Savolitinib vs. Sunitinib in MET-driven PRCC.: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Savolitinib vs. Sunitinib in MET-driven PRCC. (NCT03091192), a randomized phase 3 study. The primary outcome is Progression Free Survival (PFS) by Blinded Independent Central Review (BICR).

2. Clinical Question

3. Trial Design

This was a Phase III, randomized, controlled, multicentre, parallel-group study comparing savolitinib with sunitinib in patients with MET-driven PRCC that was unresectable and locally advanced or metastatic. The registry record describes the trial as open label and lists masking as none. The study status is active, not recruiting. Enrollment is listed as 60.

The primary outcome was PFS by BICR, with tumour assessments by MRI or CT described as occurring every 6 weeks from randomization until progression as defined by RECIST 1.1 and confirmed by BICR. The analysis population is the Full Analysis Set, which the record defines as all randomized patients. The reported arm summaries include 33.0 patients for savolitinib and 27.0 for sunitinib. The registry record does not report further details about allocation procedures or the original rationale for the group sizes.

Although the treatment comparison was open label, the primary outcome is identified as assessed by blinded independent central review. This describes the outcome review method; it does not mean that treatment assignment was masked to participants or treating teams. Further operational details about the review process are not reported in the registry record.

Secondary outcomes listed in the record include overall survival, objective response rate, duration of response, and disease control rate at 6 months and 12 months. Numerical posted results are available for objective response rate and the two disease-control outcomes. Numerical results for overall survival and duration of response are not reported in the registry record. The original statistical plan, planned sample-size justification, and any interim-analysis procedures are not reported in the registry record.

4. Participant Flow and Accounting

The registry reports how participants moved through each stage of the study. All counts below are exactly as posted; missing counts are not inferred from other stages.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Full Analysis Set (All randomised patients)

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 Progression Free Survival (PFS) by Blinded Independent Central Review (BICR), measured over RECIST tumour assessments every 6 weeks from randomisation until disease progression as defined by Recist 1.1 and confirmed by BICR..

PFS measures the time from randomization until disease progression or death in the absence of progression. Progression was defined using RECIST 1.1 and had to be confirmed by BICR. The registry description states that progression could be a sufficient increase in the summed diameters of target lesions, progression of non-target lesions, or a new lesion. It also states that participants remained eligible to have a PFS event counted even if they withdrew from randomized therapy or received another anti-cancer therapy before progression.

PFS is used in cancer trials because it captures whether disease has progressed or death has occurred over follow-up, without requiring that the study wait for every participant to die. General statistical context: a time-to-event endpoint uses both whether an event occurred and the time until it occurred. Its median is the time by which an estimated half of the population has experienced the event; it is not an average and does not describe the experience of every participant.

The registry reports median PFS of 7.0 months for savolitinib, with a reported lower confidence limit of 2.8 months and no upper limit reported in the registry record. For sunitinib, the median is 5.6 months, with a 95% confidence interval from 4.1 to 6.9 months. The comparison is also reported as a hazard ratio, which summarizes the relative event rate over follow-up under the model used. These summaries address different aspects of the time-to-event data.

7. Results Posted in the Registry

• Savolitinib QD vs Sunitinib QD: hazard ratio 0.71 (95% CI 0.37–1.36), reported p = 0.313 (as reported in the registry).

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.

Objective Response Rate (ORR) by BICR

• 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:
  • Sunitinib QD: 7.4% (2/27) versus 27.3% (9/33) on control; risk difference -19.9 percentage points (95% CI -38.0 to -1.7), risk ratio 0.27 (0.06–1.15); unadjusted two-proportion z-test p = 0.048.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the proportion with the event is statistically lower for Sunitinib QD than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

Disease Control Rate (DCR) at 6 Months by BICR

• 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:
  • Sunitinib QD: 37.0% (10/27) versus 48.5% (16/33) on control; risk difference -11.4 percentage points (95% CI -36.4 to 13.5), risk ratio 0.76 (0.42–1.40); unadjusted two-proportion z-test p = 0.373.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Sunitinib QD and the control. The interval gives the range of effects that remain compatible with the data.

