Ultrasound-Based Liver Fat Quantification (LFQ) Pilot Study: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Ultrasound-Based Liver Fat Quantification (LFQ) Pilot Study (NCT04657523), a na na study. The primary outcome is Correlation of the Liver Fat Percentage Reported From MRI-PDFF With Measurements Hepatorenal Index (HRI) of Ultrasound Biomarkers for Liver Fat..

2. Clinical Question

3. Trial Design

This was a completed, single-group pilot study of an investigational ultrasound-based liver-fat quantification feature. The registry lists allocation as not applicable and masking as none. There is no randomized treatment assignment or parallel control arm described, so the design is not intended to estimate a treatment effect by comparing groups. Instead, it evaluates whether measurements from an ultrasound feature correspond with MRI-PDFF liver-fat measurements.

The registry describes an intended series of abdominal imaging procedures using ultrasound and MRI. The plan states that subjects would undergo two investigational ultrasound examinations with the Philips EPIQ Ultrasound System and one MRI-PDFF examination according to clinical standard of care. The outcome description reports that each enrolled subject underwent up to two ultrasound scans conducted by two different operators, with 119 non-missing scans reviewed. The record does not explain whether every planned MRI examination was completed or how MRI records were matched to ultrasound measurements.

The listed study dates are 2020-10-30 for start and 2023-01-18 for primary completion; status is completed. The sponsor is Philips Clinical & Medical Affairs Global. The registry lists enrollment as 60, but does not give further details about recruitment, participant flow, exclusions, or the number included in each outcome analysis. It also does not report a statistical analysis plan, sample-size justification, or prespecified criteria for interpreting correlation. These omissions limit assessment of how the pilot’s design and analysis were set before data collection.

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

Analysis population for the primary outcome, as defined in the registry: Each enrolled subject underwent up to two ultrasound scans conducted by two different operators, resulting in a total of 119 non missing scans reviewed for this study. The primary outcome measure, the Hepatorenal Index, was calculated based on these scans. The hepatorenal index measurement from each ultrasound scan was reviewed.

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: 60 participants used for Baseline Characteristics

6. Primary Endpoint

The registry lists the primary outcome as Correlation of the Liver Fat Percentage Reported From MRI-PDFF With Measurements Hepatorenal Index (HRI) of Ultrasound Biomarkers for Liver Fat., measured over Intra-procedural (1 day).

The primary outcome concerns the association between the ultrasound Hepatorenal Index (HRI) and liver-fat percentage measured by MRI-PDFF. HRI is a continuous ultrasound-derived biomarker. The registry states that higher HRI indicates greater liver-fat accumulation, while lower HRI suggests less hepatic fat. MRI-PDFF is described as the clinical reference measurement, with its result obtained from the subject’s medical record.

Correlation is used to describe how measurements vary together across subjects. It does not, by itself, establish that the measurements agree in absolute value or can be substituted for each other. Two measures can be correlated while having systematic differences, so evidence about agreement would require appropriate analyses in addition to an association estimate.

The outcome timeframe is intra-procedural (1 day). Multiple quantitative measurements were acquired for each biomarker per subject on the same day by two operators. The outcome description says the difference between operators was calculated at the subject level by averaging multiple measurements. However, the registry record does not report the resulting HRI correlation, the operator-difference summary, or uncertainty around either result. The stated purpose is clear, but the posted record does not permit evaluation of the observed relationship or its precision.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

What does correlation measure?

Correlation is a summary of the direction and strength of a relationship between two measurements. In this study, the central pair is ultrasound HRI and liver-fat percentage measured by MRI-PDFF. A positive association would mean that larger values of one measure tend to accompany larger values of the other; a negative association would mean they tend to move in opposite directions. The registry record does not report the estimated correlation, so its direction and magnitude cannot be described from the posted results.

General statistical explanation: A correlation coefficient compresses the pattern of paired observations into one summary. That summary can be influenced by the range of values represented in the study and by unusual observations. It does not show whether one measurement can replace another, whether either measure is accurate against an external truth, or whether a change in one causes a change in the other. The record’s stated purpose is to assess correlation, not to establish causation.

Why is agreement a separate issue?

Correlation concerns co-variation; agreement concerns the closeness of paired measurements. If two measures rise and fall together but one tends to be consistently higher than the other, their correlation can still be substantial even though they do not give interchangeable values. A measurement-validation assessment may therefore examine bias and the spread of differences, alongside association. The registry record does not report an agreement analysis or results that would allow such an assessment.

General statistical explanation: Agreement methods use paired differences and consider both systematic offset and variation in those differences. Their interpretation depends on the scale and intended use of the measurement. No such details or agreement estimates are reported for this study in the registry record, so no conclusion about numerical interchangeability is supported.

What do the repeated measurements add?

The primary outcome description says multiple quantitative measurements were acquired for each biomarker per subject on the same day by two operators. It says the operator difference was calculated at the subject level by averaging multiple measurements. The outcome timeframe is intra-procedural (1 day). This design allows the study to examine whether measurements vary between operators, in addition to examining association with MRI-PDFF.

