Study Evaluating the Safety and Efficacy of KTE-C19 in Adult Participants With Refractory Aggressive Non-Hodgkin Lymphoma: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Study Evaluating the Safety and Efficacy of KTE-C19 in Adult Participants With Refractory Aggressive Non-Hodgkin Lymphoma (NCT02348216), a non-randomized phase 1, phase 2 study. The primary outcome is Phase 2 Pivotal Study (Cohorts 1 and 2): Overall Response Rate (ORR) as Assessed by Investigator Per Revised International Working Group (IWG) Response Criteria for Malignant Lymphoma.

2. Clinical Question

The registry describes a study spanning phase 1 and phase 2, with distinct pivotal and safety-management cohorts. Its objectives differed across those parts: phase 1 evaluated safety of axicabtagene ciloleucel regimens, the pivotal phase 2 study evaluated efficacy, and the safety-management study assessed approaches intended to affect the rate and severity of cytokine release syndrome and neurologic toxicities. The primary endpoint addressed the pivotal efficacy question rather than combining every cohort into a single response estimate.

The endpoint is a response proportion, not a comparison against an untreated or alternative-treatment group. It describes how often complete or partial response was recorded in the defined analysis population. The registry record does not report the primary endpoint's observed percentage or confidence interval, so the response level and its precision cannot be described from the record excerpt. The fact that the total study enrollment was 307 does not establish how many participants were included in the primary endpoint analysis.

Some cohorts involved different lymphoma groups or distinct safety-management approaches, and the registry also lists retreatment groups. Those populations do not constitute randomized comparison arms. Their separate results, where reported, would answer cohort-specific questions and should not be interpreted as direct evidence that one regimen or cohort performed better than another.

3. Trial Design

This was a phase 1/2, multicenter study with non-randomized allocation, no masking, and a sequential model. The registry lists total enrollment as 307 and status as completed. The primary completion date is 2023-07-27. The study was sponsored by Kite, A Gilead Company.

The registry describes three broad study components: phase 1; the phase 2 pivotal study, comprising cohorts 1 and 2; and the phase 2 safety-management study, comprising cohorts 3, 4, 5, and 6. The phase 1 objective was to evaluate the safety of axicabtagene ciloleucel regimens. The pivotal phase 2 objective was to evaluate efficacy. The safety-management objective was to assess the impact of prophylactic regimens or earlier interventions on the rate and severity of cytokine release syndrome and neurologic toxicities.

In the pivotal cohorts, participants received conditioning chemotherapy with fludarabine and cyclophosphamide followed by a single infusion of axicabtagene ciloleucel. Cohort 1 enrolled participants with refractory DLBCL; cohort 2 was for refractory PMBCL or TFL. The other phase 2 cohorts covered different populations and management approaches, including relapsed or refractory disease and approaches involving optional bridging or debulking therapy. These design differences mean that cohort-level outcomes should be read in the context of each cohort's eligibility and treatment strategy.

Because allocation was not randomized and there was no masked control group, differences between cohorts cannot be treated as causal treatment comparisons. Participant characteristics, disease groups, treatment context, and cohort objectives may differ. The record does not report a randomized comparison that would separate treatment effects from those sources of difference.

Participants who received an infusion were to complete the remainder of the 15 year follow-up assessments in a separate long-term follow-up study. The primary endpoint timeframe was from the first infusion through the last follow-up visit, with a maximum duration of 7.7 years. The registry does not describe the full protocol rationale for the cohort sequence, nor does it report a sample-size justification for the trial.

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: 390 participants were screened. Participants who initially responded and subsequently relapsed, became eligible for second course of conditioning chemotherapy and axicabtagene ciloleucel. Participants received the same axicabtagene ciloleucel regimen as the original target dose.

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: DLT-Evaluable Analysis Set included participants treated in Phase 1 dosing cohort who received the target dose and were followed for at least 30 days after the axicabtagene ciloleucel infusion; or dose of anti-CD19 CAR T cells (axicabtagene ciloleucel) lower than the target for that cohort and experienced a DLT during the 30 day post-infusion period.

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: The Safety Analysis Set included all participants treated with any dose of axicabtagene ciloleucel.

Only the first 6 of 8 groups are shown.

6. Primary Endpoint

The registry lists the primary outcome as Phase 2 Pivotal Study (Cohorts 1 and 2): Overall Response Rate (ORR) as Assessed by Investigator Per Revised International Working Group (IWG) Response Criteria for Malignant Lymphoma, measured over First infusion date of axicabtagene ciloleucel up to last follow-up visit (maximum duration: 7.7 years).

The primary endpoint was overall response rate (ORR), expressed as the percentage of participants achieving either a complete response or a partial response. It applied to the phase 2 pivotal study, cohorts 1 and 2, and was assessed by study investigators according to revised International Working Group response criteria for malignant lymphoma.

