← Clinical Trials
Multiple Myeloma Phase 1/2 CAR-T Therapy NCT03548207

CARTITUDE-1: Complete Statistical Analysis of JNJ-68284528 in Multiple Myeloma

An independent statistical analysis of CARTITUDE-1, a phase 1/2 single-group study of JNJ-68284528, a chimeric antigen receptor T cell therapy directed against B-cell maturation antigen, in participants with relapsed or refractory multiple myeloma.

Study period: 2018-06-29 to 2022-08-23  ·  Status: Completed  ·  Industry-sponsored
Scope of this record

This page provides an independent statistical analysis and educational interpretation of publicly reported results. ClinicalTrials.gov provides the official trial registry record.

1. Trial at a Glance

CARTITUDE-1 was a completed phase 1/2, single-group study evaluating JNJ-68284528 in participants with relapsed or refractory multiple myeloma. The registry reports 126 enrolled participants and one intervention: JNJ-68284528, a biological intervention.

126
Enrolled
Participants
1
Study arm
Single-group design
1/2
Phase
Phase 1/2
45.2
Months
Maximum primary-endpoint time frame
FeatureCARTITUDE-1
PhasePhase 1/2
ConditionMultiple Myeloma
PopulationParticipants with relapsed or refractory multiple myeloma
Design modelSingle group
MaskingNone
AllocationNA
Primary purposeTreatment
Enrollment126
Number of arms1
InterventionJNJ-68284528 (biological)
StatusCompleted
Lead sponsorJanssen Research & Development, LLC
Sponsor typeIndustry
Start date2018-06-29
Primary completion date2022-08-23
ClinicalTrials.govNCT03548207

2. Clinical Question

The central clinical-statistical question was how participants with relapsed or refractory multiple myeloma responded to treatment with JNJ-68284528 and what adverse events occurred during the study period.

Population

Participants with relapsed or refractory multiple myeloma.

Intervention

JNJ-68284528, a chimeric antigen receptor T cell (CAR-T) therapy directed against B-cell maturation antigen (BCMA).

Comparator

The registry identifies a single-group design and one study arm; no comparator arm is reported.

Primary questions

What proportion of participants achieved a partial response or better according to IMWG criteria, and what adverse events occurred by severity?

3. Trial Design

01
Enroll 126 participants
02
Single group One study arm
03
Intervention JNJ-68284528
04
Assess Response and adverse events
05
Follow-up Through 45.2 months
Design model
Single group
No randomized comparator is reported.
Masking
None
The registry reports an unmasked study.
Allocation
NA
A comparative allocation scheme is not applicable to the single-group design.
Primary purpose
Treatment
The registry classifies the primary purpose as treatment.
SINGLE STUDY GROUP · 126 ENROLLED

JNJ-68284528

  • Biological intervention
  • Chimeric antigen receptor T cell (CAR-T) therapy
  • Directed against B-cell maturation antigen (BCMA)
COMPARATOR

No comparator arm

  • The design model is single group.
  • The registry reports one study arm.
  • No randomized control group is reported.
Why the single-group design matters. A single-group trial can directly describe outcomes observed after treatment, such as the proportion achieving a specified response. It cannot, from the randomized comparison framework used in a parallel-group trial, estimate a treatment-versus-control effect because there is no concurrent comparator arm.

4. Endpoints

Primary endpointTime frameRegistry definition
Phase 1b: Number of Participants With Adverse Events as Per Severity Day 1 up to 45.2 months An adverse event was any untoward medical occurrence in a clinical study participant administered a pharmaceutical product. Severity was graded according to NCI-CTCAE version 5.0, ranging from Grade 1 (Mild) to Grade 5 (Death).
Phase 2: Overall Response Rate (ORR) Day 1 up to 45.2 months ORR was the percentage of participants who achieved partial response (PR) or better according to international myeloma working group (IMWG) criteria.

How the ORR definition works

The registry defines a partial response as a reduction of serum M-protein of greater than or equal to 50% together with reduction in 24-hour urinary M-protein by greater than or equal to 90% or to less than 200 mg per 24 hours. When serum and urine M-protein were not measurable, the criteria could instead use the difference between involved and uninvolved free light chain levels. If those measurements were also not measurable, a reduction in bone marrow plasma cells could be used when the baseline bone marrow plasma cell percentage was greater than or equal to 30%. When soft tissue plasmacytomas were present at baseline, a greater than or equal to 50% reduction in their size was additionally required.

