A Study to Investigate the Efficacy and Safety of Two Doses of GSK2857916 in Participants With Multiple Myeloma Who Have Failed Prior Treatment With an Anti-CD38 Antibody: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for A Study to Investigate the Efficacy and Safety of Two Doses of GSK2857916 in Participants With Multiple Myeloma Who Have Failed Prior Treatment With an Anti-CD38 Antibody (NCT03525678), a randomized phase 2 study. The primary outcome is Overall Response Rate (ORR) by Independent Review Committee (IRC) (Full Analysis Population).

2. Clinical Question

The clinical question concerns response proportions across treatment configurations and doses, not a direct comparison with no treatment or another medicine. The registry describes a randomized, open-label, parallel study of the frozen-liquid doses. Participants were followed for progression-free survival and overall survival as well as response-related outcomes, but the primary endpoint reported for this analysis is overall response rate.

3. Trial Design

This was a completed Phase II, open-label, randomized, parallel-group study sponsored by GlaxoSmithKline. The registry lists enrollment of 221 participants. The randomized comparison involved frozen-liquid GSK2857916 at 2.5 mg/kg and 3.4 mg/kg, administered intravenously every 3 weeks. Because the study was open label, participants and study personnel were not masked to treatment assignment according to the registry record.

The registry also describes an independent cohort receiving lyophilized GSK2857916 at 3.4 mg/kg. That cohort was to receive the lyophilized configuration once it became available and enrollment had been completed for the frozen-liquid arms. It should therefore not be read as another randomized dose arm. Differences between that cohort and the randomized arms may reflect cohort formation and other differences, not just formulation.

The brief summary describes treatment continuing until disease progression or unacceptable toxicity, with follow-up for progression-free survival and overall survival. For participants who discontinued for reasons other than progression, disease evaluations were to continue at 3-week intervals until a specified follow-up endpoint, such as confirmed progression or the end of the study. The primary endpoint timeframe is up to 48 weeks. The registry record does not report the original sample-size justification, the detailed allocation procedure, or further protocol details about the conduct of randomization.

For the primary full analysis population, all randomized participants were included whether or not randomized treatment was administered, and participants were analyzed according to their randomized treatment. The primary result therefore uses the randomized population definition stated in the record. The lyophilized cohort is presented separately and should not be interpreted as part of that randomized dose comparison.

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: A total of 221 participants were enrolled in this study. The study consisted of two phases - Main Study Phase and Post Analysis Continued Treatment (PACT) Phase. In PACT phase those participants benefiting from drug continued to receive study drug until discontinuation or withdrawal from study.

Main Study (Day 1 to 186 Weeks)

Reasons for not completing this period:

PACT Phase (From 186 Weeks to 325 Weeks)

Analysis population for the primary outcome, as defined in the registry: Full Analysis Population comprised of all randomized participants (any participant who received a treatment randomization number was considered as randomized) whether or not randomized treatment was administered. This population was based on the treatment the participant was randomized to.

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: Baseline characteristics were presented for all randomized participants including 3 randomized participants who never received treatment (2-withdrawal by physician decision and 1-death).

6. Primary Endpoint

The registry lists the primary outcome as Overall Response Rate (ORR) by Independent Review Committee (IRC) (Full Analysis Population), measured over Up to 48 weeks.

The primary endpoint is overall response rate (ORR) by independent review committee (IRC) in the full analysis population, measured as the percentage of participants with a confirmed partial response or better. The response categories counted are partial response, very good partial response, complete response and stringent complete response, using the 2016 international myeloma working group response criteria. The registry specifies a timeframe of up to 48 weeks and exact-method confidence intervals for the arm estimates.

ORR is a binary outcome at the participant level: each participant either meets the response definition or does not. Expressing it as a percentage describes how common a confirmed response was in each arm during the stated timeframe. It can be useful for comparing response proportions across groups, while not describing how long a response lasts. Duration of response is listed separately among the secondary outcomes.

