Efficacy Comparison of Dostarlimab Plus Chemotherapy Versus Pembrolizumab Plus Chemotherapy in Participants With Metastatic Non-squamous Non-small Cell Lung Cancer (NSCLC): Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Efficacy Comparison of Dostarlimab Plus Chemotherapy Versus Pembrolizumab Plus Chemotherapy in Participants With Metastatic Non-squamous Non-small Cell Lung Cancer (NSCLC) (NCT04581824), a randomized phase 2 study. The primary outcome is Objective Response Rate (ORR).

2. Clinical Question

In statistical terms, the main question is whether the observed response proportions differ between the randomized treatment groups, and how precisely that difference is estimated. The registry record reports response rates and confidence intervals for each group, as well as a recalculated unadjusted comparison. Those results describe the enrolled study population; they do not by themselves establish that the treatments are equivalent or identify a causal mechanism for any observed difference.

3. Trial Design

This was a randomized, double-blind, parallel-group phase 2 study comparing dostarlimab plus chemotherapy with pembrolizumab plus chemotherapy in metastatic non-squamous NSCLC. The registry record lists an enrollment of 243. The treatment groups followed the same described chemotherapy sequence, with the immunotherapy drug differing between groups. Both study drugs and chemotherapy were administered on Day 1 of every 21 Day cycle; cisplatin or carboplatin was used for Cycles 1 to 4 according to investigator decision.

The study began on 2020-11-19, and the primary completion date was 2022-08-04. Its registry status is terminated. The reason for termination and the date of termination are not reported in the registry record. The brief summary describes an intended study period of 5 years and approximately 240 participants; these statements describe the summary’s plan, not the observed duration or a sample-size justification.

Randomization is intended to make treatment groups comparable on average, while double masking can reduce the influence of treatment knowledge on participant care and outcome assessment. These are general design principles; the registry record does not report the randomization procedure, allocation concealment details, or masking procedures. It identifies blinded independent central review for the primary response assessment, which provides a defined assessment process for that endpoint.

The record gives no further details about stratification, protocol changes, interim analyses, or the original design assumptions. It also does not explain why the study was terminated. The enrollment and endpoint results should therefore be interpreted in the context of the design information actually documented, rather than assumptions about the intended analysis or reasons for stopping.

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.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Intent-to-treat (ITT) population included all participants who were randomized.

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.

6. Primary Endpoint

The registry lists the primary outcome as Objective Response Rate (ORR), measured over Up to approximately 20 months.

The primary endpoint was Objective Response Rate (ORR), expressed as the percentage of participants with a confirmed complete response or confirmed partial response as their best overall response. The assessment covered the period from randomization until disease progression or initiation of new anti-cancer therapy, whichever occurred earlier, and was based on blinded independent central review using RECIST version 1.1.

A complete response required disappearance of all target lesions; any pathological lymph nodes had to be less than 10 millimeter in the short axis. A partial response required at least a 30% decrease in the sum of target-lesion diameters, using the baseline sum as reference. Thus, ORR combines two categories of tumor response into a participant-level proportion. It does not describe how long a response lasted or whether disease remained controlled after the response assessment.

As a general statistical concept, a binary endpoint such as ORR makes the outcome easy to interpret as the proportion meeting a defined criterion. It is also sensitive to the response definition, assessment timing, and rules for confirmation and missing assessments. Central review can standardize assessment, but the registry record does not report the detailed handling of missing or unevaluable response assessments. ORR is one measure of tumor response and should be considered alongside other outcomes, including the listed survival endpoints, rather than treated as a complete account of clinical benefit.

7. Results Posted in the Registry

• Pembrolizumab + Chemotherapy: 36.9% (45/122) versus 46.3% (56/121) on control; risk difference -9.4 percentage points (95% CI -21.7 to 2.9), risk ratio 0.80 (0.59–1.08); unadjusted two-proportion z-test p = 0.137.

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.

Number of Participants With Serious AEs

• Endpoint type: Binary: the number of participants with the event, compared as proportions.
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • Pembrolizumab + Chemotherapy: 50.0% (61/122) versus 42.1% (51/121) on control; risk difference 7.9 percentage points (95% CI -4.6 to 20.3), risk ratio 1.19 (0.90–1.56); unadjusted two-proportion z-test p = 0.220.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Pembrolizumab + Chemotherapy and the control. The interval gives the range of effects that remain compatible with the data.

The registry lists secondary outcomes spanning survival, disease progression, treatment-emergent and immune-related adverse events, serious adverse events, and laboratory findings. The list includes Overall Survival and Progression-free Survival, along with counts of participants experiencing several safety outcomes. This broadens the outcome picture beyond tumor response: response is an assessment of tumor shrinkage, whereas survival, progression, adverse events, and laboratory results address different aspects of follow-up and treatment experience.

