Detection of Graft Versus Host Disease With [18F]F-AraG: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Detection of Graft Versus Host Disease With [18F]F-AraG (NCT03367962), a non-randomized phase 1 study. The primary outcome is Distribution of Acute GVHD Grades (MAGIC Criteria) at Time of PET/CT Scan.

2. Clinical Question

3. Trial Design

NCT03367962 is listed as a completed phase 1 study with enrollment of 9 participants. The study began on 2018-05-15, and its primary completion date is 2023-06-21. It was sponsored by CellSight Technologies, Inc. The registry record describes it as a single-center imaging study.

The design was non-randomized, unmasked, and parallel. It included three groups: people highly suspected of already having acute GVHD, people at high risk of developing acute GVHD, and healthy subjects. These groups differ in clinical status and in the circumstances surrounding their scans. Consequently, they should not be interpreted as randomized or interchangeable comparison arms.

In the highly suspected group, a biopsy was taken to confirm acute GVHD, followed by [18F]F-AraG PET-CT. In the high-risk group, the planned scans were on day 4 ± 2 after transplant and again between day 14-21 after transplant. Healthy volunteers underwent eligibility evaluation, consent, enrollment, and a PET-CT scan. The registry record also lists biodistribution and kinetic behavior in healthy volunteers as a secondary outcome.

The primary outcome is a count of participants within each acute GVHD grade, based on MAGIC criteria at the time of PET/CT. The registry record lists 3 outcome measures and 0 analyses. It does not report the observed grade distribution or analysis results. The protocol details beyond the stated study description, including any further design rationale, 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

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: Ran out of funding for Healthy Subjects

6. Primary Endpoint

The registry lists the primary outcome as Distribution of Acute GVHD Grades (MAGIC Criteria) at Time of PET/CT Scan, measured over At the time of the PET/CT scan: within 7 days of GVHD suspicion for the Highly Suspected arm, and at the Day 4 ± 2 post-transplant scan for the High-Risk arm.

The primary endpoint is the distribution of acute GVHD grades among participants at the time of PET/CT. Grades I-IV were assigned using Mount Sinai Acute GVHD International Consortium (MAGIC) criteria, drawing on organ involvement documented in the clinical chart and biopsy review. The measure is a participant count within each grade, rather than a continuous imaging measurement or a time-to-event outcome.

Grade counts provide clinical context for interpreting imaging in participants with suspected or elevated risk of GVHD. If imaging observations are compared with clinical grading, the grade distribution helps describe the severity spectrum represented at scanning. It does not, by itself, measure diagnostic accuracy, predictive accuracy, or whether imaging changes clinical outcomes. Those questions require defined analyses that relate scan findings to an appropriate reference outcome and account for the study population and timing.

The registry record gives the endpoint definition and timing but does not report the number of participants in each grade. It therefore does not show the observed grade distribution or permit an assessment of how those counts relate to PET/CT findings. The registry record also does not report a numerical primary endpoint estimate or an uncertainty interval.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

A participant-count endpoint records how many people fall into each defined category. In this study, the categories are acute GVHD Grades I-IV under MAGIC criteria. Counts describe the observed sample; they do not alone quantify how well an imaging test identifies disease. To evaluate that broader question, an analysis would need to connect PET/CT findings with a reference measure such as biopsy or a later clinical outcome, using a clearly specified definition of a positive result. Such analysis results are not reported in the registry record.

What a grade distribution can and cannot show

A grade distribution can describe the severity categories represented at the scan time. It is not inherently a comparison of treatments, because [18F]F-AraG is used as an imaging tracer and the primary endpoint is a clinical grade count. The endpoint description says grades were assigned from charted organ involvement and biopsy review. The registry record does not say how disagreements were resolved, whether graders were masked to imaging, or how incomplete grading information was handled.

Counts are often accompanied by uncertainty intervals when the goal is to estimate a broader population quantity. A confidence interval describes a range of values compatible with the observed data under a specified statistical model and procedure; it does not guarantee that the true value lies in that range. With a small cohort, estimates can be imprecise, and rare outcomes may be difficult to characterize. No confidence intervals or other uncertainty estimates for the primary endpoint are reported in the registry record.

Why the group structure matters

The study is non-randomized and unmasked. Random allocation, when used, can help balance known and unknown prognostic factors between comparison groups. That protection does not apply to this design. The study groups are defined by different clinical states, and the registry describes different scan schedules for them. A contrast across these groups would therefore be vulnerable to differences in participant characteristics, disease status, timing, and follow-up. The registry record explicitly cautions that the arms are not a like-for-like control comparison.

