Florbetapir Calibration to the Centiloid Scale: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Florbetapir Calibration to the Centiloid Scale (NCT02120664), a non-randomized phase 1 study. The primary outcome is Florbetapir SUVr Conversion to Centiloid Units.

2. Clinical Question

3. Trial Design

NCT02120664, titled Florbetapir Calibration to the Centiloid Scale, is recorded as a Phase 1, completed study sponsored by Avid Radiopharmaceuticals. Its official title describes a multicenter, multicountry study to calibrate florbetapir (18F) PET imaging data to the Centiloid scale based on 11C-PiB. The listed study dates are 2014-04 for start and 2015-09 for primary completion. The registry reports enrollment of 35.

The design is classified as non-randomized, single-masked, and parallel. The registry describes amyloid-negative subjects and amyloid-positive subjects, with the latter including participants with a range of amyloid density. The amyloid-negative description refers to approximately 10 cognitively normal young subjects. The amyloid-positive description refers to approximately 25 subjects, including people clinically diagnosed with Alzheimer’s disease and people at risk for elevated amyloid density. These are approximate group descriptions; the record does not provide a detailed accounting that reconciles them with total enrollment.

The intended study task was measurement calibration rather than a randomized evaluation of treatment efficacy. Participants were to undergo PET imaging with florbetapir and 11C-PiB, and the primary measure was designed to express florbetapir SUVr in Centiloid units. The registry lists a primary endpoint timeframe of up to 70 minutes post injection. It also lists secondary outcomes concerning correlation between florbetapir and PiB Centiloid measurements and variability of PET images in young healthy control subjects.

Some operational details are absent. The registry record does not report the specific masking procedure, the imaging and conversion algorithms in sufficient detail for independent reproduction, or the allocation process beyond stating that allocation was non-randomized. The available record therefore identifies the broad study design and calibration objective, but does not fully document how measurements were generated and analyzed.

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: All participants receiving both a florbetapir and a PiB scan per protocol. One subject in the possible AD group did not complete a florbetapir scan due to excessive movement. One subject in the at risk elderly group elected to leave the study before the second PET scan.

5. Baseline Characteristics

The table reproduces the baseline characteristics posted in the registry, which describe the participants analysed at baseline and help show whether the randomized groups were comparable.

Population: All subjects receiving at least one dose of florbetapir

6. Primary Endpoint

The registry lists the primary outcome as Florbetapir SUVr Conversion to Centiloid Units, measured over up to 70 minutes post injection.

The primary endpoint is the conversion of florbetapir (18F) PET standard uptake value ratio, or SUVr, to Centiloid units. It is recorded as a mean continuous measure in Centiloids, with standard deviation as the dispersion measure and a timeframe of up to 70 minutes post injection. The description specifies a cortical average-to-cerebellum SUVr.

The Centiloid scale is intended to provide a common way to express amyloid PET measurements across tracers. In this record, its anchors are 0 and 100, based on median 11C-PiB SUVr values for representative groups of young cognitively and medically healthy controls and PiB-positive patients diagnosed with Alzheimer’s disease. The calibration question is whether florbetapir measurements can be expressed on that reference scale, allowing measurements made with different tracers to be compared in a common metric.

This is a measurement-calibration endpoint, not a clinical outcome such as symptoms or functioning. A successful conversion would require evidence about the relationship between tracer measurements and the conversion’s precision across the studied range. The registry record does not report the primary endpoint estimate, its standard deviation, or a confidence interval. Consequently, the posted information does not establish the observed calibration relationship or its precision.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The statistical task in this study is calibration of a continuous imaging measure. The endpoint is florbetapir SUVr converted to Centiloid units, with the scale defined using 11C-PiB reference measurements. This is different from estimating whether a treatment changes a clinical outcome: the central statistical question concerns the relationship between measurements obtained with the two PET tracers.

What does a continuous mean endpoint represent?

General explanation: a continuous variable can take values across a scale rather than belonging to only a set of categories. A mean summarizes the center of the observed values, while a standard deviation describes the spread of individual measurements around that mean. These are distinct summaries. A mean alone does not show how much participants’ measurements varied, and a standard deviation alone does not show the mean level.