Complete response required disappearance of detectable clinical evidence of disease and disease-related symptoms, regression of lymph nodes and nodal masses to normal size, a normal-size and non-palpable spleen and liver without nodules, and no evidence of disease in bone marrow aspirate and biopsy. Partial response required a substantial reduction in the summed products of diameters of designated dominant nodes or masses, without growth in nodes, liver, or spleen and without new disease sites. The criteria also specified reductions for splenic or hepatic nodules and single nodules.

ORR is a binary-outcome summary: each participant is classified as having achieved response or not under the stated criteria, and the result is expressed as a proportion. It provides a direct description of response frequency in the defined population, but it does not by itself describe how long a response lasts, whether disease later progresses, or overall survival. The registry lists duration of response, progression-free survival, and overall survival among secondary outcomes.

The analysis population included participants treated with at least 1.0 x 10^6 anti-CD19 CAR T cells/kg. The registry states that the 95% confidence interval was calculated using the Clopper-Pearson method. The observed ORR and its interval are not reported in the registry record.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

What kind of endpoint is ORR?

Overall response rate is a binary endpoint summarized as a proportion. Under the registry definition, participants either met the criteria for complete or partial response, or they did not. A proportion provides a compact description of the frequency of the specified outcome in an analysis population. It does not preserve all the detail in the underlying clinical assessments, and it does not measure the length of time a response persists.

Response classification depended on revised International Working Group criteria and was made by investigators. The criteria distinguish complete response from partial response using findings such as disappearance of detectable disease, changes in lymph nodes and organ involvement, bone marrow findings, and reductions in measured lesions. Combining complete and partial response into ORR gives a single response-frequency measure, while secondary outcomes such as duration of response can address a different question about persistence.

What does a confidence interval convey?

A confidence interval expresses statistical uncertainty around an estimated quantity. For a response proportion, it reflects the uncertainty associated with observing responses in a finite analysis population; it does not guarantee that the true value falls within the particular interval, and it is not a measure of the range of individual participants' outcomes. A wider interval generally indicates less precision than a narrower interval, although the width also depends on the observed proportion and the method used.

The registry specifies a 95% confidence interval using the Clopper-Pearson method. This is an exact binomial interval method commonly used for proportions. Its calculation is based on the number of participants with the outcome and the relevant analysis denominator. Since neither the numerical ORR nor its interval is reported in the registry record, the estimate's precision cannot be assessed from this record.

Why does the analysis population matter?

The endpoint population is the modified intent-to-treat set, defined as participants treated with at least 1.0 x 10^6 anti-CD19 CAR T cells/kg. This is not automatically identical to everyone enrolled, everyone assigned to a cohort, or everyone who began screening. Defining the population is important because the denominator determines the proportion and the precision of its estimate. The registry record does not report the size of this analysis set or explain the handling of participants without an evaluable response assessment.

Results from a population defined by treatment received can differ from results based on all enrolled participants. That difference does not, by itself, establish bias; interpretation depends on the reasons some participants may be outside the analysis set and how those participants' outcomes were handled. Those details are not reported in the registry record.

How should the cohort structure affect interpretation?

The study was non-randomized, unmasked, and sequential. The pivotal cohorts and safety-management cohorts addressed different populations or objectives, and the study also lists retreatment groups. In randomized trials, allocation can help make groups comparable on average. That protection is absent for comparisons across cohorts in this study. Differences in disease group, eligibility, treatment context, or timing could contribute to observed differences. The cohort data should therefore be interpreted as descriptions of their respective groups rather than as randomized estimates of comparative effect.

The primary endpoint was specific to pivotal cohorts 1 and 2, with investigator assessment. The registry also lists an ORR endpoint in the pivotal cohorts assessed by an independent radiological review committee as a secondary outcome. Different assessors or response criteria can produce different classifications; the two endpoint descriptions should not be treated as interchangeable. Numerical results for the independent review endpoint are not reported in the registry record.

What do planning illustrations mean?

Sample-size planning uses assumptions about the outcome, the effect one wants to detect, statistical error rates, allocation, and the calculation method. A planning illustration based on an observed effect is not a trial-design justification and is not observed power. The stated illustration assumptions are 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. For time-to-event outcomes, the illustration concerns total events, while other outcomes are expressed as participants per arm.

Those assumptions are illustrative rather than a description of the actual design of this study. This study was non-randomized, and no two-arm randomized comparison is reported. The planning assumptions should not be read as the protocol's prospective sample-size plan, evidence that the study had a particular power, or a basis for interpreting an unreported primary endpoint result. The registry does not report the trial's own sample-size justification.

Why are adverse-event counts not comparative risks?