The secondary outcome record includes a comparison for participants with serious adverse events. Numerical results for other listed secondary endpoints are not reported in the registry record available for this analysis. Their presence in the outcome list does not by itself establish the direction or size of any treatment-group difference. The primary ORR result therefore remains a distinct finding rather than a substitute for the unreported secondary results.

As a general statistical principle, testing or estimating many outcomes creates more opportunities for apparently notable results to arise by chance. Unless a multiplicity strategy is prespecified and applied, individual secondary comparisons are usually less conclusive than a clearly designated primary analysis. The registry record does not report a multiplicity adjustment or a hierarchy for interpreting secondary outcomes. Secondary findings should consequently be treated as supportive or exploratory rather than as independent confirmation of the primary result.

Survival outcomes also have different statistical features from binary response proportions. In general, time-to-event analyses account for when an event occurs and may need to address censoring and follow-up. The registry record lists OS and PFS, but those analysis details and numerical findings are not reported. No inference about survival or progression should be drawn from the ORR estimates alone.

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

• Overall Survival (OS)
• Progression-free Survival (PFS)
• Number of Participants With Treatment-emergent Adverse Event (TEAEs), TEAEs Leading to Death and TEAEs Leading to Treatment Discontinuation
• Number of Participants With Treatment Emergent Immune-related Adverse Event (irAE)
• Number of Participants With Serious AEs
• Number of Participants With Worst Case Clinical Chemistry Results Relative to Normal Range Post-Baseline Relative to Baseline
• Number of Participants With Worst Case Hematology Results Relative to Normal Range Post-Baseline Relative to Baseline
• Number of Participants With Worst Case Urinalysis Results Post-Baseline Relative to Baseline

9. Statistical Interpretation

10. Statistical Methods Explained

For a binary endpoint, each participant either meets the endpoint definition or does not. ORR is reported as a percentage of participants meeting the confirmed complete-response or partial-response definition. A group response percentage estimates the underlying response proportion for a population like the one enrolled, subject to sampling variability and the endpoint’s assessment rules.

What does a confidence interval express?

As a general explanation, a confidence interval is a range of values compatible with the data under a particular statistical procedure. A 95% confidence interval describes the performance of the procedure over repeated samples; it is not a statement that there is a 95% probability that this particular fixed interval contains the true value. Wider intervals indicate less precision, while narrower intervals indicate greater precision under the same method and assumptions.

The registry gives a 95% confidence interval for each arm’s ORR. Those intervals describe uncertainty in each group’s response proportion. A separate interval is needed for the treatment contrast because uncertainty in the difference depends on both groups. Comparing arm-specific intervals by eye is not a substitute for examining a confidence interval or test for the between-group contrast.

What is the recalculated comparison?

The recalculated comparison is an unadjusted risk difference based on posted summary statistics. In general, a risk difference expresses the difference between two event proportions. An unadjusted estimate does not account for covariates or other adjustments. The recalculated comparison includes an unadjusted p-value and a confidence interval; it should not be mistaken for the sponsor’s own reported analysis. The record marks the discrepancy field as false, but it does not identify the detailed method used for the sponsor-reported result.

A p-value is calculated under a null hypothesis and a statistical model. It measures how unusual data at least as incompatible with that null hypothesis would be if the model and null hypothesis were true. It is not the probability that the null hypothesis is true, nor is it a measure of the size or clinical importance of an effect. A p-value above a conventional threshold does not prove the absence of a difference. The p-value reported for the recalculated ORR comparison is unadjusted; the registry record does not state that it accounts for multiple outcomes or analyses.

Why do endpoint definitions matter?

ORR depends on the response criteria and the way responses are assessed. Here, the definition requires confirmed CR or PR as best overall response, with central review against RECIST version 1.1. This specificity helps clarify which participants count as responders. It does not resolve all analytic questions: the registry record does not report how missing assessments, participants without a post-baseline assessment, or other special cases were handled. General statistical practice requires such rules to be specified because different rules can alter the analyzed proportion.

Blinded independent central review is relevant because response assessment can involve interpretation of imaging findings. In general, a central assessment process may reduce the influence of local treatment knowledge on endpoint classification. The registry record identifies the review basis but does not describe reviewer procedures or how disagreements were resolved.

How should the sponsor analysis be read?

The sponsor-reported result is a difference in response rate of 9.32 with an 80% confidence interval from 1.46 to 17.18. Its analysis method is not reported. The recalculated comparison is unadjusted and uses a 95% confidence interval. Different methods or confidence levels can yield results that are not directly comparable, so the two results should be kept distinct. The registry record does not report a sponsor p-value for this analysis.