Disease Control Rate (DCR) at 12 Months by BICR

• 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:
  • Sunitinib QD: 22.2% (6/27) versus 30.3% (10/33) on control; risk difference -8.1 percentage points (95% CI -30.3 to 14.1), risk ratio 0.73 (0.31–1.76); unadjusted two-proportion z-test p = 0.481.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Sunitinib QD and the control. The interval gives the range of effects that remain compatible with the data.

The posted secondary outcomes show a broadly similar numerical direction for the reported binary endpoints: ORR and disease control at both assessed time points are numerically higher for savolitinib than for sunitinib. These outcomes describe response or disease status at assessment points and do not measure the same time-to-event quantity as the primary PFS endpoint.

This directional pattern can be considered alongside the primary result, but it does not resolve the uncertainty in the PFS comparison. The primary hazard-ratio confidence interval includes 1, and the log-rank p-value is 0.313. The secondary outcomes are not interchangeable with PFS and should not be treated as a replicated primary result. Their estimates are also uncertain, as reflected by their confidence intervals.

The registry lists overall survival and duration of response among the secondary outcomes, but numerical results for those endpoints are not reported in the registry record. Thus, the posted secondary evidence is incomplete with respect to the full set of listed outcomes.

The posted secondary comparisons are unadjusted recalculations, and the registry record does not report a multiplicity strategy. General statistical context: when several outcomes are tested, the chance of at least one apparently positive result can increase unless the analysis accounts for multiple testing. This means that unadjusted secondary p-values carry less confirmatory weight than a prespecified primary analysis. The secondary pattern is informative as descriptive context, but it should not be read as proof of a treatment effect or as a correction to the primary result.

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

• Overall Survival (OS)
• Objective Response Rate (ORR) by BICR
• Duration of Response (DoR) by BICR
• Disease Control Rate (DCR) at 6 Months by BICR
• Disease Control Rate (DCR) at 12 Months by BICR

9. Statistical Interpretation

10. Statistical Methods Explained

The primary endpoint is a time-to-event outcome analyzed using a log-rank method, with the treatment comparison expressed as a hazard ratio. General statistical explanation follows; it describes how these methods work and does not add trial-specific analyses beyond those reported in the registry record.

What does the hazard ratio describe?

A hazard is the instantaneous event rate among participants who have not yet experienced the event, considered over follow-up. A hazard ratio compares those rates between groups. A value of 1 corresponds to equal rates under the comparison; a value below 1 indicates a lower estimated event rate in the first-named group, while a value above 1 indicates a higher estimated event rate. This is a relative summary over time, not a difference in median months and not a direct statement about an individual participant’s outcome.

For savolitinib versus sunitinib, the reported hazard ratio is 0.71. Its interpretation depends on the time-to-event comparison and the assumptions underlying the analysis. The registry record reports the estimate and confidence interval but does not report further diagnostics about those assumptions or the detailed model specification.

What does the log-rank test contribute?

The log-rank test compares the observed pattern of events over follow-up between groups. It accounts for the ordering and timing of events and can include participants whose follow-up ends before an event is observed. General statistical context: censoring occurs when event status is not observed after a participant’s last known follow-up; valid interpretation depends on how censoring relates to the event process. The registry record does not report detailed censoring summaries or the rules used to handle missing event information.

The sponsor-reported log-rank p-value is 0.313. A p-value measures how surprising data at least as incompatible with the null hypothesis would be under that hypothesis and the analysis assumptions. It is not the probability that the null hypothesis is true, and it does not measure the size or clinical importance of an effect. The p-value does not establish equivalence or prove that no difference exists.

How should the median and confidence interval be read?

The median PFS is the estimated time at which the time-to-event curve reaches the point corresponding to half the population remaining event-free. It describes the curve at one location and does not capture its full shape. The reported medians are 7.0 months for savolitinib and 5.6 months for sunitinib. The 95% confidence interval for the savolitinib median has a lower limit of 2.8 months; its upper limit is not reported in the registry record. The sunitinib median has a 95% confidence interval from 4.1 to 6.9 months.

