Volatilome and Single-Lead Electrocardiogram Optimize Ischemic Heart Disease Diagnosis Using Machine Learning Models: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Volatilome and Single-Lead Electrocardiogram Optimize Ischemic Heart Disease Diagnosis Using Machine Learning Models (NCT06181799), a registered phase not stated study. The primary outcome is Diagnostic Accuracy (AUC, Sensitivity, Specificity, NPV, PPV) of the Stress-ECG Test in Ischemic Heart Disease.

2. Clinical Question

3. Trial Design

This is a prospective, case-control, single-center, observational, non-randomized study. The registry lists I.M. Sechenov First Moscow State Medical University as the sponsor and gives the study status as completed. The listed start date is 2023-11-01, and the primary completion date is 2024-06-10.

The registry lists an enrollment of 80. Its group descriptions are stated as plans: the experimental group was planned to include 31 people with a myocardial perfusion defect on stress computed tomography myocardial perfusion imaging, and the control group was planned to include 49 people without that defect. These descriptions do not establish how many people were ultimately analyzed in each group. Actual analysis-group sizes are not reported in the registry record.

The diagnostic context includes physical exertion monitored with a 12-lead ECG, together with exhaled-breath volatile organic compound analysis and single-lead ECG parameters. The listed intervention is mass spectrometry using the PTR TOF-1000 trace VOC analyzer. The primary outcome concerns diagnostic accuracy of the stress electrocardiography test in ischemic heart disease, assessed during 6 months according to the registry description.

Because participants were not randomized to the groups, the group labels should be read as classifications based on the stress-imaging finding, not as randomized treatment assignments. The registry record explicitly cautions that the arms are not a like-for-like control comparison. Details about recruitment procedures, eligibility criteria, test thresholds, blinding, and the statistical analysis plan 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.

6. Primary Endpoint

The registry lists the primary outcome as Diagnostic Accuracy (AUC, Sensitivity, Specificity, NPV, PPV) of the Stress-ECG Test in Ischemic Heart Disease, measured over The study was completed on 10.06.2024; the outcome measure was assessed during 6 months for the stress electrocardiography test.

The primary endpoint is described as diagnostic accuracy of the stress-ECG test in ischemic heart disease, with area under the curve, sensitivity, specificity, negative predictive value and positive predictive value named as measures. The stated assessment timeframe is during 6 months for the stress electrocardiography test. The registry also describes the study as evaluating functional testing with physical exertion monitored by 12-lead ECG, in combination with exhaled-breath volatile organic compounds and single-lead ECG parameters.

These measures address different aspects of classification. Sensitivity describes how often a test identifies people who meet the chosen disease reference definition; specificity describes how often it classifies people without that definition as negative. Positive and negative predictive values describe the chance that a positive or negative result, respectively, corresponds to the reference classification in the studied setting. The area under the curve summarizes discrimination across possible test thresholds. These are general descriptions of diagnostic statistics, not estimates from this study.

Diagnostic-accuracy measures are used to characterize how test results correspond to a reference classification, rather than to estimate the effect of assigning a treatment. The registry names the measures but gives no numerical estimates, confidence limits, diagnostic thresholds, or detailed reference-standard procedures. Those results and details are not reported in the registry record, so the endpoint cannot be quantitatively interpreted from the record alone.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The registry identifies a diagnostic-accuracy question. In general, diagnostic studies compare a test result with a reference classification to assess how well the test distinguishes the conditions of interest. Here, the primary outcome names stress electrocardiography and several standard diagnostic measures. The numerical results and important implementation details are not reported in the registry record, so the following explains the measures generally rather than describing study-specific estimates.

What does area under the curve describe?

In general, a receiver operating characteristic curve represents the relationship between sensitivity and the false-positive rate across possible classification thresholds. The area under that curve summarizes discrimination across those thresholds: it describes how well a test tends to rank people with and without the reference condition. It does not identify a single threshold for clinical use, and it does not by itself show whether a test improves decisions or outcomes. Interpretation requires knowing how the curve was constructed and what reference definition and test outputs were used. Those details are not reported in the registry record.

How are sensitivity and specificity interpreted?

In general, sensitivity is the proportion of people meeting the reference definition who test positive; specificity is the proportion of people not meeting it who test negative. They describe classification conditional on the reference status. A change in threshold can trade sensitivity against specificity, so the selected threshold matters. The registry names both measures but does not report their estimates, the threshold used, or the underlying classification counts. Their performance in this study therefore cannot be reconstructed from the registry record.