Endpoint interpretation: ORR is a proportion, not a time-to-event measure. It answers how many participants achieved the prespecified response threshold during the defined assessment period. It does not by itself describe how long responses lasted or whether treatment changed survival.

5. Statistical Methodology

Descriptive analysis of adverse events

The phase 1b primary endpoint is the number of participants with adverse events according to severity. Because the endpoint is defined by counts of participants within severity categories, an appropriate analysis is descriptive tabulation of participants by adverse-event severity using the NCI-CTCAE version 5.0 grading framework reported in the registry.

Severity scale
Grade 1 = Mild  ·  Grade 2 = Moderate  ·  Grade 3 = Severe  ·  Grade 4 = Life-threatening  ·  Grade 5 = Death

The grade is a severity classification. An adverse event does not necessarily have a causal relationship with the intervention.

Proportion-based analysis of ORR

The phase 2 ORR endpoint is a binary response outcome summarized as the percentage of participants achieving PR or better according to IMWG criteria. For a single-group study, the natural statistical summary is the observed response proportion together with an appropriate confidence interval for that proportion.

Conceptual response proportion
ORR = responders / evaluable participants

The numerator represents participants meeting the prespecified PR-or-better definition. The denominator must correspond to the analysis population specified for the endpoint.

No treatment-effect comparison

Because CARTITUDE-1 is a single-group study, there is no randomized treatment-versus-control contrast. Consequently, a hazard ratio, risk ratio, odds ratio comparing randomized groups, or between-arm p-value is not an inherent feature of this design.

Statistical distinction: a response percentage in a single-group study is an estimate of the observed response probability under the study's eligibility, treatment, assessment, and follow-up conditions. It should not be described as a relative treatment effect because there is no concurrent randomized comparator.

6. Planned Analysis

The registry reports results for the primary endpoints but does not provide posted statistical analyses in the ClinicalTrials.gov record. For the phase 1b safety endpoint, the relevant analysis would normally summarize the number of participants experiencing adverse events within each NCI-CTCAE severity grade over the specified Day 1 up to 45.2 months time frame. For the phase 2 ORR endpoint, the relevant analysis would normally estimate the proportion of participants achieving PR or better according to IMWG criteria, accompanied by a confidence interval describing statistical precision.

Phase 1b endpoint

Adverse events are categorized by severity using NCI-CTCAE version 5.0. The primary descriptive quantity is the number of affected participants in each severity category.

Phase 2 endpoint

ORR is defined as the percentage of participants achieving PR or better according to the detailed IMWG response criteria.

Confidence intervals

For a single-group response proportion, a binomial confidence interval is a natural way to quantify uncertainty around the observed proportion.

Interpretation

The analysis describes the observed experience of the treated study population rather than a randomized treatment effect.

The registry does not report a posted formal statistical analysis for the primary endpoints. The appropriate interpretation therefore centers on the endpoint definitions, study design, observed counts or proportions when reported, and the uncertainty appropriate to a single-group analysis.

7. Statistical Methods Explained

Why does a single-group trial change the statistical interpretation?

In a randomized comparative trial, the difference between treatment groups provides the basis for estimating a comparative treatment effect. CARTITUDE-1 has a single-group design, so there is no concurrent randomized comparator. The primary statistical task is therefore estimation and description of outcomes in the treated study population rather than estimation of a randomized between-group effect.

What does ORR measure?

ORR measures the proportion of participants who achieve a partial response or better under the prespecified IMWG criteria. It is a categorical efficacy endpoint: each participant is classified according to whether the response threshold was achieved during the defined assessment period.

Why does the definition of PR matter statistically?

The response definition determines who enters the numerator of the ORR. Here, the registry uses a hierarchy of measurements involving serum and urinary M-protein, free light chains, bone marrow plasma cells, and soft tissue plasmacytomas when applicable. A response proportion is only meaningful when the classification rule is applied consistently across participants.

What does an ORR confidence interval tell us?

A confidence interval around a single-group response proportion describes statistical uncertainty about the underlying response probability under the model and sampling framework used to construct the interval. It does not describe the range of responses that individual participants will experience.

Why is an adverse-event count different from an ORR?

Adverse-event severity is organized into clinical grades, while ORR reduces response status to whether a participant meets the PR-or-better criterion. The two endpoints therefore answer different questions: one characterizes safety events by severity, while the other summarizes antitumor response.

Why is there no hazard ratio for the primary ORR endpoint?

A hazard ratio is primarily a relative measure for time-to-event outcomes. ORR is a proportion measured over a defined period. In a single-group study, there is also no second randomized group against which a comparative hazard ratio could be calculated.

