Positive Imaging Study for Distant Metastases in Patients With Castration-Resistant Prostate Cancer: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Positive Imaging Study for Distant Metastases in Patients With Castration-Resistant Prostate Cancer (NCT01981109), a registered phase not stated study. The primary outcome is Occurrence Measured in Time of an Image Showing Positive for M1 (Metastatic Disease) at Baseline to Registration.

2. Clinical Question

3. Trial Design

This was a study of imaging findings in patients with castration-resistant prostate cancer and no known distant metastases. Its stated purpose was to describe patient characteristics predictive of an imaging study positive for distant metastases. The record does not describe randomization, an assigned control group, or a like-for-like treatment comparison. Allocation is not stated, and the M0 and M1 classifications are imaging results rather than assigned study arms.

The registry lists an enrollment of 225. At baseline, 206 patients had evaluable data, with 151 classified as non-metastatic (M0) and 55 classified as metastatic (M1). The record says that patients in the primary analysis population had a current diagnosis of M0 disease at baseline, underwent imaging, and were assessed as M1 or M0. The relationship between that stated population definition and the reported M1 count is not explained further.

The study started in 2012-01, had a primary completion date of 2017-07, and is listed as terminated. The reason for termination and the point at which termination occurred are not reported in the registry record. The record explicitly states that the study terminated early and that no further statistical analysis was performed. It does not describe follow-up beyond the baseline imaging assessment, nor does it report a completed analysis of which patient characteristics predicted positive imaging.

The record lists one outcome measure. It does not provide a detailed statistical analysis plan, a sample-size justification, or an explanation of whether the enrollment figure was a target or a final count beyond its listing as enrollment. The reported design information therefore supports a descriptive reading of baseline imaging classifications, not a conclusion about comparative treatment effects or validated prediction.

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: There were 206 patients at baseline that had a valuable data that were assessed. The results were 151 patients that had non-metastatic (M0) and 55 patients that had metastatic disease (M1) disease. The study terminated early therefore no further statistical analysis was performed.

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 Occurrence Measured in Time of an Image Showing Positive for M1 (Metastatic Disease) at Baseline to Registration, measured over Each enrolled subject will be evaluated at baseline for the occurrence of metastatic disease by positive imaging scan..

The primary endpoint is described in the endpoint definition as whether an imaging study at baseline was positive for M1 metastatic disease, recorded as yes or no. The registry describes the primary analysis population as patients with a current M0 diagnosis at baseline who underwent imaging and were assessed as M1 or M0. This binary classification is intended to capture the presence or absence of distant metastatic disease on the baseline scan.

Such an endpoint can address the study’s stated descriptive aim: characterizing baseline imaging findings and identifying patient characteristics that might be associated with a positive scan. The reported counts are 151 patients classified as M0 and 55 as M1 among 206 patients with evaluable baseline data. No estimates of associations between characteristics and imaging classification are reported.

There is an important reporting inconsistency. The endpoint title refers to occurrence measured in time from baseline to registration, with years as the unit, while the endpoint description defines a yes-or-no baseline scan result. The registry also labels the outcome as continuous and specifies a mean and standard deviation. Those metadata do not match the binary classification in the description and reported counts. No mean, standard deviation, or time-to-event estimate is reported, so the endpoint should be interpreted cautiously as a baseline imaging classification rather than as a measured duration.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The registry gives a binary endpoint definition but also labels the outcome continuous, with mean as the summary measure, standard deviation as the dispersion measure, and years as the unit. These descriptions point to different statistical approaches. A binary endpoint records membership in one of two categories, such as M1 or M0. A continuous endpoint records a quantity that can vary across a range, for which a mean and standard deviation may be suitable summaries. The record does not clarify the discrepancy or report a continuous measurement. The analysis described by the endpoint definition is therefore best understood as classification at baseline, while recognizing that the registry’s metadata are inconsistent.

What do the reported counts describe?

The reported values are counts in the assessed baseline group: 151 patients were classified as M0 and 55 as M1, among 206 patients with evaluable data. Counts preserve the number of people in each reported category. They do not quantify an association between patient characteristics and scan classification. No cross-tabulations by characteristic, model coefficients, or group-specific estimates are reported.

General statistical explanation: for a binary outcome, an analysis might describe the proportion in each category and quantify uncertainty around that proportion. A confidence interval can communicate the range of values reasonably compatible with the observed data under a specified statistical model. Neither a proportion estimate nor a confidence interval is reported in the registry record, and no additional value should be inferred from the counts alone in this account.

What would a predictive analysis address?