For this endpoint, the registry specifies a mean in Centiloid units and standard deviation as the dispersion measure. It does not report either numerical result. It also does not report a confidence interval for the mean or for a conversion parameter. Therefore, readers cannot evaluate the observed central estimate or quantify its uncertainty from the posted results.

Why use a reference scale?

General explanation: a calibration scale provides a shared framework for interpreting measurements made using different instruments or methods. A conversion maps a measurement from one scale to another according to a defined reference. Here, the registry describes the Centiloid scale as anchored at 0 and 100 using median 11C-PiB SUVr values from representative young healthy controls and PiB-positive patients diagnosed with Alzheimer’s disease. Florbetapir measurements are the values intended to be expressed in those units.

A conversion can be evaluated in more than one way. Its average mapping, variation around the mapping, and performance across the measurement range may each matter. A high correlation between two measurements would mean that their values tend to move together; it would not necessarily show close agreement or absence of systematic differences. The registry lists correlation as a secondary outcome but gives no estimate or uncertainty for it.

What does the timeframe tell the reader?

The primary endpoint timeframe is up to 70 minutes post injection. This identifies the time window associated with the endpoint as recorded, but it does not, by itself, describe the exact acquisition schedule, image-processing steps, or timing rules. Those details are not reported in the registry record. The endpoint description does specify a cortical average-to-cerebellum SUVr.

How should the study design affect interpretation?

General explanation: random allocation is used in many trials to balance measured and unmeasured participant characteristics between intervention groups. This study is classified as non-randomized, and the arms are not a like-for-like control comparison. That is consistent with a calibration objective, where paired measurement values rather than treatment assignment are central. It also means treatment-effect interpretations are not appropriate for the design described.

The primary outcome population is defined as participants receiving both florbetapir and PiB scans per protocol. Paired measurements can be useful because the two tracer values can be related within the same participant. But the registry notes that one possible AD subject missed florbetapir imaging because of excessive movement and one at-risk elderly subject left before the second PET scan. The number of participants ultimately included in each analysis and the statistical handling of incomplete pairs are not reported.

What do the planning assumptions mean?

The registry’s sample-size assumptions describe an illustration based on the observed effect, not the trial’s original design. The illustration uses a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation; its target effect is a stated fraction of the observed effect. For time-to-event outcomes, the stated convention is total events, while other outcomes use participants per arm. This trial’s primary outcome is a continuous measure.

Such planning calculations are hypothetical estimates of sample size under stated assumptions. They are not observed power, and they do not replace the study’s outcome estimates or uncertainty intervals. The numerical planning results and the original sample-size justification are not reported in the registry record. The registry also does not report the protocol’s statistical analysis plan, interim analyses, or any adjustment strategy for multiple outcomes.

What can be learned from missing numerical results?

General explanation: an outcome estimate and its uncertainty are both needed to assess what the data indicate and how precisely they indicate it. Without a numerical estimate, there is no basis for judging the observed magnitude or direction. Without an uncertainty measure, there is no basis for judging how stable that estimate may be across plausible values. Because neither is reported for the primary calibration endpoint, the posted record describes the statistical question but does not answer it numerically.

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("Clincally Diagnosed AD", "Possible AD", "Mild Cognitive Impairment (MCI)", "At Risk Elderly", "Young Healthy Controls (YHC)")
n   <- c(10, 3, 7, 3, 10)
m   <- c(82.44, 50.75, 81.19, 14.93, 5.24)
sp  <- c(45.623, 60.383, 66.357, 35.885, 11.981)
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  Clincally Diagnosed AD  10  82.440  45.623  49.803  115.077
CARM  Possible AD  3  50.750  60.383  -99.250  200.750
CARM  Mild Cognitive Impairment (MCI)  7  81.190  66.357  19.820  142.560
CARM  At Risk Elderly  3  14.930  35.885  -74.213  104.073
CARM  Young Healthy Controls (YHC)  10  5.240  11.981  -3.331  13.811

11. Multiplicity, Missing Data and Analysis Populations

The registry lists three outcome measures and three analyses, but does not provide enough detail to identify the analysis corresponding to each listed outcome or to determine how the analyses were structured. It lists the primary conversion of florbetapir SUVr to Centiloid units and secondary outcomes concerning correlation with 11C-PiB Centiloid and variability of PET images in young healthy controls.