More generally, the confidence level sets the long-run coverage target of the interval procedure. An 80% interval and a 95% interval need not have the same width, even when based on the same data and method. Interpreting one as if it were the other would misstate the uncertainty. The sponsor-reported interval also does not establish how the analysis was prespecified, because that information is not reported in the registry record.

Why does over-dispersion matter in general?

Over-dispersion is a general issue in statistical models for count or proportion data. It occurs when observed variability exceeds what a model assumes. If ignored, standard errors may be too small and confidence intervals or tests may be too optimistic. The posted primary outcome is a binary participant-level response proportion, and the registry record does not identify an over-dispersion adjustment. There is no basis to infer that such an adjustment was required or performed for this analysis; the concept is relevant as a general modeling consideration, not as a reported feature of this trial.

Overall, the primary figures are best read as descriptive arm-level estimates plus an uncertain unadjusted contrast. The endpoint definition and analysis population are stated, but key details about missing data and the sponsor’s analysis method are not reported in the registry record. Those omissions limit how fully the calculation can be reproduced or checked from the record alone.

11. Confidence Intervals and Precision

The arm-specific ORR estimates are 46.0% (95% confidence interval, 37.2% to 55.6%) for dostarlimab plus chemotherapy and 37.0% (95% confidence interval, 28.3% to 46.1%) for pembrolizumab plus chemotherapy. The intervals extend across a range of plausible response proportions, so neither group’s rate is known with high precision. The apparent difference between the point estimates should be considered together with uncertainty in the comparison, rather than read as a precise estimate of a treatment effect.

The recalculated comparison’s confidence interval spans zero. This means the interval includes the value corresponding to no difference under that risk-difference scale, as well as values on either side. In general, this breadth signals uncertainty about both the direction and magnitude of the contrast. The associated unadjusted p-value is 0.137; it does not quantify the probability that there is no difference and does not prove equal effects.

The sponsor-reported interval is an 80% interval, whereas the recalculated comparison uses a 95% interval. Confidence intervals with different coverage levels should not be compared as if they represented identical uncertainty procedures. The sponsor result is reported as 9.32, with limits of 1.46 and 17.18; its method and sign convention are not reported in the registry record. The different confidence level and absent method details limit direct interpretation alongside the recalculation.

As a general matter, precision depends on the amount of information, variability, endpoint definition, and analysis method. A confidence interval does not capture all sources of uncertainty, including possible bias from missing assessments or deviations from the planned analysis. Details about these issues are not reported in the registry record. The intervals therefore describe statistical uncertainty under their respective procedures, not every uncertainty affecting interpretation.

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 figures are planning illustrations based on the observed effect, not the trial’s original design assumptions. Because the primary endpoint is a participant-level binary outcome, the listed figures are participants per arm, not total events. The calculation assumes a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. For each planning scenario, the effect to detect is the stated fraction of the observed effect.

The 80% power column gives the planned number of participants per arm for a calculation with an 80% chance of detecting the specified effect under the listed assumptions, if that effect and the model assumptions are correct. The 90% power column uses a higher target chance of detection for that same specified effect, and therefore requires more participants per arm. These are planning quantities, not a description of the actual study’s operating characteristics.

For the scenario using 100% of the observed effect, the illustration lists 431 participants per arm for 80% power and 577 participants per arm for 90% power. For 75% of the observed effect, it lists 773 and 1,034 participants per arm, respectively; for 50% of the observed effect, it lists 1,751 and 2,344 participants per arm. Smaller effects require more participants because they are harder to distinguish from random variation.

These values are not observed power. Observed power calculated from an estimated effect does not provide an additional measure of evidence beyond the estimate and its uncertainty. The planning illustration is sensitive to the observed effect used as its starting point, and the record does not report that these figures were used to design the terminated study.

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 = 0.463; treatment proportion p₁ = 0.369; average p̄ = 0.416.
2. First term: z_1−α/2 × √[2p̄(1−p̄)] = 1.96 × 0.6970 = 1.3661.
3. Second term: z_1−β × √[p₀(1−p₀) + p₁(1−p₁)] = 0.842 × 0.6938 = 0.5840.
4. n = (1.3661 + 0.5840)² / (-0.0940)² = 430.8, rounded up to 431 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 <- 0.462809917355372                          # control proportion
p1 <- 0.368852459016393                          # 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(4581824)
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 = 431 (80% power), 577 (90% power)
 75% of observed effect: n per arm = 773 (80% power), 1,034 (90% power)
 50% of observed effect: n per arm = 1,751 (80% power), 2,344 (90% power)
power.prop.test: 430.8 
Simulated power at n = 431 per arm: 0.804 (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. Analyses Reported by the Sponsor

The registry reports a sponsor analysis of the difference in response rate as 9.32, with an 80% confidence interval from 1.46 to 17.18. The record lists the groups in the order dostarlimab plus chemotherapy and pembrolizumab plus chemotherapy. It does not report the analysis method or sign convention, and no p-value is listed for the sponsor analysis. The result should therefore be described as recorded without inferring unreported methodological details.

