (6.2%), sepsis (5.5%).

The adverse-event summary lists 128 participants at risk for serious adverse events and 86 affected. It also lists 128 at risk for other adverse events and 128 affected. These counts describe the registry’s safety summary; they do not provide the primary NLR–PFS result. The deaths entry is not reported in the registry record.

The listed serious adverse-event terms and percentages are pulmonary embolism, 10.2%; pyrexia, 9.4%; vomiting, 7.8%; diarrhoea, 6.2%; sepsis, 5.5%; and dehydration, 5.5%. These are the selected terms presented in the record, not a complete description of every safety outcome. The record does not give further details about timing, duration, severity grades, participant-level overlap between terms, or attribution to treatment. The percentages should therefore be read as registry-reported event figures, not as proof that treatment caused the events.

General statistical context: adverse-event counts depend on how events are defined, collected and attributed, and multiple events may occur in the same participant. The summary does not establish how many participants experienced more than one listed event. Counts and percentages also do not, by themselves, compare safety with another treatment, because this study had no separate control group.

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

The study’s primary statistical question is whether baseline NLR category was associated with PFS among participants receiving a bevacizumab-based treatment strategy. That is a prognostic association question within a single-group cohort. The registry record specifies the two NLR categories, the PFS definition and the intended HR measure, but does not report the HR or 95% confidence interval. Consequently, it does not permit a conclusion about the direction, magnitude or precision of the primary association.

The design also does not test bevacizumab against a control treatment. There was no randomized comparator, and the treatment course included chemotherapy combinations that changed between Phase A and Phase B. The record’s description of treatment does not establish that all participants received an identical regimen at every point. Any association between NLR and PFS would not, on its own, demonstrate a causal effect of NLR or show that one treatment is better than another.

General statistical context: a marker can be associated with outcome because it reflects other participant or disease characteristics. Without a reported adjustment strategy, it is not possible to determine whether the primary analysis attempted to account for such factors. And even an adjusted association would not by itself prove causality.

The registry lists secondary outcomes and gives a safety summary, but the outcome names alone are not evidence of the results. Serious adverse-event counts and selected term percentages describe reported safety events; they cannot establish comparative risk in the absence of a control group. The numerical primary efficacy result, secondary outcome estimates, detailed missing-data methods and multiplicity approach are not reported in the registry record.

The most defensible interpretation is therefore about study intent and design: this prospective, open-label study was structured to examine NLR as a prognostic marker during bevacizumab-based treatment. The posted registry information does not provide the numerical primary result required to judge whether an association was observed or how precisely it was estimated. That limitation should not be replaced with an inference from the endpoint definition, enrollment count, or safety figures.

13. Why This Trial Matters Statistically

This trial is statistically notable for asking a marker-based prognostic question in a prospective treatment cohort rather than comparing randomized treatment groups. It links a baseline laboratory ratio, NLR, to a time-to-event outcome, PFS, defined by progression or death. This design makes the marker association the central question while treatment is described as a two-phase strategy involving bevacizumab and chemotherapy combinations.

The distinction between prognosis and treatment effect is important. A marker may help describe how outcomes vary across participants, but an association in a single-group study cannot establish that the marker changes treatment benefit. A randomized treatment comparison would address a different question. The absence of a control group is not simply a missing detail: it defines what causal comparisons this design can support.

The study also illustrates why outcome definitions and numerical results need to be read together. The record describes routine-practice progression assessment and specifies a 95% confidence interval for the HR, yet the HR and interval are not reported in the registry record. Knowing what was intended for analysis is valuable, but it does not reveal the estimated association or its precision.

Finally, the registry lists many outcomes and analyses while omitting details of the analysis plan, missing-data handling and multiplicity approach. These gaps matter when considering how stable and selective any reported finding might be. The available information supports a careful account of the research question, design and safety summary, but not a numerical conclusion about the primary association.

14. 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 primary statistical result is not reported in the registry record: there is no HR estimate or 95% confidence interval for the NLR comparison. Without these, the association’s direction, size and precision cannot be evaluated. Numerical results for the listed secondary outcomes are also not reported in the registry record.

The study is single-group and open-label, with no randomized control arm. The NLR categories therefore do not constitute treatment groups, and the design cannot isolate a causal effect of NLR or compare bevacizumab-based treatment with an alternative. Potential confounding is especially relevant to a prognostic association, but the record does not report the covariates or adjustment methods, if any, used in the analysis.

The primary population is identified as FAS and as evaluable participants, but the number analyzed is not stated. The record does not describe how missing baseline laboratory values, incomplete follow-up, or unevaluable progression assessments were handled. It also does not report the statistical model, censoring rules, assessment schedule, model diagnostics, or sensitivity analyses. The extent to which investigator assessments followed a consistent process across participants is not described beyond routine clinical practice.

The registry lists 22 outcome measures and 5 analyses, but does not report whether multiplicity was addressed or provide a detailed statistical analysis plan. A sample-size justification, interim-analysis plan and numerical planning illustration are not reported in the registry record. Finally, the selected adverse-event terms do not provide full details about event timing, attribution, severity or overlap. These limitations restrict interpretation of the numerical evidence and should not be filled by assumptions about the protocol or analyses.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: A Translational Study of Bevacizumab in Participants With Metastatic Colorectal Cancer — NCT01588990.
• Registry record last updated 2019-01-22; status last verified 2018-08.
• Results first posted 2019-01-22.
• 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.

18. Record Summary

NCT01588990 is a completed phase 4, prospective, open-label, multicenter, single-group study titled “A Translational Study of Bevacizumab in Participants With Metastatic Colorectal Cancer.” The registry lists Hoffmann-La Roche as sponsor and enrollment of 128. The study began on 2012-06-26, with primary completion on 2016-09-15.

The treatment strategy comprised Phase A bevacizumab with XELOX or mFOLFOX6 until first progression, followed by Phase B FOLFIRI plus bevacizumab until second progression. The primary outcome was the association between baseline NLR category and PFS, expressed as an HR with a 95% confidence interval. The HR and interval are not reported in the registry record. The adverse-event summary lists 86 participants affected by serious adverse events and 128 affected by other adverse events; a death count is not reported in the registry record.