Cohort-level adverse-event figures give counts of participants at risk and participants affected. Such counts can describe the recorded experience within a group, but comparing raw counts across groups is not a reliable causal comparison when group sizes and participant characteristics differ. A randomized comparator and a prespecified analysis would be needed to support a controlled estimate of relative risk. The registry lists neither a randomized control group nor an adjusted comparative safety analysis.

Serious adverse events and other adverse events are distinct categories in the record. A count of participants affected does not necessarily correspond to the number of events, because a participant may experience more than one event. The registry lists deaths separately by cohort; the counts alone do not establish cause or attribution. The record also lists event terms with percentages, but the denominator for those percentages is not reported in the registry record. They should not be read as a population-wide patient risk without that context.

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("Phase 2 (Pivotal Study): Cohort 1", "Phase 2 (Pivotal Study): Cohort 2")
n      <- c(77, 24)
events <- c(64, 20)
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  Phase 2 (Pivotal Study): Cohort 1  77  64  83.1  72.9  90.7  73.2  89.9  73.2  89.8
ARM  Phase 2 (Pivotal Study): Cohort 2  24  20  83.3  62.6  95.3  64.1  93.3  63.9  93.2

11. Multiplicity, Missing Data and Analysis Populations

The registry lists 28 outcome measures, including the primary ORR and secondary outcomes involving response, duration of response, progression-free survival, overall survival, and best overall response. Considering multiple outcomes can create a multiplicity issue: when many statistical questions are tested, the chance of at least one apparently notable result can increase if no adjustment or hierarchy is used. Whether and how this issue was addressed depends on the prespecified analysis plan. A multiplicity adjustment, testing hierarchy, or other control strategy is not reported in the registry record.

The registry gives the primary endpoint's confidence-interval method but does not report a statistical testing plan for the collection of outcomes. It also lists no analysis entries. That absence does not establish what analyses may have been conducted outside the registry; it means those details are not reported in the registry record. No interim-analysis plan is reported in the registry record.

Missing-data handling is also not reported in the registry record. In a response analysis, missing or unevaluable assessments can affect the denominator or classification, depending on the prespecified rules. The record defines the modified intent-to-treat population by treatment received, but does not state how missing response assessments, withdrawals, or other unavailable endpoint information were handled. The analysis-set size is not reported.

The listed primary population is not all participants enrolled; it includes those treated with at least 1.0 x 10^6 anti-CD19 CAR T cells/kg. The registry does not report the number meeting that definition or a participant flow accounting for inclusion in the analysis. These omissions limit independent appraisal of the denominator and the possible effect of missing outcomes. They also make it impossible to reconstruct the primary response estimate from the registry record.

12. Safety Results

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

Deaths

Other (non-serious) adverse events

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

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: myelodysplastic syndrome (100.0%), sepsis (50.0%), b-cell lymphoma (50.0%), headache (50.0%), respiratory failure (50.0%).

The serious adverse event figures are organized by study group. Each group has a count of participants at risk and a count affected by serious adverse events; the record separately lists participants affected by other adverse events and deaths. For example, phase 2 pivotal cohort 1 lists 77 participants at risk, 40 affected by serious adverse events, and 49 deaths. Pivotal cohort 2 lists 24 at risk, 14 affected by serious adverse events, and 13 deaths. These are cohort-level counts, not a controlled comparison.

Group sizes vary, and the registry includes small retreatment groups as well as the main study cohorts. A raw count should therefore be interpreted alongside its stated at-risk group, not compared across cohorts without accounting for the differences in group composition and design. Death counts do not establish that deaths were caused by the study treatment. The registry record does not report a randomized safety comparator or a causal attribution analysis.

Listed serious-event terms include Myelodysplastic syndrome (100.0%), Sepsis (50.0%), B-cell lymphoma (50.0%), Headache (50.0%), Respiratory failure (50.0%), and Pneumonia (28.6%). The denominator for these term percentages is not reported in the registry record. They should not be interpreted as percentages of all enrolled participants. Event counts and participant counts are also not necessarily the same, because a participant may have multiple events.

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

The primary endpoint was designed to describe the proportion of treated participants in the pivotal phase 2 analysis population who achieved a complete or partial response by investigator assessment. The registry provides a specific outcome definition and identifies the confidence-interval method. Those features clarify the intended meaning of the endpoint, but the numerical ORR and confidence interval are not reported in the registry record. The response result and its precision therefore cannot be stated.

Even if a response proportion is available, ORR is not equivalent to a direct comparison with another treatment. This study was non-randomized and the registry does not list a randomized control arm. The response proportion describes observed outcomes in the defined study population; it does not, by itself, establish what would have happened to similar participants under a different treatment or without treatment.