The study’s stated purpose refers to patient characteristics predictive of a positive imaging study. In general, prediction requires an explicit outcome definition and a method relating candidate characteristics to that outcome. A fitted model can estimate how the predicted outcome varies with recorded characteristics; its usefulness also depends on adequate data, clearly defined predictors, and assessment of how well predictions perform. The registry record does not report the candidate characteristics, model form, selection procedures, estimated associations, or predictive performance. It states that no further statistical analysis was performed, so none of those steps can be attributed to this study’s posted results.

General statistical explanation: an association in a fitted model is not automatically a reliable prediction for new patients. Model assessment can examine how closely predictions correspond to observed outcomes and whether performance holds in data not used to fit the model. No such assessment is reported. There is also no basis in the record for distinguishing a prespecified predictor analysis from exploratory examination of characteristics.

Why does the endpoint mismatch matter?

A mean and standard deviation summarize a continuous measurement by describing its center and spread. For a yes-or-no endpoint, those summaries do not directly express the two categories in the way counts or a proportion do. A time-to-event analysis is different again: it requires a defined event and timing information, and may need to account for people whose event status is not observed during follow-up. The title refers to time from baseline to registration, but the description refers to a baseline M1 yes-or-no scan result. No event times, follow-up structure, or continuous endpoint results are reported. It is therefore not possible to determine whether the title, the outcome metadata, or the endpoint description best reflects the intended analysis.

General statistical explanation: over-dispersion matters when outcome variation is greater than a model’s assumed variability. It can make standard errors too small if ignored, leading to overly narrow uncertainty intervals or misleading tests. Whether over-dispersion is relevant depends on the outcome and model. Because this record reports no fitted model or inferential analysis, over-dispersion cannot be evaluated as a feature of the reported analysis.

What can be said about planning calculations?

The registry record includes assumptions for a planning illustration: two-sided significance level 0.05, equal allocation between two arms, and a normal approximation, with the effect to detect expressed as a stated fraction of the observed effect. It distinguishes total event counts for time-to-event illustrations from participants per arm for other illustrations. These are planning assumptions, not evidence that the trial used those methods in its design. The record does not report the trial’s own sample-size justification, and no planning count is stated in the registry record for this account. Any such illustration is not observed power and does not replace the absent analysis of the baseline imaging results.

10. Confidence Intervals for Each Arm

No two-arm comparison is shown for this trial, but each arm's mean can be given with a t-based 95% confidence interval computed from the posted mean, spread and number of participants.

R code and output

arm <- c("M0- No Distant Metastases", "M1-Distant Metastases")
n   <- c(151, 55)
m   <- c(10.9, 8.8)
sp  <- c(6.4, 5.9)
sd  <- sp      # posted standard deviations
for (i in seq_along(n)) {
  se <- sd[i] / sqrt(n[i]); h <- qt(0.975, n[i] - 1) * se
  cat(sprintf("CARM|%s|%d|%.3f|%.3f|%.3f|%.3f\n", arm[i], n[i], m[i], sd[i], m[i] - h, m[i] + h))
}

CARM  M0- No Distant Metastases  151  10.900  6.400  9.871  11.929
CARM  M1-Distant Metastases  55  8.800  5.900  7.205  10.395

11. Multiplicity, Missing Data and Analysis Populations

The record lists one outcome measure and zero analyses. It does not report a multiplicity adjustment, a hierarchy of endpoints, or a plan for controlling error across multiple tests. A single listed outcome does not establish that only one hypothesis or analysis was considered, particularly because the study’s stated purpose also refers to patient characteristics that might predict positive imaging. The record does not describe how many characteristics were examined or planned for examination.

The reported baseline results include 206 patients with evaluable data, compared with enrollment of 225. The reason for the difference is not reported in the registry record. The record does not say whether data were missing, whether patients were excluded, whether scans were unavailable, or whether another issue affected evaluability. It also does not give the characteristics or imaging status of patients without evaluable baseline data. The possible impact on representativeness and estimates cannot therefore be assessed.

The primary analysis population is described as patients with a current M0 diagnosis at baseline who underwent imaging. However, the reported results include both M0 and M1 classifications. The record does not explain how the M0 eligibility description relates to the inclusion of M1 classifications in the reported baseline results. It also does not state whether all enrolled patients met the stated eligibility description or how population membership was determined.

No missing-data method, imputation procedure, sensitivity analysis, or analysis-population accounting is reported. The study’s early termination and lack of further statistical analysis mean that no completed inferential analysis can be evaluated for its treatment of missing observations or multiple testing.

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

The result documents baseline imaging classifications among the patients with evaluable data: 151 were reported as M0 and 55 as M1, out of 206 assessed. This is a descriptive account of the classifications recorded in that group. It does not provide an estimate of the probability of a positive scan in a wider population, because no uncertainty interval or inferential analysis is reported.