When several outcomes or analyses are examined, general statistical practice may address the increased chance of finding an apparently notable result through prespecified priorities or multiplicity adjustments. Whether such methods were planned or used is not reported in the registry record. The record also does not report a detailed statistical analysis plan, prespecified hypotheses for each secondary outcome, or interim analyses. No adjustment method should be inferred from the existence of multiple listed outcomes.

The primary outcome population is participants receiving both scans per protocol. The record identifies one possible AD subject who did not complete florbetapir imaging because of excessive movement and one at-risk elderly subject who left before the second PET scan. It does not report the final analysis population, the amount of missing endpoint data by group, or how incomplete scan pairs were treated. It also does not report whether any imputation or sensitivity analyses were performed.

These omissions matter because paired tracer measurements require the relevant measurements to be available for the same participants. General explanation: excluding participants with incomplete measurements can affect which participants contribute to an estimate, while imputation or other missing-data methods rely on assumptions that should be described. The registry’s brief account identifies missing scans but does not show what methods, if any, were used to address them.

12. Safety Results

The registry reports serious adverse events as affected participants divided by participants at risk.

Other (non-serious) adverse events

Participants with at least one non-serious event at or above the registry's 5% reporting threshold.

The adverse-event record separates participants by exposure group: Florbetapir Only, PiB Only, Both, and Total. For serious adverse events, the record lists 34 participants at risk in the Florbetapir Only group, 35 in the PiB Only group, 34 in the Both group, and 34 in the Total group. The number affected is recorded as 0 in each of these groups.

For other, non-serious adverse events, the at-risk counts are also listed as 34, 35, 34, and 34 for those same groups, and the number affected is recorded as 0 in each. These figures describe what the registry reports for the listed categories; they should not be treated as estimates of risk beyond the study population or as comparisons between randomized groups. The study was non-randomized, and the exposure categories are not a like-for-like control comparison.

Deaths are recorded as null rather than as a numerical count. A null entry does not establish a count of zero; deaths are not reported in the registry record. The record also lists no top serious-event terms. The adverse-event figures do not supply detailed event definitions, follow-up duration, or additional context about ascertainment, so the categories should be read narrowly as posted.

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

The study’s stated purpose was to demonstrate conversion of florbetapir (18F) PET SUVr measurements to Centiloid units, with 11C-PiB providing the reference scale. The record explains the scale anchors and identifies the primary outcome as a mean continuous measure in Centiloids. This establishes the intended measurement question and the broad framework used to address it.

The posted results do not include numerical estimates for the primary endpoint or the listed secondary outcomes. As a result, the registry record does not show the observed conversion, the strength of the relationship between the tracers, or the variability of the measurements. It also does not give confidence intervals or other uncertainty summaries for these outcomes. The conversion’s magnitude and precision therefore cannot be judged from the posted information.

This does not mean that no calibration was performed. It means that the numerical findings needed to assess the calibration are not reported in the registry record. Nor does the study’s completed status supply those missing results. A completed study can have a stated objective and an outcome definition without the registry entry presenting the estimate needed for statistical interpretation.

The design also sets limits on the kind of conclusion that is appropriate. This was a non-randomized imaging calibration study, not a randomized test of a treatment. Its group descriptions and scan procedures do not establish a treatment effect. The endpoint concerns measurement conversion, not a clinical benefit or harm outcome.

The record reports enrollment of 35 and notes two participants who did not complete both required scans for the primary outcome population. It does not state the final number analyzed or how incomplete scan data were handled. That leaves uncertainty about the analysis population in addition to uncertainty about the missing numerical results.

In statistical terms, the record defines a question but does not provide the estimates required to answer it. Conclusions about how well florbetapir SUVr maps onto Centiloid units, how reliable the mapping is, or how it performs across participants would require outcome results and analysis details not reported in the registry record.

