FlowMet-R Blood Flow Measurement for the Diagnosis of Peripheral Artery Disease and Critical Limb Ischemia: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for FlowMet-R Blood Flow Measurement for the Diagnosis of Peripheral Artery Disease and Critical Limb Ischemia (NCT04120610), a registered phase not stated study. The primary outcome is Sensitivity and Specificity of FlowMet-R for Diagnosis of PAD and CLI.

2. Clinical Question

3. Trial Design

This was a non-randomized, multi-center, longitudinal study of healthy participants and participants with PAD. The brief study description says participants were scheduled for ABI, TBI, and either Duplex Ultrasound or angiographic assessments in a vascular clinic. The registry lists FlowMet-R, ABI, and TBI as diagnostic-test interventions. It describes separate healthy and PAD cohorts, rather than randomized treatment groups, and states that the arms are not a like-for-like control comparison.

The healthy cohort was described as an age-matched population over 40 years old without a history of PAD or suspected PAD. This cohort was to receive FlowMet-R, ABI, and TBI measurements. The PAD cohort comprised all-comers to the vascular lab scheduled for PAD assessment or with a planned endovascular or surgical intervention to address PAD. That cohort was to receive FlowMet-R in addition to routine standard of care. These descriptions do not establish that the groups were comparable for evaluating a treatment effect.

The primary analysis description says healthy and PAD participants were combined into a cohort. ABI and TBI measurements were then used to assign PAD (1), PAD (2), and CLI categories. FlowMet-R data were used to create a predictive model of PAD severity, and ROC curves were generated for PAD and independently for CLI. The listed assessment times were initial, three-month, and six-month time points.

The registry lists enrollment as 195, study start as 2019-12-09, and primary completion as 2022-11-21. The status is TERMINATED. The reason for termination, recruitment details, allocation procedures, and any changes to the planned study are not reported in the registry record. The record also does not report a randomized control group or a treatment comparison.

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.

Before assignment: 205 subjects were consented in the study. Of the 205 subjects, 10 subjects were screen fails due to I/E criteria, resulting in 195 total subjects enrolled. There were limitations discovered in the study design during enrollment due to classification of PAD and CLI from respective sites. Patients were categorized based on observed ABI, TBI and Rutherford Classification regardless of enrollment cohort.

Overall Study

Analysis population for the primary outcome, as defined in the registry: Sensitivity, specificity and area under the curve (AUC) were calculated for PAD (1) vs. not PAD (1), PAD (2) vs not PAD (2), and CLI vs. not CLI at initial, three-month and six-month timepoints.

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: Site-reported demographic and medical history data are presented based on the assigned cohort using the definitions for Healthy and PAD (2) in the Groups and Interventions Section. Baseline information was not collected for one participant; therefore, a total of 194 participants were included in baseline analysis.

6. Primary Endpoint

The registry lists the primary outcome as Sensitivity and Specificity of FlowMet-R for Diagnosis of PAD and CLI, measured over Initial, three month, and six month time points..

The primary endpoint concerns diagnostic performance: sensitivity and specificity of FlowMet-R for diagnosing PAD and CLI. The registry also describes calculating AUC from the ROC curves. Sensitivity describes the proportion of people classified as positive by the test among those who meet the reference classification; specificity describes the proportion classified as negative among those who do not meet it. These measures address different forms of diagnostic error, so neither alone captures all aspects of performance.

An ROC curve represents the trade-off between sensitivity and specificity across possible decision thresholds. The registry says the reported peak sensitivity and peak specificity were selected at the point on the curve closest to the upper-left corner of the unit square. This is a rule for choosing a point on the ROC curve, not evidence that a clinically preferred threshold was established. The threshold values themselves are not reported in the registry record.

AUC summarizes discrimination across the ROC curve rather than at a single threshold. The study description specifies PAD and CLI ROC analyses at the initial, three-month, and six-month time points, with ABI and TBI used to define diagnostic categories. Numerical sensitivity, specificity, AUC, and confidence interval results are not reported in the registry record. Consequently, the endpoint definition explains what was to be evaluated, but the registry information does not establish the magnitude or precision of diagnostic performance.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The primary analysis is framed as a diagnostic-accuracy analysis. The registry says FlowMet-R blood-flow measurements and/or analysis of blood-flow waveforms were used to create a predictive model of PAD severity. The model was then used to generate ROC curves for PAD and, independently, for CLI. The registry describes analysis at initial, three-month, and six-month time points. The model-building procedure, predictor selection, handling of repeated measurements, and validation approach are not reported in the registry record.