The result does not answer the stated predictive question. No patient characteristics are linked to M1 classification in the posted results, and no predictive model or validation assessment is reported. The study record says that it terminated early and that no further statistical analysis was performed. As a result, the descriptive counts cannot be treated as a completed analysis of predictors.

The endpoint’s differing descriptions also limit what the result can mean. The definition describes a baseline yes-or-no scan outcome, while the title and metadata refer to time, a continuous mean, and years. No mean, standard deviation, or time-based outcome is reported. The counts support interpretation as categories, but do not establish a duration or rate of occurrence.

These data do not show that a treatment did or did not work. The record does not describe randomized treatment assignment or a formal comparator, and the two reported categories are imaging classifications rather than treatment groups. Nor do the results establish diagnostic accuracy: no reference standard or accuracy measures are reported. They also do not establish that any characteristic causes metastatic disease or a positive scan.

General statistical explanation: a sample result is an observation in the assessed group. Drawing conclusions beyond that group requires attention to how participants were selected, how much information is available, and how uncertainty was quantified. The record does not provide enough information to assess those issues fully. The counts remain informative as the reported baseline classifications, but broader conclusions should not be inferred from them alone.

13. Why This Trial Matters Statistically

Statistically, the record illustrates the distinction between a study question and a completed analysis. Its stated purpose is to describe characteristics predictive of positive imaging, but the posted results give only baseline M0 and M1 counts and state that no further statistical analysis was performed. A prediction question ordinarily needs more than a count of outcome categories: it requires information relating characteristics to the outcome, an analysis method, and some assessment of the resulting predictions. Those results are not reported.

The endpoint description also demonstrates why outcome definitions matter. A yes-or-no imaging classification, a continuous measure summarized by a mean and standard deviation, and a time-based outcome are not interchangeable. The registry’s wording contains elements of each, without reporting a result that resolves the inconsistency. Clear alignment among the endpoint name, definition, measurement unit, and analysis is important for interpreting any estimate.

The difference between enrollment and the evaluable baseline group is another statistical consideration. Without an account of why not all enrolled participants appear in the reported assessment, readers cannot determine whether evaluability was related to disease status or other characteristics. This affects how confidently descriptive findings can be generalized beyond the assessed group.

Finally, early termination and the absence of further analysis limit the record’s ability to address its predictive aim. The reported counts remain a useful description of the baseline classifications recorded, while the lack of uncertainty estimates and predictor analyses defines what cannot be concluded. This distinction—between observed data and the question the study set out to address—is central to careful interpretation.

14. 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 record describes early termination but does not report the reason, timing, or decision process. Without that information, it is not possible to determine how termination affected enrollment, data collection, or the intended analysis. The record states directly that no further statistical analysis was performed, leaving the predictive aim unaddressed in the posted results.

The endpoint is internally inconsistent across its title, metadata, and definition. The title refers to occurrence measured in time from baseline to registration; the unit is years; the outcome is labeled continuous with mean and standard deviation; but the description defines a baseline yes-or-no M1 imaging result. The reported findings are M0 and M1 counts, not a time-based or continuous estimate. The record does not explain this mismatch.

The analysis population description also needs clarification. It identifies the primary analysis population as patients with an M0 diagnosis at baseline, while the results report 55 patients classified as M1 among the assessed group. The registry does not explain how those classifications fit together or provide further population-accounting details.

Although enrollment is listed as 225, the baseline results cover 206 patients with evaluable data. Reasons for non-evaluability, missing-data handling, and the characteristics of those not assessed are not reported. No confidence intervals, hypothesis tests, model estimates, predictive performance measures, or reference-standard accuracy results are reported. The record does not provide enough detail to assess precision, potential selection effects, or the performance of any prediction approach.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: Positive Imaging Study for Distant Metastases in Patients With Castration-Resistant Prostate Cancer — NCT01981109.
• Registry record last updated 2019-08-28; status last verified 2019-07.
• Results first posted 2019-08-28.
• 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.

18. Record Summary

The record concerns a study of baseline imaging for distant metastases in patients with castration-resistant prostate cancer and no known distant metastases. Its stated aim was to describe characteristics predictive of positive imaging. Enrollment is listed as 225; 206 patients had evaluable baseline data, with 151 classified as M0 and 55 as M1.

The study is listed as terminated, with a start date of 2012-01 and primary completion date of 2017-07. The record states that it terminated early and that no further statistical analysis was performed. It lists one outcome measure and zero analyses.

The endpoint description defines a yes-or-no M1 imaging result at baseline, while the title and metadata refer to time, a continuous mean, standard deviation, and years. The record does not resolve this inconsistency. It reports no analysis of predictive characteristics, no inferential estimates, and no explanation for the difference between enrollment and the number with evaluable baseline data.