Independent review is relevant because treatment allocation was open label. The IRC assessment offers an assessment distinct from investigator assessment, although the registry record does not provide further details about review procedures. A response endpoint also does not, by itself, establish an effect on progression-free survival, overall survival, quality of life or other outcomes. Those require their own results and interpretation.

7. Results Posted in the Registry

• Main Study: GSK2857916 3.4 mg/kg (Frozen Liquid): 34.3% (34/99) versus 30.9% (30/97) on control; risk difference 3.4 percentage points (95% CI -9.7 to 16.5), risk ratio 1.11 (0.74–1.66); unadjusted two-proportion z-test p = 0.610.
• Main Study: GSK2857916 3.4 mg/kg (Lyophilized): 48.0% (12/25) versus 30.9% (30/97) on control; risk difference 17.1 percentage points (95% CI -4.6 to 38.7), risk ratio 1.55 (0.94–2.57); unadjusted two-proportion z-test p = 0.109.

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.

Overall Response Rate by Investigator Assessment (IA) (Full Analysis Population)

• 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:
  • Main Study: GSK2857916 3.4 mg/kg (Frozen Liquid): 32.3% (32/99) versus 33.0% (32/97) on control; risk difference -0.7 percentage points (95% CI -13.8 to 12.5), risk ratio 0.98 (0.66–1.46); unadjusted two-proportion z-test p = 0.921.
  • Main Study: GSK2857916 3.4 mg/kg (Lyophilized): 52.0% (13/25) versus 33.0% (32/97) on control; risk difference 19.0 percentage points (95% CI -2.7 to 40.7), risk ratio 1.58 (0.98–2.53); unadjusted two-proportion z-test p = 0.079.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Main Study: GSK2857916 3.4 mg/kg (Frozen Liquid) and the control. The interval gives the range of effects that remain compatible with the data.
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Main Study: GSK2857916 3.4 mg/kg (Lyophilized) and the control. The interval gives the range of effects that remain compatible with the data.

Overall Response Rate by Investigator Assessment (Efficacy Population)

• 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:
  • Main Study: GSK2857916 3.4 mg/kg (Frozen Liquid): 25.8% (17/66) versus 32.8% (21/64) on control; risk difference -7.1 percentage points (95% CI -22.7 to 8.6), risk ratio 0.78 (0.46–1.35); unadjusted two-proportion z-test p = 0.377.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Main Study: GSK2857916 3.4 mg/kg (Frozen Liquid) and the control. The interval gives the range of effects that remain compatible with the data.

Clinical Benefit Rate (CBR) by Investigator Assessment (Full Analysis Population)

• 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:
  • Main Study: GSK2857916 3.4 mg/kg (Frozen Liquid): 38.4% (38/99) versus 35.1% (34/97) on control; risk difference 3.3 percentage points (95% CI -10.2 to 16.8), risk ratio 1.10 (0.76–1.58); unadjusted two-proportion z-test p = 0.629.
  • Main Study: GSK2857916 3.4 mg/kg (Lyophilized): 60.0% (15/25) versus 35.1% (34/97) on control; risk difference 24.9 percentage points (95% CI 3.5 to 46.4), risk ratio 1.71 (1.13–2.60); unadjusted two-proportion z-test p = 0.023.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Main Study: GSK2857916 3.4 mg/kg (Frozen Liquid) and the control. The interval gives the range of effects that remain compatible with the data.
  • The 95% confidence interval excludes no difference, so the proportion with the event is statistically higher for Main Study: GSK2857916 3.4 mg/kg (Lyophilized) than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

The posted secondary outcomes broaden the response assessment beyond the primary independent-review ORR. They include investigator-assessed ORR in the full analysis and efficacy populations, clinical benefit rate assessed by investigators or an independent review committee, and duration of response. The timeframes and populations differ across outcomes, so they answer related but not identical questions.