What do predictive values add?

In general, positive predictive value is the proportion of positive test results that correspond to the reference condition, while negative predictive value is the proportion of negative test results that correspond to its absence. Unlike sensitivity and specificity, predictive values depend on how common the reference condition is in the population being assessed, as well as on the test’s classification performance. A case-control sample may have a distribution of cases and controls that differs from the population in which a test might be used. Predictive values from such a sample therefore need careful interpretation and may not transfer directly to another setting. The registry does not report these values or the relevant underlying counts.

Why do confidence intervals matter?

In general, a confidence interval expresses statistical uncertainty around an estimate under the assumptions of the method used. A narrower interval indicates greater precision under those assumptions; a wider interval indicates more uncertainty. For diagnostic measures, precision depends in part on the number of participants in the relevant reference-status and test-result categories. The primary outcome lists a 95% confidence interval as its dispersion measure, but interval bounds are not reported in the registry record. Without estimates, interval bounds and analysis counts, neither the magnitude nor the precision of the diagnostic findings can be evaluated.

What can the design establish?

Prospective data collection can define a planned sequence for obtaining measurements, but the registry’s description alone does not establish how completely that sequence was followed. In a case-control design, groups are classified according to disease-related status rather than assigned at random to treatments. Such a design can examine diagnostic separation between groups, but it does not estimate the causal effect of an intervention. The registry explicitly notes that the arms are not a like-for-like control comparison. The precise sampling process and the relationship between the study groups and a broader clinical population are not reported in the registry record.

How should the sample-size illustrations be understood?

In general, a planning calculation based on an observed effect is an illustration of the sample size that might be needed under stated assumptions; it is not observed power and does not show that a study had a particular probability of detecting an effect after the fact. The record’s stated illustration assumptions are a two-sided significance level of 0.05, equal allocation between arms, and a normal approximation, with the effect to detect described as a stated fraction of the observed effect. These are assumptions for planning illustrations, not evidence that the observational study used equal allocation or that its original design was based on that calculation. The listed planned groups are not equally sized. Numerical illustration results are not reported in the registry record, and the study’s own sample-size justification is not reported in the registry record.

For event outcomes, the record’s general note distinguishes total-event calculations from calculations expressed as participants per arm. That distinction is a general feature of sample-size methods and does not supply an observed endpoint result for this study. Because the primary diagnostic estimates and analysis details are absent, no recalculated diagnostic effect or precision estimate can be interpreted from the registry record.

10. Multiplicity, Missing Data and Analysis Populations

The registry lists 7 outcome measures and 0 analyses. It does not identify which outcomes were treated as primary beyond the named primary diagnostic-accuracy outcome, nor does it describe a strategy for controlling the chance of false-positive findings across multiple outcomes or test combinations. A multiplicity adjustment or other prespecified approach is not reported in the registry record. Without that information and without numerical analyses, the impact of multiple testing cannot be assessed.

Missing-data handling is also not described. The registry does not report the number of missing or uninterpretable stress-ECG results, breath measurements, single-lead ECG measurements, or reference classifications. It does not state whether participants with incomplete measurements were excluded, how any missing values were handled, or whether sensitivity analyses were performed. These details are not reported in the registry record.

The analysis populations are not specified. In particular, the record does not give the number of people included in each endpoint calculation or explain whether all enrolled participants, only participants with complete diagnostic assessments, or another defined population contributed to the estimates. Planned group descriptions should not be taken as analyzed denominators.

These omissions limit assessment of completeness, selection effects and the stability of any diagnostic estimates. They do not establish that missing data or multiplicity problems occurred; the relevant procedures and counts are simply not reported in the registry record.

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 figures are recorded separately for the experimental and control groups. For the experimental group, the registry lists 31 at risk for serious events and 0 affected; it also lists 31 at risk for other events and 0 affected, with 0 deaths. For the control group, it lists 49 at risk for serious events and 0 affected; it also lists 49 at risk for other events and 0 affected, with 0 deaths. No serious-event terms are listed.

These are the recorded figures, not evidence that no adverse events of any kind occurred in every participant or that all possible safety outcomes were assessed. The registry record does not describe the event-collection procedures, reporting window, completeness of follow-up, or definitions used for these counts. The groups are observational and are not a randomized, like-for-like safety comparison. The figures therefore should not be interpreted as a comparative estimate of risk or as proof of safety.