The separate recalculation is a simple unadjusted comparison derived from the posted summary statistics. It uses a 95% confidence interval and reports an unadjusted p-value of 0.137. Because its method and confidence level differ from the sponsor result, the estimates should not be treated as interchangeable or as competing versions of a single fully specified analysis. The recalculated interval spans zero, while the sponsor interval is reported with an 80% confidence level; the record does not provide enough method detail to reconcile them.

The discrepancy indicator is false. The registry record does not report an explanation of the difference in analysis presentation. The sponsor result and the recalculation should remain clearly distinguished: one is the registry’s reported analysis, and the other is an unadjusted re-calculation from posted summary statistics. Neither should be used to imply an analysis plan or adjustment that the record does not describe.

15. Multiplicity, Missing Data and Analysis Populations

The registry identifies ORR as the primary outcome and lists multiple secondary outcomes, including survival, progression, adverse-event, and laboratory measures. When many outcomes or comparisons are examined, the chance of at least one apparently unusual result can increase. A multiplicity plan can control or organize that risk, for example by defining which hypotheses are confirmatory and how testing proceeds. The multiplicity strategy is not reported in the registry record, so it is not possible to determine whether or how secondary analyses were adjusted.

The record defines the primary population as the intent-to-treat population, including all randomized participants. This is an important analysis-population description, but it does not specify how every randomized participant contributed to the ORR calculation. Rules for missing or unevaluable response assessments, handling of new anti-cancer therapy, and other deviations are not reported in the registry record. Nor are the reasons for any missing assessments or the extent of missingness.

As a general matter, missing outcome data can affect both the estimate and its uncertainty. The direction of potential bias depends on why data are missing and on the analysis assumptions; it cannot be resolved by knowing the population label alone. Sensitivity analyses can explore alternative assumptions, but the registry record does not report whether such analyses were conducted.

The record does not report the full statistical analysis plan, the original sample-size justification, interim analyses, or adjustment methods for the sponsor analysis. These gaps matter because they limit assessment of whether the reported analyses followed prespecified decisions and how much emphasis should be placed on individual results. The posted findings can be described, but these unreported details should not be inferred from the outcome labels or from the numerical results.

16. 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: pneumonia (12.3%), anaemia (7.4%), dyspnoea (4.1%), pneumonitis (4.1%), febrile neutropenia (3.3%).

For serious adverse events, the registry reports 51 participants in the dostarlimab plus chemotherapy group and 61 participants in the pembrolizumab plus chemotherapy group. The corresponding group sizes listed for this outcome are 121 and 122. The recalculated comparison is an unadjusted risk difference with a confidence interval that spans zero; its unadjusted p-value is 0.22. This indicates uncertainty in the comparison, not proof that serious adverse-event experience was the same in both groups.

The record also lists participants with other adverse events and deaths, but those counts do not establish that every event or death was caused by study treatment. The registry’s most frequently listed serious-event terms include pneumonia, anaemia, dyspnoea, pneumonitis, febrile neutropenia, and COVID-19. These event listings provide context but are not a complete safety assessment.

As a general principle, safety counts should be read with their definitions, observation periods, exposure, severity, and attribution rules in mind. The record’s serious-event timeframe is up to approximately 46 months. Detailed definitions and exposure-adjusted analyses are not reported in the registry record, so the counts should not be interpreted as a causal comparison or as a complete description of risk.

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

The posted primary result describes the percentage of randomized participants meeting the registry’s confirmed-response definition in each treatment group. The point estimates were 46.0% for dostarlimab plus chemotherapy and 37.0% for pembrolizumab plus chemotherapy. The recalculated unadjusted comparison is uncertain: its confidence interval spans zero, and its unadjusted p-value is 0.137. These results do not establish that the treatments are equivalent, nor do they provide strong statistical evidence of a difference under that recalculation.

A confidence interval gives a range of values compatible with the data under the method used; it is not a guarantee about the result in another population. The observed rates may differ from rates in future settings because of sampling variability, participant characteristics, outcome assessment, or other factors. The registry record does not provide all details needed to assess missing outcome data or the sponsor analysis method.