A confidence interval conveys uncertainty in an estimate under the statistical method used. General statistical explanation: a 95% confidence procedure would cover the underlying parameter in 95% of repeated applications under its assumptions. It does not mean that there is a 95% probability that this particular interval contains the parameter. For the treatment comparison, the hazard-ratio interval from 0.37 to 1.36 includes 1, so values consistent with the data include no relative difference as well as effects in either direction.

What role does blinded central review play?

The primary endpoint is assessed by BICR, while the trial is listed as open label with no masking. Central review can apply a common assessment process to imaging outcomes and, when reviewers are blinded to treatment assignment, can reduce the influence of knowledge of treatment on outcome classification. The registry identifies the review as blinded and independent. Further details about reviewer procedures, adjudication, or agreement are not reported in the registry record.

Open-label treatment assignment remains relevant to the overall study context. Knowledge of treatment can influence participant or care-team behaviour, while central review addresses the review of tumour assessments. These are separate aspects of trial conduct; the existence of BICR does not make the treatment comparison masked.

What can be inferred from the secondary comparisons?

The posted binary secondary outcomes are reported as percentages with 95% confidence intervals. The separate simple comparisons are unadjusted recalculations from the posted summaries; no sponsor-reported analyses for those secondary outcomes are listed. A risk difference is the difference between the two group proportions under the stated ordering. Its confidence interval describes uncertainty in that difference. An interval that includes zero is compatible with no difference under the method used, as well as with differences in either direction.

The unadjusted p-values for secondary outcomes do not account for the fact that multiple outcomes are examined. General statistical context: testing multiple hypotheses increases the chance of at least one apparently unusual result even when no underlying effect exists. A prespecified multiplicity method can control that risk, but such a method is not reported in the registry record. The unadjusted secondary comparisons should therefore be interpreted as exploratory evidence rather than as a set of confirmatory tests.

The registry record does not report the full statistical analysis plan, details of missing-data handling, or interim analyses. Those omissions limit evaluation of how outcome definitions and analysis decisions were prespecified. The posted primary analysis remains the sponsor-reported log-rank comparison, and the secondary recalculations should not be mistaken for additional sponsor analyses.

11. Confidence Intervals and Precision

The primary comparison is imprecise. The reported hazard ratio is 0.71, and its 95% confidence interval extends from 0.37 to 1.36. Because the interval includes 1, the estimate is compatible with no relative difference and with effects in either direction. It does not support a precise estimate of the magnitude of any treatment difference.

The median estimates also have uncertainty. For savolitinib, the reported lower confidence limit is 2.8 months, while an upper limit is not reported in the registry record. For sunitinib, the 95% confidence interval is 4.1 to 6.9 months around a median of 5.6 months. The difference in how much interval information is posted means the precision of the two median estimates cannot be fully compared from the listed limits alone.

General statistical context: confidence intervals tend to be wider when the data contain less information about an endpoint. In time-to-event analyses, the amount and timing of observed events, as well as censoring, contribute to information. The registry record does not provide all event and follow-up details needed to assess those contributors directly. The reported arm summaries are based on 33.0 and 27.0 participants, which also places the estimates in the context of a modest-sized study.

Precision is not the same as whether a result is statistically significant. The p-value of 0.313 and the interval spanning 1 both indicate that the primary comparison does not yield a statistically clear result. Even when an estimate points in one direction, uncertainty can remain broad enough that materially different effects are compatible with the data.

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 planning illustrations use different units for the primary and secondary outcomes. For the time-to-event PFS comparison, the figures are total numbers of events, not participants per arm. For the binary secondary outcomes, the figures are participants per arm. The trial’s own original sample-size assumptions and justification are not reported in the registry record.

The listed planning assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. The target effect is set at the stated fraction of the observed effect. The 80% power column describes the planned information level that, under these assumptions and a specified effect, would give an 80% chance of detecting that effect using the stated test. The 90% power column uses a higher chance of detection under the same planning framework.