What does an ROC curve describe?

As a general statistical explanation, an ROC curve shows how sensitivity and specificity change as a test's decision threshold changes. A more permissive threshold may classify more observations as positive, affecting both true-positive and false-positive classifications. The curve therefore describes a range of possible operating points rather than a single result. The registry says peak sensitivity and peak specificity were selected at the point closest to the upper-left corner of the unit square. This selection rule identifies a point on the curve, but it does not tell readers the corresponding FlowMet-R threshold or whether that threshold was chosen before examining the data.

Sensitivity and specificity answer different questions. Sensitivity concerns positive classifications among those in the reference-positive category; specificity concerns negative classifications among those in the reference-negative category. Neither measure by itself describes the proportion of all classifications that are correct. Their interpretation depends on the reference categories and on which observations were available for analysis. The registry states that ABI and TBI were used to assign PAD (1), PAD (2), and CLI categories, but the category thresholds and detailed classification rules are not reported in the registry record.

What does AUC add?

As a general explanation, AUC summarizes the ROC curve over possible thresholds. It is a measure of discrimination across that range, rather than performance at the single point selected for peak sensitivity and specificity. AUC does not identify a particular operational threshold, and it does not by itself describe calibration, clinical consequences of errors, or how the test would perform in a different population. The registry says AUC was calculated, but numerical AUC estimates and their confidence intervals are not reported in the registry record.

How should confidence intervals be read?

The primary outcome specifies a 95% confidence interval. In general, a confidence interval conveys sampling uncertainty around an estimate under the statistical method used. A narrower interval indicates greater precision under that method; a wider interval indicates more uncertainty. An interval does not remove uncertainty about the study design, reference classification, missing observations, or whether the study population is relevant to another setting. The actual intervals for sensitivity, specificity, and AUC are not reported in the registry record, so their precision cannot be assessed from the available information.

Diagnostic estimates can also depend on the range of participants and disease severity represented in a study. As a general point, a study that includes both healthy participants and people attending a vascular lab may have a different case mix from another clinical setting. The record describes those two cohorts and their eligibility, but does not report the distribution of severity within the analyzed population or provide enough numerical diagnostic results to assess the effect of case mix.

What limits interpretation of the model?

The registry describes developing a predictive model from FlowMet-R data, but does not report whether performance was evaluated in data separate from the data used to create the model. As a general statistical concern, evaluating a model on the same observations used to fit it can make apparent performance optimistic. This is a methodological consideration, not a claim that the study used a particular validation method. Without details of model development and validation, readers cannot assess that possibility. The record also does not describe the model's predictors, parameter estimation, threshold selection beyond the stated ROC-curve rule, or handling of repeated observations.

There are several categories and time points in the stated primary population: PAD (1) versus not PAD (1), PAD (2) versus not PAD (2), and CLI versus not CLI, with assessment at initial, three-month, and six-month time points. These comparisons are described in the registry, but the approach to accounting for dependence among measurements from the same participant is not reported. The statistical method for calculating confidence intervals, and whether estimates were adjusted for any covariates, are also not reported in the registry record.

The registry lists 7 outcome measures and 10 analyses. Those counts alone do not establish how many statistical tests were conducted for a particular question or whether a multiplicity adjustment was used. The adjustment strategy, if any, is not reported in the registry record. As a general matter, examining multiple outcomes, categories, or time points can increase the chance of apparently notable findings when results are selected or interpreted without accounting for the number of analyses. The numerical primary results and analysis plan are unavailable here, so the practical impact cannot be evaluated.

The planning assumptions described for sample-size illustrations are not the same as a prospectively specified trial design. The record says the illustrations use an observed effect, a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation, with the effect to detect set as a stated fraction of the observed effect. It also distinguishes total events for time-to-event calculations from participants per arm for other calculations. These are planning illustrations, not observed power and not evidence that the trial was designed or analyzed according to those assumptions. The registry does not report the original sample-size justification or the statistical plan.