14. Why This Trial Matters Statistically

Calibration studies address a statistical problem that arises when measurements from different tracers are intended to be interpreted on a common scale. A clearly defined reference scale can make measurement values easier to compare across imaging approaches. In this study, the Centiloid scale is anchored using 11C-PiB SUVr values for representative young healthy controls and PiB-positive patients diagnosed with Alzheimer’s disease, and florbetapir measurements are the target of conversion.

The study is therefore relevant as an example of the distinction between a measurement question and a treatment-effect question. Its stated aim is not to compare assigned therapies, but to relate imaging measurements. A non-randomized design may be appropriate to the described calibration purpose, while still requiring careful evaluation of the conversion method, its uncertainty, and the participants whose paired measurements contribute to the analysis.

Statistically, a conversion is more informative when its estimation method and performance are documented. The mean conversion, variability, association with the reference measure, and precision can answer different questions. A correlation outcome alone would not necessarily establish agreement. The registry lists correlation and image variability as secondary outcomes, but does not provide numerical findings for these measures.

The record also illustrates why endpoint definitions and analysis populations matter. It specifies the cortical average-to-cerebellum SUVr and identifies participants with both scans per protocol as the primary outcome population. It records two incomplete scan situations, but does not report the final analyzed sample or missing-data approach. Such details affect how readers understand the basis of a calibration estimate.

The statistical importance of the record is thus in its documented calibration objective and endpoint framework, rather than in a numerical result that can be evaluated. The registry entry does not report the estimates needed to assess whether the conversion was accurate or precise. Its design and scale description should not be mistaken for evidence about the conversion’s observed performance.

15. 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 most consequential limitation for interpreting the posted findings is that numerical results for the primary calibration endpoint are not reported in the registry record. The primary mean, its standard deviation, and a confidence interval are absent. Numerical findings for the listed secondary outcomes are also not reported. This prevents assessment of the observed conversion, its uncertainty, the association between tracer measurements, or the variability of images in young healthy controls.

The registry defines the primary outcome population as participants who received both scans per protocol, and it notes two incomplete scan situations. It does not report the number included in the final analysis, the handling of incomplete pairs, or any missing-data sensitivity analysis. The arm descriptions use approximate group sizes, while the record reports total enrollment of 35; a detailed participant flow or reconciliation is not reported.

The record’s design description is broad. It identifies a non-randomized, single-masked, parallel study, but does not report the specific masking procedure, the full allocation process, or the detailed image-processing and calibration algorithms. These omissions limit assessment of reproducibility and potential sources of measurement variation.

The registry lists three outcome measures and three analyses but does not provide their full analysis specifications. A prespecified statistical analysis plan, multiplicity approach, interim analyses, and original sample-size justification are not reported in the registry record. The planning assumptions described for sample-size illustrations do not establish what was planned for the trial or what was observed in its analyses.

Finally, the design addresses imaging calibration rather than a randomized treatment comparison. The posted information should therefore not be interpreted as evidence of treatment efficacy. The absent numerical results and incomplete analysis details sharply limit conclusions about the statistical performance of the intended conversion.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Florbetapir Calibration to the Centiloid Scale — NCT02120664.
• Registry record last updated 2017-03-09; status last verified 2017-01.
• Results first posted 2017-03-09.
• 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.

19. Record Summary

NCT02120664 is a completed Phase 1 study sponsored by Avid Radiopharmaceuticals, with enrollment of 35 and study dates listed as 2014-04 to 2015-09. The title and official title describe calibration of florbetapir (18F) PET imaging data to the Centiloid scale based on 11C-PiB. The study is classified as non-randomized, single-masked, and parallel.

The primary outcome is a mean continuous measure in Centiloid units, defined through conversion of florbetapir SUVr using a cortical average-to-cerebellum ratio, with a timeframe of up to 70 minutes post injection. Secondary outcomes concern correlation between florbetapir and PiB Centiloid measurements and variability of PET images in young healthy controls.

The registry describes the scale anchors, approximate study groups, and two incomplete scan situations. It does not report numerical results for the primary or secondary outcomes, the final analysis population, or the statistical handling of missing measurements. The record therefore documents the calibration objective and design but does not permit evaluation of the observed conversion or its precision.