For example, the PFS planning figures at 100% of the observed effect are 268 total events for 80% power and 359 total events for 90% power. The 75% and 50% effect scenarios require more events. For ORR, the corresponding planning figures are participants per arm; at 100% of the observed effect they are 56 for 80% power and 75 for 90% power. The DCR planning illustrations follow the same per-arm convention.

Smaller effects are harder to distinguish from random variation, so the planning illustrations require more participants or events when the assumed effect is reduced. These figures are planning illustrations based on the observed results and stated assumptions, not the trial’s original design targets. They are not observed power and do not show the probability that the trial’s conclusion is correct.

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

d is the total number of events needed (Schoenfeld's formula) with 1:1 allocation and proportional hazards.

Worked calculation for this trial

1. Hazard ratio HR = 0.710, so ln(HR) = -0.3425.
2. d = 4 × (1.96 + 0.842)² / [ln(0.710)]² = 267.7, rounded up to 268 events in total 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
hr <- 0.71                              # observed hazard ratio
events_needed <- function(power, frac = 1) ceiling(round(4 * (z_a + qnorm(power))^2 / log(hr^frac)^2, 6))   # Schoenfeld, 1:1 allocation
for (f in c(1, 0.75, 0.5))
  cat(sprintf("%3.0f%% of observed effect: events = %s (80%% power), %s (90%% power)\n", 100 * f,
              format(events_needed(0.80, f), big.mark = ","), format(events_needed(0.90, f), big.mark = ",")))

# Simulation: given d total events under proportional hazards, events in the treatment arm are Binomial(d, hr / (1 + hr)).
# The score (log-rank type) statistic is (D1 - d / 2) / sqrt(d / 4).
set.seed(3091192)
d    <- events_needed(0.80)
reps <- 20000
D1 <- rbinom(reps, d, hr / (1 + hr))
z  <- (D1 - d / 2) / sqrt(d / 4)
cat(sprintf("Simulated power with %s events: %.3f (%d replicates)\n", format(d, big.mark = ","), mean(abs(z) > z_a), reps))

Output from running the code

100% of observed effect: events = 268 (80% power), 359 (90% power)
 75% of observed effect: events = 476 (80% power), 637 (90% power)
 50% of observed effect: events = 1,071 (80% power), 1,434 (90% power)
Simulated power with 268 events: 0.780 (20000 replicates)

The simulation uses the fact that, with d events in total and proportional hazards, the number of events in the treatment arm is binomial. It is an approximation that conditions on the total number of events, and it should land close to the target power.

14. Group-Sequential Design Analysis

Many time-to-event trials, especially in oncology, stop or report early when interim data are overwhelming, so a single fixed-sample number understates the real design problem. The calculation below is an illustrative group-sequential design built from this trial's posted control median and hazard ratio. The registry record does not state the planned number of looks or the spending function, so the design is a planning illustration, not a reconstruction of the protocol. Assumed: one-sided α = 0.025, 90% power, three equally spaced looks on information, Lan-DeMets O'Brien-Fleming spending, 24 months of accrual, 60 months in total and a 1% monthly dropout hazard.

Under these assumptions the design needs about 400 participants in total, with the boundaries below.

Fixed design versus group-sequential alternatives

Group-sequential designs cost a small increase in maximum sample size and events compared with the fixed design, in exchange for the chance to stop early for overwhelming efficacy. O'Brien-Fleming spending keeps early boundaries very conservative and the final boundary close to the fixed-design value, whereas Pocock spending spends more alpha early.

Sensitivity to a smaller true effect

Planning for exactly the observed effect is optimistic, because observed effects in published trials tend to be somewhat larger than the true effect. The table repeats the main design (three looks, O'Brien-Fleming spending, 90% power) for smaller true hazard ratios.

For context, this trial enrolled 60 participants.

Cross-check, upper boundaries: gsDesign 3.71, 2.511, 1.993 | rpact 3.71, 2.511, 1.993. Agreement between the two packages confirms the boundary calculation.