Across the posted ORR and clinical benefit results, the pattern is not identical across dose, formulation and assessment source. The investigator-assessed estimates do not simply reproduce the primary independent-review estimates, and the lyophilized cohort remains distinct from the randomized frozen-liquid groups. This variation is a reason to consider the total pattern and the design of each comparison rather than treating any one secondary result as a definitive confirmation of the primary finding.

Secondary endpoints can help describe consistency, alternative assessment approaches and aspects of response not captured by ORR alone. However, multiple outcomes and comparisons create more opportunities for apparently notable results to arise by chance. The posted recalculated p-values are unadjusted, and the registry record does not report a multiplicity strategy or identify which secondary tests, if any, were intended to be confirmatory. As a result, secondary findings carry less inferential weight than a prespecified, multiplicity-controlled analysis would carry.

The secondary results should therefore be read as supportive descriptive evidence, not as independent proof that a treatment contrast exists. The endpoint definitions, population, assessment method, follow-up timeframe and cohort allocation all matter to whether a result can be meaningfully compared with the primary endpoint.

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

• Overall Response Rate by Investigator Assessment (IA) (Full Analysis Population)
• Overall Response Rate by Investigator Assessment (Efficacy Population)
• Clinical Benefit Rate (CBR) by Investigator Assessment (Full Analysis Population)
• Clinical Benefit Rate by Investigator Assessment (Efficacy Population)
• Clinical Benefit Rate by Independent Review Committee (Full Analysis Population)
• Clinical Benefit Rate by Independent Review Committee (Efficacy Population)
• Duration of Response (DoR) by Investigator Assessment (Full Analysis Population)
• Duration of Response by Investigator Assessment (Efficacy Population)

9. Statistical Interpretation

10. Statistical Methods Explained

Binary response outcome. ORR is a binary outcome because each participant is classified according to whether a confirmed response meeting the stated criteria occurred. The arm percentage is the proportion meeting that definition. This makes the result straightforward to describe, but the percentage remains an estimate based on the participants assessed in that population and timeframe.

Confidence intervals. A confidence interval expresses statistical uncertainty around an estimate. In repeated use of the same method under its assumptions, the stated confidence level describes the method’s long-run coverage, not the probability that a particular fixed interval contains the true value. For the primary arm percentages, the registry specifies 97.5% confidence intervals based on the exact method. Exact methods are commonly used for proportions when normal approximations may be less reliable, especially with limited event counts. The record does not report a separate rationale for choosing this interval level.

Independent review and investigator assessment. The primary endpoint is assessed by an independent review committee, while several secondary outcomes use investigator assessment. These are different assessment sources and can yield different response classifications. Open-label conduct may be relevant to investigator assessment because treatment is known, while independent review provides a separate assessment. The record does not report enough procedural detail to characterize how disagreements were resolved or how the review was organized.

Full analysis population. For the primary endpoint, the full analysis population includes all randomized participants, whether or not randomized treatment was administered, analyzed according to randomized treatment. This population definition preserves the randomized assignment for the frozen-liquid comparison. It does not mean that every participant necessarily had complete endpoint information; the handling of missing response assessments is not reported in the registry record.

Risk difference. A risk difference describes the difference between two event proportions. A value near zero indicates that the estimated proportions are close; positive or negative values indicate the direction under the ordering used for the comparison. The recalculated comparisons use the 2.5 mg/kg frozen-liquid group as the reference. Their intervals describe uncertainty in the estimated difference, and an interval crossing zero indicates that the data do not rule out no difference under that calculation. The recalculations are unadjusted and should not be mistaken for the sponsor’s planned analysis.

p-values. A p-value is a measure of how incompatible the observed data are with a specified null model, assuming that model and the analysis assumptions hold. It is not the probability that the null hypothesis is true, nor does it measure the size or clinical importance of a difference. The p-values accompanying recalculated comparisons are unadjusted; no multiplicity adjustment or confirmatory interpretation is reported for them.