General statistical explanation: Repeated observations from the same person are related rather than independent. Treating every scan or measurement as if it came from a different person can make uncertainty appear smaller than it is, because within-person information is counted repeatedly. Analyses can address this clustering by summarizing at the participant level or by using a model that accounts for repeated observations. The registry describes a subject-level averaging step for the operator difference, but does not report the full analysis method for the primary correlation or the other outcomes.

How should scan counts be understood?

The outcome description reports 119 non-missing scans reviewed for the study and says each enrolled subject underwent up to two ultrasound scans by two different operators. These statements describe scan availability, not necessarily the number of complete pairs used in each analysis. A primary correlation requires paired ultrasound and MRI-PDFF information for the relevant subjects. The record does not report how many complete pairs were analyzed, whether some scans lacked matching MRI values, or whether participant-level averages or individual scan values entered the correlation analysis.

General statistical explanation: Missing observations can affect precision and can also affect the estimated relationship if missingness is associated with measurement values or participant characteristics. Reporting the analysis population and the reasons for missing data helps readers assess this risk. The record does not provide enough detail to determine the missing-data mechanism or to assess any adjustment for it.

What does a confidence interval contribute?

General statistical explanation: A confidence interval communicates uncertainty around an estimate. It helps distinguish a precise estimate from one compatible with a broad range of possible values. The width of an interval depends on the amount and variability of usable information as well as the analysis method. The registry record does not report a confidence interval for the primary correlation, so the precision of the estimate cannot be assessed.

A p-value, if reported, would address compatibility of the data with a specified null hypothesis under assumptions of the test. It would not measure the strength of association, prove equivalence, or describe measurement agreement. No primary p-value is reported in the registry record. The study’s clinical feasibility objective should therefore not be interpreted as a statistical conclusion based solely on the existence of a completed study record.

What is the role of the sample-size illustration?

The record describes a planning illustration based on the observed effect, using a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation; the effect to detect is a stated fraction of the observed effect. Such illustrations estimate the participants per arm that would be needed under those assumptions in a hypothetical two-arm design. They are not observed power and do not establish the design rationale for this single-group measurement study. The registry record does not report that this illustration was the study’s own sample-size plan or provide a sample-size justification for the pilot.

General statistical explanation: A planning calculation depends on the effect size and statistical assumptions chosen. When based on an effect observed in the same study, it is a retrospective planning exercise, not evidence that the original design had a particular probability of detecting an effect. The stated illustration also assumes a two-arm allocation that does not describe the registry’s single-group design. Its assumptions should not be mistaken for the analysis plan used for the primary endpoint.

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("Investigational Ultrasound Imaging for Liver Fat Q")
n   <- c(60)
m   <- c(1.6)
sp  <- c(0.7)
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  Investigational Ultrasound Imaging for Liver Fat Q  60  1.600  0.700  1.419  1.781

11. Multiplicity, Missing Data and Analysis Populations

The registry lists 9 outcome measures and 15 analyses. It identifies the primary objective as examining correlation between MRI-PDFF liver-fat percentage and ultrasound HRI, and lists secondary objectives involving same-day inter-operator variability and data-acquisition failure for HRI, acoustic attenuation, and tissue stiffness. The record does not state how statistical testing across these outcomes or analyses was handled. A multiplicity adjustment, testing hierarchy, or distinction between confirmatory and exploratory analyses is not reported in the registry record.

This matters because examining several outcomes or analyses creates multiple opportunities to observe apparently notable results. General statistical explanation: If many hypotheses are tested, the chance of at least one small p-value can increase unless the analysis plan accounts for the number and structure of tests. Without the stated testing strategy and results, it is not possible to assess how multiplicity should affect interpretation of any findings.

The outcome description reports 119 non-missing ultrasound scans, but the registry does not report the complete-case count for the primary correlation or the amount of missing MRI-PDFF and biomarker data. It does not state whether missing measurements were excluded, imputed, or otherwise handled. The population used in each analysis, reasons for missingness, and any sensitivity analyses are not reported in the registry record. As a result, the effect of incomplete data on the estimates cannot be evaluated.

The record gives enrollment as 60 and describes the scan-level information, but does not provide a participant flow account or analysis-population definitions. It is therefore unclear from the registry record how many participants contributed to each endpoint, whether exclusions occurred, or whether all enrolled participants were included in every analysis. No statistical analysis plan or interim-analysis approach is reported.

12. 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 registry lists one group, Investigational Ultrasound Imaging for Liver Fat Quantification. For this group, 60 participants were reported at risk for serious adverse events and 0 were reported as affected; deaths are recorded as 0. The registry also reports 60 at risk for other adverse events and 0 affected. No serious adverse-event terms are listed.

These figures describe what was recorded in the registry for the reported categories. A count of 0 means that the record reports no affected participants in that category; it does not prove that no symptoms or unwanted events of any kind occurred. The registry record does not provide event descriptions, severity details beyond the serious-event category, timing, relatedness assessments, or a fuller account of adverse-event collection.