10. Multiplicity, Missing Data and Analysis Populations

The registry describes multiple diagnostic classifications and assessment times: PAD (1) versus not PAD (1), PAD (2) versus not PAD (2), and CLI versus not CLI, at the initial, three-month, and six-month time points. It also lists sensitivity, specificity, and AUC as measures derived from ROC curves. The record lists 7 outcome measures and 10 analyses, but those counts do not show which analyses were designated primary, how they were prioritized, or whether any adjustment was made for multiple comparisons. A multiplicity strategy is not reported in the registry record.

Missing-data handling is also not reported. The registry does not describe how incomplete FlowMet-R measurements, unavailable ABI or TBI classifications, or missed follow-up assessments were handled. It does not report imputation methods, exclusions, or the number of observations contributing to each time point and outcome. The stated primary population combines the healthy and PAD groups before classification, but the record does not give further analysis-population rules.

The registry does not report whether the analysis followed an intention-to-diagnose principle, whether measurements had to meet quality criteria, or how unusable or indeterminate test results were treated. These choices matter because diagnostic estimates depend on which observations enter each calculation and how unresolved classifications are handled. The record also does not specify an analysis plan for dependence among repeated measurements. Without these details, the degree to which missingness or analytic choices may affect the estimates cannot be assessed.

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.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: amputation (2.1%), femoral artery bypass (0.7%), femorofemoral bypass (0.7%), non-healing left leg ulcer (0.7%), cerebrovascular accident (0.7%).

The registry reports serious adverse events and other adverse events separately for the healthy and PAD groups. In the healthy group, 52 were listed as at risk for serious adverse events and 0 as affected; 52 were listed as at risk for other adverse events and 0 as affected. In the PAD group, 143 were listed as at risk for serious adverse events and 6 as affected; 143 were listed as at risk for other adverse events and 2 as affected. The registry reports 0 deaths in the healthy group and 4 in the PAD group.

The listed serious-event terms include Amputation at 2.1%, and Femoral Artery Bypass, Femorofemoral Bypass, Non-Healing Left Leg Ulcer, and Cerebrovascular Accident at 0.7% each. These are registry-reported event terms and percentages; they should not be read as evidence that FlowMet-R caused the events. The record does not report event attribution, timing, event-level overlap, or further details about how the percentages were calculated.

The groups were not randomized treatment arms and are not described as a like-for-like control comparison. Consequently, their adverse-event counts do not establish a comparative safety effect. The registry status is TERMINATED, but the reason for termination is not reported. No inference about the reason for termination or the relationship of events to study procedures can be made from these figures alone.

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

The record establishes the intended diagnostic question and the broad analysis approach: FlowMet-R measurements were used to create a predictive model, with ROC curves for PAD and CLI and sensitivity, specificity, and AUC evaluated at initial, three-month, and six-month time points. ABI and TBI measurements were used to assign the diagnostic categories. This describes what the study aimed to assess, not the size or reliability of any diagnostic effect.

Numerical primary results are not reported in the registry record. Sensitivity, specificity, AUC, confidence intervals, and the selected thresholds are therefore unavailable for interpretation. Without them, readers cannot determine how well the test separated the registry's diagnostic categories, how precisely the estimates were measured, or whether performance differed across the listed time points. The record also does not report the model-validation details needed to judge how well its estimates might generalize.

The study was non-randomized and included a healthy cohort and a PAD cohort with different eligibility descriptions. The registry states that the arms are not a like-for-like control comparison. Group differences should not be treated as causal effects of FlowMet-R, and the diagnostic analysis should not be confused with a randomized treatment comparison. The registry lists enrollment as 195 and status as TERMINATED, but does not report the reason for termination or the number contributing data to each analysis.

Serious and other adverse events are reported for the two groups, along with deaths and selected event terms. These are descriptive safety counts. The record does not report whether events were judged related to FlowMet-R or study procedures, and the group structure does not support a randomized safety comparison.