Planning illustrations. The sample-size illustrations use a two-sided significance level of 0.05, equal allocation between two arms and a normal approximation. The illustrated effect is a stated fraction of the observed effect. Such calculations show how sample requirements change under those planning choices; they do not describe the original trial’s design assumptions, which are not reported in the registry record. They also do not provide observed power, which is not a useful substitute for interpreting the estimates and their uncertainty.

Dispersion and outcome type. For a binary percentage, the uncertainty is about the estimated event proportion and is represented by an interval rather than a standard deviation of a continuous measurement. General statistical methods must match the outcome type and the amount of information in the data. For other outcomes, such as counts with over-dispersion, methods that allow variability beyond a simple reference model may be important; that general consideration does not establish that such an outcome or method was used for this primary endpoint.

Limits of the reported analysis. The registry record lists numerous outcome measures and secondary response analyses, but detailed statistical models, covariate adjustments, missing-data procedures, interim analyses and the original sample-size justification are not reported in the registry record. The posted summary statistics support descriptive reading and the stated unadjusted recalculations, but cannot answer how unreported protocol analyses were specified or conducted.

11. Confidence Intervals and Precision

The primary arm estimates are 31%, 34% and 48%, with 97.5% confidence intervals of 20.8 to 42.6, 23.9 to 46.0 and 25.5 to 71.1. These intervals show that the point estimates should not be read as precise response probabilities. The lyophilized estimate has the widest interval and comes from 25 participants, compared with 97 and 99 in the frozen-liquid arms. The interval width indicates substantial uncertainty around the lyophilized proportion.

The randomized frozen-liquid arms have intervals that overlap considerably. For their unadjusted recalculated risk difference, the interval extends from -0.097 to 0.165. It includes zero and spans both directions, so the estimate does not pin down a clear difference between those arms. The lyophilized comparison interval, -0.046 to 0.387, also includes zero and extends over a broad range. Its interpretation is further limited by the nonrandomized cohort structure.

Intervals are not guarantees about the range of future study results, and a broad interval does not prove that there is no difference. It indicates that the estimate is relatively uncertain under the method used. Conversely, a higher point estimate alone does not establish a reliable difference, particularly when the cohort is small and not randomized. Precision should be considered alongside design and endpoint assessment, rather than inferred from the point estimate alone.

The primary arm intervals use the exact method stated in the registry. The recalculated comparison intervals are reported as intervals, but their confidence level is not specified in the registry record. They should therefore be read as the uncertainty ranges accompanying those unadjusted recalculations, without assigning them an unreported confidence level.

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 expressed as participants per arm for 80% power and participants per arm for 90% power. They apply to binary outcomes in these comparisons, so the values are participants in each of two equally allocated arms, not total events. The assumptions are a two-sided significance level of 0.05, equal allocation between the two arms and a normal approximation. The assumed effect to detect is the stated fraction of the observed effect.

The 80% power column gives the number of participants per arm that the planning calculation indicates would be needed to have an 80% chance of detecting the specified effect under those assumptions, if the assumptions hold. The 90% power column uses a higher probability of detecting that effect and therefore generally requires more participants. Smaller effects need more participants to detect because they are harder to distinguish from random variation.

The rows for 100%, 75% and 50% of the observed effect are alternative planning scenarios, not three observed results. The observed effect is used only as a reference for those illustrations. The registry record does not report that these assumptions were used to design the original trial, and the lyophilized cohort was not randomized alongside the frozen-liquid arms. These figures are planning illustrations, not observed power.

Very large or small outputs can arise when a planning calculation uses an observed effect that is itself uncertain, particularly when the comparison is based on limited data or an estimate close to zero. The figures should be interpreted as consequences of the stated assumptions, not as evidence that a future study of that size would necessarily produce a particular result.