ORR measures whether a participant had a confirmed complete or partial response by the specified criteria. It does not by itself measure duration of response, overall survival, progression-free survival, quality of life, or the overall balance of benefits and harms. Although the registry lists some of these as secondary outcomes, their numerical findings are not reported in the record available for this analysis. The response estimates should not be used as a stand-in for those outcomes.

The study is listed as terminated, but the reason for termination is not reported in the registry record. Termination alone does not explain the observed estimates or establish that a treatment was effective, ineffective, safe, or unsafe. The most defensible reading is limited to the recorded endpoint estimates, their uncertainty, the specified analysis population, and the methodological information documented in the record.

This is a statistical interpretation of posted trial results, not medical advice. Treatment decisions require consideration of the full clinical evidence and an individual’s circumstances by qualified clinicians.

18. Why This Trial Matters Statistically

This trial is statistically informative as a randomized comparison of two immunotherapy-plus-chemotherapy regimens using a clearly defined tumor-response endpoint. The double-blind, parallel-group design and central review basis for ORR are relevant features when interpreting response assessment. The record also identifies an intent-to-treat population, making the primary analysis population explicit at a broad level.

The results illustrate why an observed difference in point estimates is not enough to establish a treatment contrast. Each group’s response percentage has a confidence interval, and the recalculated comparison has an interval spanning zero. Statistical interpretation therefore needs both an estimate and its precision, rather than relying only on which point estimate is larger or on a single p-value.

The record also demonstrates the importance of keeping sponsor-reported and recalculated analyses distinct. The sponsor’s difference is reported with an 80% confidence interval, while the re-calculation is unadjusted and uses a 95% interval. With the sponsor method and sign convention unreported, readers cannot fully reconcile the two analyses from the registry record alone.

Finally, ORR sits within a wider set of outcomes that includes survival, progression, safety, and laboratory findings. Those outcomes address different questions, and their numerical results are not all reported in the available record. The terminated status and unreported analysis details reinforce the value of distinguishing what the registry states from what cannot be established from the posted results.

19. 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 registry status is terminated, and the reason for termination is not reported in the registry record. This limits interpretation of how the study’s conduct and final follow-up relate to the original plan. The record lists an enrollment of 243 and describes approximately 240 participants as intended in the brief summary, but does not report the original statistical justification for the planned enrollment.

The sponsor-reported response-rate difference lacks a reported analysis method and sign convention. The recalculation is unadjusted, and the registry record does not give a full analysis plan that would allow readers to verify model choices, adjustments, or prespecification. The two reported analyses also use different confidence levels, which limits direct comparison.

Although the primary population is defined as all randomized participants, detailed rules for missing, unevaluable, or unconfirmed response assessments are not reported in the registry record. The record also does not report the amount or pattern of missing endpoint data, sensitivity analyses, or an approach to multiplicity. These omissions constrain assessment of robustness.

The primary endpoint is a response measure with a timeframe of up to approximately 20 months. The registry lists survival and progression outcomes, but their numerical results and detailed analyses are not reported in the record available for this analysis. The serious adverse-event timeframe is up to approximately 46 months, while details about exposure, attribution, and event ascertainment are not reported. These limitations restrict conclusions to the outcomes and methods documented, rather than supporting a complete assessment of comparative effectiveness or safety.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: Efficacy Comparison of Dostarlimab Plus Chemotherapy Versus Pembrolizumab Plus Chemotherapy in Participants With Metastatic Non-squamous Non-small Cell Lung Cancer (NSCLC) — NCT04581824.
• Registry record last updated 2025-08-08; status last verified 2025-07.
• Results first posted 2023-09-18.
• 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.

23. Record Summary

NCT04581824 is a randomized, double-blind, parallel-group phase 2 study of dostarlimab plus chemotherapy versus pembrolizumab plus chemotherapy in metastatic non-squamous NSCLC. The primary endpoint was ORR, defined by confirmed complete or partial response under RECIST version 1.1 and assessed by blinded independent central review. The intent-to-treat population included all randomized participants.

The recorded ORR was 46.0% for dostarlimab plus chemotherapy and 37.0% for pembrolizumab plus chemotherapy. The recalculated unadjusted comparison has a confidence interval spanning zero and an unadjusted p-value of 0.137. The sponsor-reported difference in response rate was 9.32, with an 80% confidence interval from 1.46 to 17.18; its method is not reported in the registry record.

The study is listed as terminated, and the reason is not reported. Secondary outcomes include survival, progression, safety, and laboratory measures; numerical results for most are not reported in the record available for this analysis.