Unmasked imaging may also affect assessment or interpretation if readers know clinical details or other results. Whether image interpretation was independent, whether readers were blinded to biopsy, and whether assessment procedures were standardized are not reported in the registry record. These are relevant considerations for a diagnostic imaging study, but their absence from the record does not establish that the procedures were or were not used.

Analysis plans and multiplicity

The registry record lists 3 outcome measures but 0 analyses. It does not report a statistical test, model, adjustment strategy, or prespecified analysis plan. It also does not report how multiple outcomes, grades, scan times, or group comparisons were handled. When many outcomes or comparisons are examined, the chance of a seemingly notable result can increase. Methods to address multiplicity may be considered in general, but no such method is documented for this study.

The registry record includes planning assumptions for an illustrative sample-size calculation: a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation, with the effect to detect expressed as a fraction of an observed effect. These assumptions describe a planning illustration, not the trial’s own design or an observed-power analysis. The record does not report the sample-size justification used for the trial. A planning calculation cannot replace reported effect estimates and uncertainty, and the registry record does not report the observed effect needed to evaluate such an illustration.

Safety counts and denominators

For adverse events, a denominator identifies how many participants were considered at risk in a group. The posted summary gives the at-risk counts separately for the highly suspected and high-risk groups, along with counts of affected participants and deaths. Those figures should be read as group-specific descriptive counts, not as a comparative risk analysis. The healthy-subject group has no participants at risk in this adverse-event summary; its zero counts therefore cannot be read as evidence of no risk among healthy volunteers.

In general, over-dispersion matters when count data vary more than a statistical model assumes; standard errors from an inadequate model can then be too small. This study’s primary measure is a distribution of participant counts across grades, and no fitted model or dispersion assessment is reported. It would not be appropriate to infer that a count model, over-dispersion correction, or any other specific statistical method was used.

10. Multiplicity, Missing Data and Analysis Populations

The registry record lists 3 outcome measures and 0 analyses. It describes a primary outcome concerning the distribution of acute GVHD grades and a secondary outcome concerning biodistribution and kinetic behavior in healthy volunteers. It does not report how the outcomes were prioritized in statistical testing beyond identifying the primary outcome, whether multiple testing was considered, or whether any adjustment for multiple comparisons was planned or applied. The number and structure of analyses are not reported in the registry record.

Missing-data handling is also not reported in the registry record. The record does not say whether participants with unavailable scans, incomplete clinical grading, or missing biopsy information were included in any analysis, nor does it describe imputation, exclusions, or sensitivity analyses. Because the grade counts and scan results are not reported, the extent of any missing outcome information cannot be assessed from the posted results.

Analysis populations are not described in detail. The study description identifies three groups and their planned procedures, but it does not state which participants were included in each outcome analysis or how deviations from the planned scan schedules were handled. The record does not report an intention-to-diagnose population, a per-protocol population, or another formally defined analysis set. These omissions limit assessment of whether outcome summaries used all enrolled participants or a subset.

These details matter because outcome selection, missingness, and analysis population can affect interpretation, particularly in a small, non-randomized imaging study. The absence of these details in the registry record should not be taken as evidence that the investigators did not consider them; it means the posted record does not document their handling.

11. 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 0% reporting threshold.

The adverse-event summary separates serious events, other events, deaths, and the number of participants at risk by group. In the highly suspected group, 7 participants were at risk, with 0 affected by serious adverse events, 0 affected by other adverse events, and 0 deaths. In the high-risk group, 2 participants were at risk, with 0 serious adverse events, 0 other adverse events, and 0 deaths. For healthy subjects, the summary gives 0 participants at risk and zero event counts.

The healthy-subject counts cannot be interpreted as a finding of no events among an observed at-risk group, because the summary lists no participants at risk for that group. The record does not explain this denominator. The counts also do not provide a comparison with an untreated or randomized control group. They are descriptive registry figures rather than evidence that an intervention is free of risk. The registry record does not report further details about event definitions, observation periods, or event attribution.

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

The registry record establishes that this completed phase 1 study investigated [18F]F-AraG PET-CT in a single-center setting and defined acute GVHD grade distribution at scanning as its primary outcome. It also describes intended comparisons with biopsy and follow-up in participants at high risk. Those aims are relevant to evaluating an imaging tracer, but the aims alone do not establish that the tracer identifies disease sites or predicts later GVHD.