R code and output

suppressWarnings(suppressMessages({ library(gsDesign); library(rpact) }))
med_control <- 5.6          # control-arm median, months (converted from the posted unit)
hr          <- 0.71          # hazard ratio used for planning
lambdaC     <- log(2) / med_control
eta         <- 0.01               # dropout hazard per month (assumption)
acc_months  <- 24                 # accrual duration (assumption)
study_months<- 60                 # total study duration (assumption)

design <- function(k, sfu, beta, h = hr) gsSurv(k = k, test.type = 1, alpha = 0.025, beta = beta, sfu = sfu,
  lambdaC = lambdaC, hr = h, hr0 = 1, eta = eta, gamma = 1, R = acc_months, T = study_months, minfup = study_months - acc_months)

# Main design: 3 looks, Lan-DeMets O'Brien-Fleming spending, one-sided alpha 0.025, 90% power
x <- design(3, sfLDOF, 0.10)
cat(sprintf("N_TOTAL|%s\n", format(ceiling(sum(x$gamma * x$R)), big.mark = ",")))
for (i in 1:3) cat(sprintf("GS|%d|%.1f|%.0f|%.3f|%.5f|%.5f\n", i, x$T[i], x$n.I[i], x$upper$bound[i], pnorm(-x$upper$bound[i]), cumsum(x$upper$spend)[i]))
cat("\n"); print(gsBoundSummary(x))

# Comparison of designs
cmp <- list(list("Fixed design, one analysis", 1, sfLDOF), list("Two looks, O'Brien-Fleming", 2, sfLDOF),
            list("Three looks, O'Brien-Fleming", 3, sfLDOF), list("Three looks, Pocock", 3, sfLDPocock))
cat("\n")
for (b in c(0.20, 0.10)) for (d in cmp) {
  y <- design(d[[2]], d[[3]], b)
  cat(sprintf("DES|%s|%d|%.0f|%s|%.1f\n", d[[1]], round(100 * (1 - b)), tail(y$n.I, 1), format(ceiling(sum(y$gamma * y$R)), big.mark = ","), tail(y$T, 1)))
}

# Sensitivity: what if the true effect is smaller than the observed one? (3 looks, O'Brien-Fleming, 90% power)
cat("\n")
for (fr in c(1, 0.75, 0.5)) {
  y <- design(3, sfLDOF, 0.10, exp(log(hr) * fr))
  cat(sprintf("SENS|%.2f|%.3f|%.0f|%s\n", fr, exp(log(hr) * fr), tail(y$n.I, 1), format(ceiling(sum(y$gamma * y$R)), big.mark = ",")))
}

# Independent cross-check of the boundaries with rpact
r <- getDesignGroupSequential(kMax = 3, alpha = 0.025, beta = 0.10, sided = 1, typeOfDesign = "asOF", informationRates = x$timing)
cat(sprintf("\nCross-check, upper boundaries: gsDesign %s | rpact %s\n", paste(round(x$upper$bound, 3), collapse = ", "), paste(round(r$criticalValues, 3), collapse = ", ")))

Method: LachinFoulkes 
    Analysis               Value Efficacy
   IA 1: 33%                   Z   3.7103
      N: 252         p (1-sided)   0.0001
 Events: 121        ~HR at bound   0.5088
   Month: 15    P(Cross) if HR=1   0.0001
             P(Cross) if HR=0.71   0.0338
   IA 2: 67%                   Z   2.5114
      N: 400         p (1-sided)   0.0060
 Events: 242        ~HR at bound   0.7237
   Month: 24    P(Cross) if HR=1   0.0060
             P(Cross) if HR=0.71   0.5603
       Final                   Z   1.9930
      N: 400         p (1-sided)   0.0231
 Events: 362        ~HR at bound   0.8109
   Month: 60    P(Cross) if HR=1   0.0250
             P(Cross) if HR=0.71   0.9000



Cross-check, upper boundaries: gsDesign 3.71, 2.511, 1.993 | rpact 3.71, 2.511, 1.993

15. Analyses Reported by the Sponsor

For the primary PFS outcome, the registry reports the sponsor’s log-rank analysis: a hazard ratio of 0.71, a 95% confidence interval from 0.37 to 1.36, and a p-value of 0.313. This is the primary inferential result to use when describing the sponsor’s analysis. The recalculation material does not replace or supersede this reported analysis.