The study is single-group, so these safety counts do not provide a randomized comparison against another group. They should not be read as a comparative estimate of risk or as evidence that the investigational procedure is free from adverse events. The record’s adverse-event summary is limited to the categories and counts stated above.

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

The study’s central statistical question is whether ultrasound HRI varies in association with MRI-PDFF liver-fat percentage. The registry identifies this objective and describes how measurements were collected, including scans by different operators. However, the primary correlation result and its uncertainty are not reported in the registry record. The posted information therefore does not show how strong or precise the observed association was.

Even if an association estimate were available, correlation would not by itself establish that ultrasound and MRI-PDFF measurements agree closely enough to be interchangeable. Nor would it establish that ultrasound measurement predicts future clinical outcomes, changes care, or improves health. Those are separate questions requiring evidence beyond the correlation objective described in this record.

The study is labeled a pilot and is described as evaluating clinical feasibility of an investigational technique. Its design is single-group, without randomization or masking. That structure fits a measurement-focused assessment, but it does not support a treatment-versus-control interpretation. The registry record also does not report the detailed analytic methods, complete matched-data count, or numerical findings for inter-operator variability and acquisition failure.

The reported 119 non-missing scans provide information about ultrasound data reviewed, but do not establish how many complete ultrasound–MRI pairs were analyzed. Likewise, the absence of reported serious adverse events is limited to the registry’s recorded safety categories and does not establish the absence of every possible unwanted event. The most defensible reading is that the registry describes a planned and completed pilot with relevant measurement objectives, while leaving the principal quantitative evidence unavailable for independent interpretation.

14. Why This Trial Matters Statistically

This study addresses a measurement question that is important in biostatistics: whether a new measurement method relates to an established imaging measure. The distinction between association and agreement is central. A biomarker can track variation in a reference measure without reproducing its values closely, so the statistical target must match the intended use of the measurement. The registry states a correlation objective, but does not report enough numerical results to determine what the study found.

The study also highlights the structure of data from repeated scans and multiple operators. Measurements collected from the same participant are clustered, and operator variation is a separate component of measurement performance. The registry describes subject-level averaging for the operator difference and lists variability and acquisition-failure objectives for several biomarkers. Those objectives show why a single correlation statistic would not fully characterize measurement quality.

For readers assessing a pilot study, transparency about the analysis population, paired measurements, missing data, uncertainty, and multiplicity is essential. Those details allow the reader to distinguish a descriptive feasibility exercise from a precise validation claim. Here, the registry identifies enrollment, scan information, and outcome topics, but does not report the primary estimate or supporting uncertainty measures.

The statistical value of this record is therefore in clarifying the design and the questions that a measurement study needs to answer. It also illustrates why the presence of a stated primary objective is not a substitute for the associated result. Without numerical estimates and their precision, readers cannot evaluate the strength of the evidence for the ultrasound measure’s relationship to MRI-PDFF.

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 important limitation for interpretation is that the registry record does not report the primary correlation estimate or its uncertainty. Without these results, the central objective cannot be judged quantitatively. The record also does not provide the primary analysis population, number of complete ultrasound and MRI-PDFF pairs, or the approach used to handle missing measurements. Although 119 non-missing scans are reported, scan availability does not establish complete paired data for the correlation.

The outcome description provides only a partial account of how observations were summarized. It says multiple measurements were collected by two operators and that the difference was averaged at the subject level, but does not specify the full statistical model or method for the primary association. The registry record does not report whether the primary correlation used subject-level averages or scan-level values, or how dependence among repeated observations was addressed.

Several outcomes and analyses are listed, but the approach to multiplicity is not reported. The record does not describe confidence intervals, hypothesis tests, prespecified thresholds, sensitivity analyses, or an analysis plan. It also does not give detailed eligibility criteria, participant characteristics, flow through the study, or the number included in each analysis. These omissions constrain judgments about representativeness and robustness.

Finally, correlation is not equivalent to agreement. The registry’s primary question is about association, and no agreement results are reported. The single-group, unmasked design is not a randomized comparison and does not estimate comparative treatment effects. These limitations are specific to interpreting this study’s measurement objective and the information recorded in the registry.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Ultrasound-Based Liver Fat Quantification (LFQ) Pilot Study — NCT04657523.
• Registry record last updated 2026-01-23; status last verified 2024-10.
• Results first posted 2026-01-23.
• 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

NCT04657523 is a completed, single-group pilot study of investigational ultrasound-based Liver Fat Quantification software. The registry describes comparison of ultrasound HRI measurements with MRI-PDFF liver-fat percentage and lists additional objectives involving operator variability and data-acquisition failure. Enrollment was 60; the outcome description reports 119 non-missing ultrasound scans reviewed. The primary correlation estimate, its uncertainty, and the detailed analysis population are not reported in the registry record. Consequently, the study’s stated measurement question can be described, but its numerical answer and precision cannot be assessed from the posted record.