ORR also does not answer every question about disease course. It does not indicate how long responses lasted, whether participants experienced later progression, or how long they lived. The registry lists duration of response, progression-free survival, and overall survival as secondary outcomes, but numerical results for those outcomes are not reported in the registry record. The primary endpoint's follow-up timeframe extended from first infusion to last follow-up, with a maximum duration of 7.7 years; participants were to complete the remainder of the 15 year follow-up assessments in a separate long-term follow-up study.

The safety counts describe events recorded in the study groups, but do not by themselves show that treatment caused an event or death. The different cohorts are not exchangeable randomized arms. Their results may reflect differences in disease population, treatment context, or other factors not resolved by the design. The event-term percentages also lack a reported denominator in the registry record.

Accordingly, the registry supports understanding the study question, its response criteria, its cohort organization, and the type of safety information recorded. It does not support a numerical conclusion about the primary response outcome, a causal comparison against another treatment, or a complete assessment of long-term efficacy and safety from the information reported.

14. Why This Trial Matters Statistically

This trial is statistically notable for combining a pivotal efficacy component with multiple safety-management cohorts in a non-randomized, sequential design. That structure illustrates why it is important to match each result to the population and question it addresses. An efficacy endpoint in pivotal cohorts should not be conflated with outcomes from safety cohorts that had different objectives or treatment contexts.

The primary outcome is a clinically interpretable binary measure: response under stated lymphoma criteria. The registry specifies investigator assessment, the modified intent-to-treat population, and a Clopper-Pearson confidence interval. These details are important for understanding what a reported estimate would represent and how uncertainty was to be quantified. However, the numerical primary outcome and interval are not reported in the registry record, so the response magnitude and precision remain unknown from this record.

The design also highlights the distinction between descriptive outcomes and comparative evidence. A single-arm or non-randomized cohort can document outcomes in participants treated within that study context. Without randomized allocation to a comparable control group, differences between groups cannot be interpreted as an unbiased treatment contrast. This is particularly relevant when cohorts differ in disease category, eligibility, or safety-management strategy.

Finally, safety reporting requires attention to denominators and event definitions. Counts of participants affected, counts of deaths, and listed term percentages each have different meanings. Cohort-specific counts are not interchangeable with causal risk estimates, and percentages without a clear denominator are difficult to interpret. These statistical considerations are essential for reading complex, multi-cohort trial records without overstating what the data can establish.

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 registry record does not report the primary ORR estimate, its confidence interval, or the size of the modified intent-to-treat analysis set. As a result, the primary efficacy endpoint cannot be quantified from the reported information. Although total enrollment is 307, that figure does not identify the denominator used for the primary endpoint.

The non-randomized, sequential design has no listed randomized comparator and no masking. It therefore cannot provide a like-for-like randomized estimate of treatment effect against another intervention. Cohorts differ in their stated populations and objectives, which limits direct interpretation of cohort-to-cohort differences. Baseline characteristics and adjusted comparisons are not reported in the registry record.

Missing-outcome rules, participant flow into the primary analysis set, and methods for handling unevaluable response assessments are not reported in the registry record. Nor are a multiplicity strategy, interim-analysis plan, or the trial's own sample-size justification reported. These gaps limit assessment of the analysis decisions and the robustness of outcome interpretation.

The registry lists secondary outcomes, including duration of response, progression-free survival, overall survival, and independent review assessments, but their numerical results are not reported in the registry record. Adverse-event and death counts are given by cohort, without a randomized safety comparison. The denominator for listed serious-event term percentages is not reported in the registry record. In addition, long-term follow-up assessments were to continue in a separate study, so the primary trial record alone does not provide the remainder of that planned follow-up.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Study Evaluating the Safety and Efficacy of KTE-C19 in Adult Participants With Refractory Aggressive Non-Hodgkin Lymphoma — NCT02348216.
• Registry record last updated 2024-06-04; status last verified 2024-05.
• Results first posted 2021-11-23.
• 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

NCT02348216 was a completed, multicenter phase 1/2 study of axicabtagene ciloleucel in adults with refractory aggressive non-Hodgkin lymphoma. The registry lists enrollment of 307, non-randomized allocation, no masking, and a sequential model. It describes phase 1, phase 2 pivotal cohorts, and phase 2 safety-management cohorts.

The primary endpoint was investigator-assessed ORR in pivotal cohorts 1 and 2, among participants treated with at least 1.0 x 10^6 anti-CD19 CAR T cells/kg. ORR combined complete and partial response under revised IWG criteria; its 95% confidence interval was to use the Clopper-Pearson method. The numerical primary result and interval are not reported in the registry record.

Cohort-level serious adverse event and death counts are listed, but the non-randomized design does not support treating cohorts as comparative control groups. Numerical findings for the listed secondary outcomes are not reported in the registry record.