The secondary endpoint comparisons are different in status. No sponsor-reported analyses are listed for the posted ORR and DCR results. The simple comparisons described for those outcomes are unadjusted recalculations from posted group summaries. They can help readers understand the size and uncertainty of the observed group differences, but they are not sponsor analyses and do not carry the same role as the reported primary analysis.

The registry record does not report a full statistical analysis plan that would allow a more detailed comparison of planned and posted methods. It also does not report sponsor analyses for overall survival or duration of response. The primary interpretation should therefore remain anchored to the reported hazard ratio, its confidence interval, and the log-rank p-value, with secondary comparisons treated as additional descriptive information.

16. Multiplicity, Missing Data and Analysis Populations

The registry record identifies one primary endpoint, PFS by BICR, and lists several secondary outcomes. However, it does not report whether the statistical analysis plan controlled the overall type I error across multiple endpoints or specified an order for testing them. A multiplicity procedure is therefore not reported in the registry record.

General statistical context: when researchers examine several outcomes, each test creates an opportunity to observe a result that appears unusual by chance. A prespecified method can control the probability of false-positive conclusions across a defined family of tests. Without information about such a method, unadjusted secondary p-values should be treated as exploratory rather than as independent confirmatory evidence. The record’s posted secondary results do not include sponsor-reported analyses.

The primary population is identified as the Full Analysis Set, defined as all randomized patients. The registry record does not report further detail about exclusions from analyses, the handling of participants with incomplete assessments, or how missing outcome data were addressed. For PFS, it gives the event definition and assessment schedule, but not detailed rules for censoring or loss to follow-up.

Interim analyses are not reported in the registry record. Nor are the full statistical analysis plan, prespecified decision thresholds beyond the information in the posted analysis, or any plan for managing missing data. These omissions limit the ability to assess whether the posted analyses followed a complete prespecified framework. They do not change the numerical result that is reported for the primary outcome, but they restrict how fully its analysis process can be independently evaluated from the registry record.

17. 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 (3.7%), thrombocytopenia (3.7%), atrial fibrillation (3.7%), cardiac failure (3.7%), myocardial infarction (3.7%).

The registry lists serious adverse events affecting 8 participants in the savolitinib group and 8 in the sunitinib group. The at-risk counts are 33 and 27, respectively. These are reported counts; the registry summary does not provide an adjusted comparison of serious adverse-event risk. The most commonly listed serious-event terms are pneumonia, thrombocytopenia, atrial fibrillation, cardiac failure, myocardial infarction, and hypertensive urgency, each reported as 3.7%. The record does not specify the group attribution for these listed terms in the summary.

Other adverse events are listed for 25 participants in the savolitinib group and 27 in the sunitinib group. Deaths are listed for 9 and 13 participants, respectively. These counts do not by themselves establish that treatment caused an event or death, and they do not account for differences in follow-up time or exposure. Detailed event definitions, severity distributions, exposure-adjusted comparisons, and causal assessments are not reported in the registry record. The safety counts should therefore be read as descriptive registry summaries, not as a comparative safety analysis.

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

The posted results describe a randomized comparison of savolitinib and sunitinib for MET-driven, unresectable and locally advanced or metastatic PRCC. For the primary PFS endpoint, the reported medians are 7.0 months and 5.6 months, while the sponsor-reported hazard ratio is 0.71 with a 95% confidence interval from 0.37 to 1.36 and a p-value of 0.313. The point estimate is in one direction, but the uncertainty is substantial and the interval includes 1.

This result does not establish that the groups are equivalent, and it does not establish a clear difference in PFS. A non-significant result is not proof of no effect; it means the analysis did not provide sufficiently clear evidence against the null hypothesis under the reported method. The interval also does not identify a precise effect size.

The numerical pattern in the posted response and disease-control outcomes is additional descriptive evidence, not a substitute for the primary time-to-event analysis. Those secondary comparisons are unadjusted, and multiplicity handling is not reported. Numerical results for overall survival and duration of response are not reported in the registry record.