What does "adverse event" mean?

The registry defines an adverse event as any untoward medical occurrence in a participant administered a pharmaceutical product. Importantly, the definition does not require that the event be causally related to the intervention. Severity is separately graded from Grade 1 through Grade 5.

8. Safety

The registry reports serious adverse events by study population as the number affected divided by the number at risk. These figures span the Phase 1b US population, Phase 2 US population, and Phase 2 Japan population.

Study populationSerious adverse eventsAffected / at risk
Phase 1b (US Population)Participants with serious adverse events11/29
Phase 2 (US Population)Participants with serious adverse events42/68
Phase 2 (Japan Population)Participants with serious adverse events1/9
Serious adverse events · affected / at risk
Phase 1b US
11/29
Phase 2 US
42/68
Phase 2 Japan
1/9
Important denominator issue: the serious-adverse-event figures use population-specific denominators of 29, 68, and 9. They should not be interpreted as though all three rows represent mutually exclusive randomized treatment arms. They are population-specific safety summaries within a single-group study.

Adverse-event severity as a statistical endpoint

The phase 1b primary endpoint is explicitly defined as the number of participants with adverse events according to severity. The registry specifies five severity grades: Grade 1, Mild; Grade 2, Moderate; Grade 3, Severe; Grade 4, Life-threatening; and Grade 5, Death related to the adverse event.

For safety interpretation, the distinction between frequency and severity is important. The number of participants experiencing an event and the grade assigned to that event provide different information. A participant-level count also differs conceptually from a count of the total number of events, because one participant can experience more than one adverse event.

9. Clinical Interpretation vs Statistical Interpretation

Statistical interpretation

The primary phase 2 efficacy endpoint is a single-group response proportion defined by IMWG criteria. The phase 1b primary endpoint is a participant count classified by adverse-event severity.

Clinical interpretation

The endpoints describe whether participants achieved a prespecified response and what adverse events occurred during follow-up. Neither endpoint alone summarizes every dimension of clinical benefit or risk.

The distinction is especially important because a single-group ORR does not establish how the intervention compares with another treatment. Historical comparisons may be informative in some research settings, but they are vulnerable to differences in eligibility criteria, disease characteristics, response assessment, follow-up, supportive care, and other factors. A concurrent randomized comparator provides a stronger basis for attributing differences to treatment assignment.

10. Interpreting a Single-Group Response Rate

What an ORR estimate means

An ORR estimate represents the observed proportion of participants who achieved PR or better according to the prespecified IMWG criteria during the defined assessment period. It is a descriptive estimate of response in the study population.

What ORR does not mean

ORR does not mean that the same proportion of all patients with relapsed or refractory multiple myeloma would necessarily respond outside the study. It also does not measure the duration of response, overall survival, or a randomized treatment effect.

Why the confidence interval matters

Because ORR is estimated from a finite number of participants, the observed percentage is subject to sampling uncertainty. A confidence interval quantifies that uncertainty; it does not represent a prediction interval for individual patients.

Why the p-value is not the effect size

If a hypothesis test were performed for a single-group response rate, the p-value would quantify evidence against a specified null hypothesis under the chosen testing framework. It would not measure the magnitude of the response itself. The response proportion remains the direct measure of the observed effect.

11. Understanding the IMWG Response Definition

The registry's PR definition is not simply a general clinical judgment that a tumor has improved. It is a structured classification rule based on specified laboratory and disease measurements.

Measurement pathwayRegistry criterion for PR
Serum M-protein and urinary M-protein measurableGreater than or equal to 50% reduction of serum M-protein and reduction in 24-hour urinary M-protein by greater than or equal to 90% or to less than 200 mg per 24 hours.
Serum and urine M-protein not measurableDecrease of greater than or equal to 50% in the difference between involved and uninvolved free light chain levels.
Serum and urine M-protein and serum free light assay not measurableGreater than or equal to 50% reduction in bone marrow plasma cells, provided baseline bone marrow plasma cell percentage was greater than or equal to 30%.
Soft tissue plasmacytomas present at baselineGreater than or equal to 50% reduction in the size of soft tissue plasmacytomas was additionally required.

This hierarchy illustrates why endpoint definitions matter in clinical-trial statistics. The response classification is constructed from clinically relevant measurements, but the measurement pathway can differ depending on what is measurable at baseline. Statistical analysis begins only after those clinical classification rules have been applied consistently.