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 = 0.309; treatment proportion p₁ = 0.343; average p̄ = 0.326.
2. First term: z_1−α/2 × √[2p̄(1−p̄)] = 1.96 × 0.6631 = 1.2996.
3. Second term: z_1−β × √[p₀(1−p₀) + p₁(1−p₁)] = 0.842 × 0.6627 = 0.5577.
4. n = (1.2996 + 0.5577)² / (0.0342)² = 2,957.0, rounded up to 2,958 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 <- 0.309278350515464                          # control proportion
p1 <- 0.343434343434343                          # 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(3525678)
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 = 2,958 (80% power), 3,959 (90% power)
 75% of observed effect: n per arm = 5,222 (80% power), 6,991 (90% power)
 50% of observed effect: n per arm = 11,669 (80% power), 15,620 (90% power)
power.prop.test: 2957 
Simulated power at n = 2,958 per arm: 0.806 (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 registry lists 62 outcome measures and multiple response analyses, including overall response rate, clinical benefit rate and duration of response, with some outcomes assessed by investigators and others by an independent review committee. Several analyses also use different populations and timeframes. Examining many outcomes and comparisons increases the chance that some estimates or unadjusted p-values may appear notable even when there is no underlying effect.

The registry record does not report whether a multiplicity adjustment was planned, how the primary and secondary hypotheses were ordered, or whether any secondary endpoints were designated as confirmatory. It also does not report a detailed statistical analysis plan. The recalculated comparisons are unadjusted and should not be interpreted as if they had controlled the overall chance of false-positive findings across the outcome set.

The primary endpoint is defined for the full analysis population, which includes all randomized participants regardless of whether treatment was administered, analyzed by randomized assignment. The efficacy population appears in secondary results, but its membership definition is not reported in the registry record. The record also does not explain how participants without an evaluable response assessment were handled, whether missing observations were imputed, or what sensitivity analyses were planned.

These omissions limit assessment of robustness. Missing-data handling can matter for a percentage endpoint if participants with unknown response status are treated differently, and population definitions affect which participants contribute to an estimate. The registry’s primary population definition is clear, but it does not resolve these other methodological questions. Interim analyses and their possible relationship to error control are not reported in the registry record.

15. 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: electrocardiogram t wave inversion (100.0%), mitral valve disease (100.0%), anaemia (50.0%), gastrointestinal haemorrhage (50.0%), basal cell carcinoma (50.0%).

The registry reports serious adverse event counts among participants at risk in the main-study groups: 43 of 95 in the 2.5 mg/kg frozen-liquid group, 53 of 99 in the 3.4 mg/kg frozen-liquid group and 15 of 24 in the 3.4 mg/kg lyophilized group. It also lists counts for the PACT phase: 1 of 1 in the 2.5 mg/kg frozen-liquid group and 2 of 2 in the 3.4 mg/kg frozen-liquid group. The PACT groups are very small, so those counts should not be read as stable estimates of frequency.

The record separately reports other adverse events and deaths as counts. These are descriptive safety data and do not, by themselves, establish that treatment caused an event or death. The registry record does not report the detailed attribution procedures or enough context to interpret the deaths as treatment-related. The listed serious-event terms include electrocardiogram T wave inversion, mitral valve disease, anaemia, gastrointestinal haemorrhage, basal cell carcinoma and viral gastroenteritis; the percentages attached to terms do not provide event counts in the posted summary.

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

The primary result describes the proportion of participants who met a defined response threshold by independent review, within the stated timeframe and population. It offers information about observed response frequency in this study. It does not establish the duration of response, the timing of response, or whether responses translated into longer progression-free or overall survival.

The randomized comparison is between two frozen-liquid doses. Its point estimates are close, and the unadjusted recalculated risk-difference interval includes zero and allows for differences in either direction. This does not prove that the doses are equivalent, nor does it establish a difference. The study’s reported summary statistics and the uncertainty around the estimates are the appropriate basis for interpreting what the data resolve.

The lyophilized cohort has a higher primary point estimate than the frozen-liquid 2.5 mg/kg arm, but it was an independent cohort, had a smaller population, and has a broad interval. The comparison is not a randomized formulation contrast. Differences between cohorts cannot be attributed solely to formulation using these results.