Overall, the registry record describes a diagnostic study design and endpoint definitions, but does not report the numerical primary findings. The record therefore cannot support a conclusion about the magnitude, precision, or clinical usefulness of FlowMet-R diagnostic performance. Planning illustrations based on observed effects should be read as hypothetical sample-size calculations, not as observed power or evidence of diagnostic performance.

13. Why This Trial Matters Statistically

This study raises a statistical question that differs from a treatment-effect question. Its primary aim was to assess how FlowMet-R measurements classify PAD and CLI categories, using sensitivity, specificity, and AUC. Those measures focus on discrimination against reference classifications, rather than on whether a treatment changes an outcome. Keeping those questions separate prevents a diagnostic study from being misread as a randomized comparison of clinical benefit.

The design also makes the reference classification central to interpretation. The registry says ABI and TBI were used to assign PAD (1), PAD (2), and CLI categories, while the brief summary mentions scheduled Duplex Ultrasound or angiographic assessments. Since detailed reference definitions and the role of those assessments are not reported in the registry record, readers cannot fully reconstruct what counted as a positive or negative classification. Diagnostic accuracy is meaningful only in relation to the classification standard used.

The description of a predictive model makes model validation another important issue. As a general statistical principle, estimates assessed on the same observations used to fit a model may be optimistic. The registry record does not report the validation approach, so it is not possible to assess this concern for the study. Repeated assessment times and multiple diagnostic categories also raise questions about dependence among measurements and multiplicity; the handling of these issues is not reported.

Finally, the record is informative about intended methods but does not give numerical diagnostic results. This distinction matters: endpoint definitions and a planned analysis do not establish the observed performance or its precision. Reporting estimates with confidence intervals, thresholds, reference-standard details, missing-data information, and model-validation methods would be necessary to evaluate the diagnostic evidence in depth.

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.

Several study-specific limitations of interpretation arise from details that are not reported in the registry record. The primary analysis uses ABI and TBI to assign diagnostic categories, but category thresholds and the full reference-classification process are not described. The brief summary mentions Duplex Ultrasound or angiographic assessments, yet their role in the primary endpoint is not specified. This leaves uncertainty about the standard against which FlowMet-R was evaluated.

The study combines healthy and PAD groups for the primary analysis, although their eligibility and measurement descriptions differ. The registry says the arms are not a like-for-like control comparison. It does not report the distribution of participants across disease categories or the number contributing to each analysis. The record also does not provide numerical sensitivity, specificity, AUC, confidence intervals, or the selected operating thresholds.

Model construction and validation are not described in sufficient detail to assess whether performance was evaluated independently of model fitting. The record does not report predictor selection, model specification, repeated-measure handling, or whether threshold selection was prespecified. Nor does it report missing-data procedures, exclusions, quality-control rules, or analysis-population definitions for each time point.

The registry lists the study as TERMINATED but does not report why. It does not give the statistical analysis plan or the original sample-size justification. Consequently, the planning illustrations should not be interpreted as the trial's design rationale or observed power. These omissions do not establish that a particular analysis was inappropriate; they limit what can be concluded from the registry record.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: FlowMet-R Blood Flow Measurement for the Diagnosis of Peripheral Artery Disease and Critical Limb Ischemia — NCT04120610.
• Registry record last updated 2025-04-11; status last verified 2024-03.
• Results first posted 2025-04-11.
• 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 registry describes NCT04120610, titled “FlowMet-R Blood Flow Measurement for the Diagnosis of Peripheral Artery Disease and Critical Limb Ischemia.” Medtronic Endovascular is listed as the sponsor. The study is described as non-randomized, multi-center, and longitudinal, with healthy and PAD cohorts. Enrollment is listed as 195; the study start is 2019-12-09, primary completion is 2022-11-21, and status is TERMINATED.

The primary objective was to assess FlowMet-R for diagnosing PAD and CLI using sensitivity, specificity, and AUC from ROC curves at initial, three-month, and six-month time points. ABI and TBI were used to assign diagnostic categories. Numerical primary results, confidence intervals, and the reason for termination are not reported in the registry record. The record also lists adverse-event counts for the healthy and PAD groups, but does not report event attribution.