The result applies to the study population and analysis as recorded, not automatically to every person with PRCC or to other settings. The registry record describes an open-label trial, central review for the primary outcome, and a Full Analysis Set of all randomized patients. It does not report enough detail to assess all eligibility, missing-data, or analysis-plan decisions. These features matter when judging how far the results can be generalized or how completely the analysis can be evaluated.

These statistics describe the posted trial evidence. They do not provide medical advice or determine an individual treatment choice.

19. Why This Trial Matters Statistically

This trial is statistically informative because it compares two treatments in a defined MET-driven PRCC population using randomization and a time-to-event primary endpoint. Randomization supports a group comparison by reducing systematic differences at assignment, while the PFS endpoint captures progression or death over follow-up. The use of BICR is relevant in an open-label study because the primary outcome depends on tumour assessments.

The results also illustrate why an effect estimate must be read together with its precision. The reported hazard ratio is 0.71, but its 95% confidence interval from 0.37 to 1.36 includes 1, and the log-rank p-value is 0.313. The point estimate alone would give an incomplete picture; the interval shows that the comparative effect remains uncertain.

The secondary results highlight a further statistical issue. Several outcomes can provide useful descriptive context, but unadjusted comparisons across multiple endpoints have less confirmatory force than a prespecified primary test. The registry record does not report a multiplicity strategy, and some listed secondary outcomes lack numerical results in the posted record. A careful interpretation distinguishes what was measured, what was analyzed, and what remains unavailable.

Finally, the planning illustrations show how required information depends on the assumed effect size and desired power. They are not the original trial design rationale, which is not reported in the registry record, and they are not observed power. Keeping those planning calculations separate from the trial’s actual evidence is essential for understanding what the results can and cannot support.

20. 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 reported group summaries are based on 33.0 participants in the savolitinib arm and 27.0 in the sunitinib arm, and the primary hazard-ratio interval is broad enough to include 1 and values on either side. This limits precision. The savolitinib median PFS has a reported lower confidence limit but no upper limit reported in the registry record, leaving its interval description incomplete.

The study is listed as open label with no masking. The primary endpoint was assessed by BICR, which addresses the review of tumour outcomes, but does not remove every possible influence of open-label treatment assignment. Details of how the central review was conducted are not reported in the registry record.

The record does not report the original sample-size justification, detailed missing-data or censoring procedures, interim analyses, or a complete statistical analysis plan. It also does not report a multiplicity strategy for the secondary outcomes. Numerical results are not reported for overall survival or duration of response, despite those outcomes being listed. These gaps restrict assessment of the full analysis framework and of the broader outcome profile.

The safety information is presented as counts and listed event terms, without exposure-adjusted comparisons or detailed follow-up context. The record also does not provide enough information in its summary to attribute each listed serious-event term to a particular treatment group. These limitations mean the safety counts should not be interpreted as a definitive comparative assessment.

21. Related Tutorials

22. Related Statistical Calculators

23. Sources

• ClinicalTrials.gov: Savolitinib vs. Sunitinib in MET-driven PRCC. — NCT03091192.
• Registry record last updated 2026-07-14; status last verified 2026-07.
• Results first posted 2020-07-10.
• 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.

24. Record Summary

The registry describes a Phase III, randomized, open-label, parallel-group study of savolitinib versus sunitinib in patients with MET-driven, unresectable and locally advanced or metastatic PRCC. Enrollment is listed as 60, and the status is active, not recruiting.

The primary endpoint is PFS by BICR. The sponsor-reported log-rank analysis gives a hazard ratio of 0.71, a 95% confidence interval from 0.37 to 1.36, and a p-value of 0.313. The interval includes 1, so the primary comparison remains uncertain.

Posted secondary results show numerical differences in ORR and disease control, but these comparisons are unadjusted. Numerical overall-survival and duration-of-response results are not reported in the registry record. The record also does not report multiplicity handling, the original sample-size rationale, detailed missing-data procedures, or interim analyses.