These findings apply to the trial population described in the registry, including participants with heavily pretreated multiple myeloma and prior anti-CD38 antibody failure. They should not be generalized automatically to other populations, treatment settings or comparators. The registry record does not establish a comparison with another active treatment or a control without GSK2857916.

Nothing in the response percentages alone determines an individual participant’s likely outcome. Nor do these data, on their own, establish overall benefit, harm, or the balance between them. Safety counts and other endpoints require their own assessment, and clinical decisions require evidence and context beyond this statistical summary.

17. Why This Trial Matters Statistically

This trial illustrates how design and assessment determine the meaning of a treatment comparison. Randomization supports a more direct comparison between the frozen-liquid doses because treatment assignment was randomized. The separate lyophilized cohort does not have that same protection against differences between groups, despite sharing a comparator in the posted calculations.

It also shows why point estimates need to be read with intervals. The primary percentages differ numerically across groups, but the interval widths—particularly for the smaller lyophilized cohort—show substantial uncertainty. A larger point estimate is not automatically a more reliable estimate of a treatment contrast.

The endpoint and assessor matter as well. The primary outcome is confirmed ORR by independent review, while secondary response outcomes include investigator assessment and clinical benefit rate. Related measures can provide context, but they are not interchangeable. The open-label design makes it especially important to distinguish independent review from investigator assessment without assuming that either is error-free.

Finally, the registry contains multiple outcomes and comparison opportunities, while its multiplicity strategy and detailed statistical plan are not reported. This is a practical reminder that unadjusted secondary findings should be interpreted cautiously and that statistical significance, effect size, uncertainty and study design answer different questions.

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 open-label design means treatment assignment was not masked. The primary endpoint was assessed by an independent review committee, but several secondary outcomes were investigator assessed; the record does not describe review procedures in detail. Differences across assessment sources may reflect assessment as well as underlying response.

The lyophilized group was an independent cohort enrolled after the frozen-liquid arms had completed enrollment, rather than a randomized arm. Its comparison with a randomized frozen-liquid group is therefore vulnerable to differences between cohorts. Its sample size is also smaller, and its primary interval is broad.

The registry record does not report the original sample-size justification, detailed randomization procedures, a complete statistical analysis plan, interim analyses, multiplicity control or the handling of missing response assessments. The efficacy population used for secondary analyses is not defined in the record. These omissions make it difficult to assess whether all reported analyses followed prespecified procedures or how sensitive the results are to analytic choices.

The primary endpoint is a response proportion over a specified timeframe and does not by itself capture response duration or survival. Although follow-up for progression-free and overall survival is described, those outcomes are not the focus of the primary result interpreted. The safety summary provides counts but limited detail about event attribution and context. These limitations constrain what can be concluded from the registry results.

19. Related Tutorials

20. Related Statistical Calculators

21. Sources

• ClinicalTrials.gov: A Study to Investigate the Efficacy and Safety of Two Doses of GSK2857916 in Participants With Multiple Myeloma Who Have Failed Prior Treatment With an Anti-CD38 Antibody — NCT03525678.
• Registry record last updated 2025-09-03; status last verified 2025-08.
• Results first posted 2020-04-28.
• 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

NCT03525678 is a completed, randomized, open-label, parallel Phase II study sponsored by GlaxoSmithKline. It enrolled 221 participants with multiple myeloma after prior treatment, including failure of an anti-CD38 antibody. The randomized arms received frozen-liquid GSK2857916 at 2.5 mg/kg or 3.4 mg/kg; a separate cohort received the 3.4 mg/kg lyophilized configuration.

The primary endpoint was ORR by independent review committee in the full analysis population, through up to 48 weeks. The registry reports arm estimates with exact 97.5% confidence intervals. The randomized frozen-liquid comparison’s unadjusted recalculation is uncertain, while the lyophilized cohort is not a randomized comparison. The record lists secondary response outcomes and safety counts, but does not report the original sample-size justification, detailed analysis plan, multiplicity strategy or missing-data procedures.