The posted results do not include the primary grade counts, PET/CT findings, biopsy comparisons, or analyses. Therefore, the record does not support a conclusion about diagnostic accuracy, prediction, correlation with T-cell accumulation, or clinical utility. It also does not permit an assessment of precision for those outcomes because the relevant estimates and uncertainty measures are not reported.

The absence of a randomized comparator means the study is not designed as a like-for-like comparison of treatment effects. Its groups represent different clinical populations and scan schedules. Even if group-level differences were available, they would require interpretation in light of those design features and would not by themselves show that imaging caused a clinical difference.

The adverse-event figures report no affected participants in the two groups with nonzero at-risk counts, and no deaths in the listed groups. These are limited descriptive observations in the recorded sample, not proof that adverse events cannot occur. The healthy-subject safety denominator is recorded as zero, and the record does not explain the reason.

In short, the registry record documents an exploratory study framework and some event counts, but does not report the outcome evidence needed to establish what the imaging results showed. Any conclusion beyond the recorded design and descriptive figures would go beyond the posted results.

13. Why This Trial Matters Statistically

This study raises statistical questions that are central to early diagnostic imaging research. The target is not simply whether a tracer can be administered, but whether its imaging signal relates to clinically meaningful evidence of GVHD and whether it may help identify disease sites or signal later disease in people at high risk. Those questions call for clearly defined reference standards, timing, outcome definitions, and analysis populations.

The registry record defines a clinical grading outcome at the time of PET/CT and describes biopsy review and follow-up as part of the study rationale. It also distinguishes participants with strong suspicion of disease, participants at risk, and healthy volunteers. Keeping those groups distinct is important: different clinical states and scan schedules can affect observed results, and a non-randomized design does not make the groups exchangeable.

Small early studies can be useful for characterizing feasibility and generating evidence for later research, but estimates from limited samples can be uncertain. A participant-count outcome can describe the observed distribution, while claims about accuracy or prediction require analyses that connect the scans with appropriate reference outcomes. Reporting estimates with uncertainty, and explaining missing data and analysis choices, is essential for judging how much the findings support those claims.

Here, the registry record specifies the question and design but does not report the grade distribution or analyses. That makes the distinction between a study’s scientific aim and its available evidence especially important. The record can be read as documenting an exploratory approach, not as a statistical demonstration of imaging performance.

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 outcome information available in the registry record is incomplete for evaluating the primary question. Although the grade-count endpoint is defined, grade-specific participant counts are not reported. PET/CT findings, biopsy comparisons, and high-risk follow-up results are also not reported in the registry record. The absence of these results prevents direct assessment of the study’s stated imaging aims.

The enrollment was 9, and the groups were non-randomized and clinically different. The highly suspected and high-risk groups also had different scan timing. These features limit direct comparisons and make it difficult to separate effects associated with disease status, timing, or other participant characteristics. The healthy-subject group has 0 participants at risk in the adverse-event summary, which prevents interpreting that group’s event counts as observed safety rates.

The registry record does not report the image-reading procedure, masking of image interpreters, prespecified thresholds for a positive scan, or details of how biopsy and imaging findings were compared. It does not report the statistical analysis plan, confidence intervals, missing-data approach, or analysis populations. These items are relevant to evaluating diagnostic performance and the robustness of any estimates, but their absence from the registry record does not establish what procedures were used outside the record.

The record describes a single-center study and does not report information about recruitment representativeness or external validation. Generalizability beyond the enrolled participants therefore cannot be assessed from the registry record. These limitations concern the amount and detail of evidence documented, not proof that the study procedures were inadequate.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: Detection of Graft Versus Host Disease With [18F]F-AraG — NCT03367962.
• Registry record last updated 2025-12-08; status last verified 2025-11.
• Results first posted 2025-12-08.
• 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

NCT03367962 is a completed, single-center, non-randomized, unmasked phase 1 imaging study sponsored by CellSight Technologies, Inc., with enrollment of 9 participants. It assessed [18F]F-AraG PET-CT in participants highly suspected of having acute GVHD, participants at high risk of developing acute GVHD, and healthy volunteers.

The primary outcome is the distribution of MAGIC acute GVHD grades at the time of PET/CT. The registry record lists 3 outcome measures and 0 analyses, but does not report grade counts, imaging results, biopsy comparisons, or predictive analyses. The adverse-event summary records 0 serious events, 0 other events, and 0 deaths in the two groups with participants at risk; the healthy-subject group has 0 participants at risk in that summary. The posted record therefore describes the study design and selected safety counts but does not provide results sufficient to assess imaging performance.