12. Analysis Populations and Denominators

The registry reports 126 enrolled participants overall, while the serious-adverse-event summaries use population-specific denominators. Those denominators should be preserved when interpreting the safety results.

QuantityReported valueStatistical significance
Total enrollment126Overall study population size
Number of arms1Single-group design
Phase 1b US safety denominator29Population-specific denominator
Phase 2 US safety denominator68Population-specific denominator
Phase 2 Japan safety denominator9Population-specific denominator

Differences between an overall enrollment count and endpoint-specific denominators are not automatically errors. Clinical-trial endpoints can use different analysis populations or denominators depending on how the endpoint is defined and how participants are classified. The denominator is therefore part of the statistical result and should always accompany a reported proportion or affected-participant count.

13. Follow-Up and Time Frame

2018-06-29

Study start

The registry lists 2018-06-29 as the study start date.

Day 1 to 45.2 months

Primary endpoint assessment window

Both registered primary endpoints use a time frame of Day 1 up to 45.2 months.

2022-08-23

Primary completion

The registry lists 2022-08-23 as the primary completion date.

Completed

Study status

The registry identifies the study as completed.

The endpoint time frame is important because a response percentage is meaningful only in relation to the period over which response was assessed. Likewise, adverse-event summaries depend on the observation period during which participants were monitored.

14. Why This Trial Matters Statistically

CARTITUDE-1 is a useful statistical teaching case because it illustrates how the appropriate analysis depends fundamentally on the design and endpoint. The study is not a two-arm randomized comparison; it is a single-group phase 1/2 study with a response proportion as a phase 2 primary endpoint and a severity-based adverse-event endpoint in phase 1b.

ConceptHow it appears in CARTITUDE-1
Single-group designOne study arm is reported.
Phase 1/2 developmentThe study contains phase 1b and phase 2 primary endpoints.
Binary efficacy endpointORR classifies participants according to whether PR or better was achieved.
Response criteriaPR is defined using IMWG criteria and specified laboratory and disease measurements.
Safety severity gradingAdverse events are graded from Grade 1 through Grade 5 under NCI-CTCAE version 5.0.
Denominator disciplineSafety results are reported with population-specific denominators.
Confidence intervalsAppropriate for describing uncertainty around an estimated single-group proportion.
No randomized comparatorA between-arm treatment effect cannot be estimated from this design.
Time frameBoth primary endpoints use Day 1 up to 45.2 months.

15. Important Limitations and Interpretation Issues

16. What a Statistical Comparison Would Require

Understanding what is absent from a design can be as important as understanding what is present. A randomized treatment comparison requires at least two concurrently assigned groups and a prespecified method for comparing their outcomes. CARTITUDE-1 instead provides a single treated group.

What is estimable

The study can estimate the observed proportion achieving a defined response and summarize adverse events within the study population.

What is not a direct design feature

A randomized treatment-versus-control effect is not directly estimable because there is no comparator arm.

Why historical controls differ

A historical comparison does not reproduce the protection against confounding provided by concurrent randomization.

Why endpoint definition matters

Even a numerical comparison across studies can be misleading if response criteria, populations, assessment windows, or analysis populations differ.

17. Statistical Concepts in This Trial

The statistical structure of CARTITUDE-1 can be understood through a few core concepts.

18. Sources

Continue through the Clinical Biostats knowledge graph

Explore the broader collection of statistical tutorials, calculators, and clinical trial analyses.

19. Record Summary

CARTITUDE-1 is a phase 1/2, single-group study of JNJ-68284528 in participants with relapsed or refractory multiple myeloma. The registry reports 126 enrolled participants, one study arm, no masking, and a treatment primary purpose. Its registered primary endpoints address two distinct statistical domains: phase 1b adverse events classified by NCI-CTCAE version 5.0 severity and phase 2 overall response rate defined as the percentage of participants achieving partial response or better according to IMWG criteria.

The key statistical lesson is that study design determines what can be learned from an endpoint. A single-group ORR is an estimate of observed response in the treated study population, while the absence of a concurrent comparator means that the design does not directly estimate a randomized treatment effect. Similarly, safety results must be interpreted with their population-specific denominators and severity definitions. Careful attention to endpoint definitions, analysis populations, denominators, time frames, and statistical uncertainty is therefore essential when interpreting the study.

Clinical Biostats methodology: The most informative trial analysis connects the clinical endpoint to its statistical estimand, identifies the denominator and observation period, distinguishes descriptive estimation from comparative inference, and explains what the design permits the reader to conclude.