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

The record describes a completed prospective observational study examining diagnostic information from stress electrocardiography, exhaled-breath volatile organic compounds and single-lead ECG parameters in the context of ischemic heart disease. It identifies the primary diagnostic question and names area under the curve, sensitivity, specificity, negative predictive value and positive predictive value as outcome measures. However, the numerical estimates and their confidence intervals are not reported in the registry record. The posted information therefore does not establish the magnitude or precision of diagnostic performance.

The group descriptions distinguish people with and without a myocardial perfusion defect on stress computed tomography imaging. Since the study is non-randomized and the registry cautions that the groups are not a like-for-like control comparison, the design should not be read as testing the causal effect of a treatment. Nor do the listed group plans establish the number of people included in the final endpoint analyses.

The results that are absent from the registry should not be interpreted as zero values or as evidence of either favorable or unfavorable test performance. The absence of numerical estimates simply means that the record does not permit a quantitative conclusion. The same caution applies to test thresholds, missing data, analysis populations, model-building details and multiplicity: these are not reported in the registry record.

More broadly, diagnostic accuracy is not identical to clinical usefulness. In general, a test’s measured classification performance does not alone show whether its use changes decisions, patient outcomes or resource use. Such questions require relevant evidence beyond the diagnostic measures named in this record. No treatment effect or clinical benefit can be inferred from the registry information.

13. Why This Trial Matters Statistically

This study is statistically relevant because it frames a diagnostic question involving several types of measurement: stress electrocardiography, exhaled-breath volatile compounds and single-lead ECG parameters. Combining information from different measurement sources can raise questions about how each test is evaluated, how any combined model is constructed, and whether its performance is assessed in data independent of model development. The registry record does not describe those procedures, so it cannot establish how such questions were handled.

The design also matters for interpretation. A prospective case-control study can compare test classifications across groups defined by a reference finding, but it is not a randomized comparison of assigned treatments. The registry explicitly warns that the arms are not a like-for-like control comparison. That distinction prevents a diagnostic study from being mistakenly read as a treatment-effect study.

Finally, the difference between naming an endpoint and reporting its results is consequential. The registry specifies recognized accuracy measures and a 95% confidence interval as the dispersion measure, yet the corresponding numerical results are not reported in the registry record. Without estimates, interval bounds, thresholds and analysis denominators, readers cannot judge the size or precision of the reported diagnostic performance. Clear reporting of these items is central to assessing whether a diagnostic finding is robust and applicable beyond the study setting.

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 trial-specific limitations arise from the information recorded. The study is single-center, observational and non-randomized. Its case-control groups are planned according to whether a stress-imaging myocardial perfusion defect is present, and the registry cautions that they are not a like-for-like control comparison. This limits the interpretation of between-group contrasts and does not support a causal treatment comparison.

The registry lists an enrollment of 80 and planned group descriptions of 31 and 49, but it does not report final analysis counts, endpoint denominators or the number of participants with complete measurements. The diagnostic estimates, confidence-interval bounds, thresholds and underlying classification counts are also not reported in the registry record. As a result, neither the performance nor its precision can be appraised from the record.

The record does not provide enough detail to assess how the stress-ECG reference comparison was implemented, whether assessors were blinded, how the breath and ECG measurements were processed, or whether any combined model was evaluated independently of its development data. These details are not reported in the registry record. The number of listed outcomes makes the absence of a reported multiplicity approach relevant, while the lack of missing-data procedures and analysis-population definitions further limits assessment of potential bias.

These are limits of what can be concluded from the registry record; they do not establish that a particular methodological problem occurred. They indicate which information is not reported and therefore cannot be evaluated.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: Volatilome and Single-Lead Electrocardiogram Optimize Ischemic Heart Disease Diagnosis Using Machine Learning Models — NCT06181799.
• Registry record last updated 2025-08-15; status last verified 2023-10.
• Results first posted 2025-08-15.
• 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 record describes NCT06181799, a completed, prospective, case-control, single-center, observational, non-randomized study sponsored by I.M. Sechenov First Moscow State Medical University. It lists an enrollment of 80 and concerns stress electrocardiography, exhaled-breath volatile organic compounds and single-lead ECG parameters in assessing ischemic heart disease.

The primary outcome names diagnostic accuracy measures for the stress-ECG test, including area under the curve, sensitivity, specificity, negative predictive value and positive predictive value. The registry lists 7 outcome measures and 0 analyses. Numerical primary-outcome results, confidence-interval bounds, analysis counts and key analysis procedures are not reported in the registry record. The study is completed, but its diagnostic findings cannot be quantitatively assessed from the posted record.