This page provides an independent statistical analysis and educational interpretation of publicly reported results. ClinicalTrials.gov provides the official trial registry record. Numerical trial results on this page are restricted to the results posted in the ClinicalTrials.gov record; calculations added by Clinical Biostats are labelled as such.
1. Trial at a Glance
This page covers the registry results for Imagio Feasibility Multi-Reader, Multi-Case Study (NCT03708393), a randomized na study. The primary outcome is The Gain in Imagio IUS Alone Specificity vs Imagio (IUS+OA) Specificity at Fixed 95-99% Sensitivity (fSp), Cohort 1.
| Feature | Registry value |
|---|---|
| ClinicalTrials.gov identifier | NCT03708393 |
| Phase | NA |
| Condition | Breast Cancer |
| Allocation | Randomized |
| Design model | Sequential |
| Masking | Single |
| Enrollment | 155 |
| Arms in the primary outcome | n/a |
| Lead sponsor | Seno Medical Instruments Inc. |
| Results posted | Yes |
2. Clinical Question
Population
Cohort 1 was the intent-to-diagnose population used for the primary analysis and included 120 patients/masses. The registry record also describes Cohort 2, with 30 masses for assessing false negative rate, and Cohort 3, with 5 masses including DCIS and lymphoma cases for assessing how readers scored. Those cohorts have different stated purposes; the primary endpoint is specified for Cohort 1.
Intervention
The combined imaging condition was Imagio Ultrasound plus optoacoustic imaging, identified as Imagio (IUS+OA). In the second read, readers assessed mammography as available together with Imagio Ultrasound and optoacoustic imaging.
Comparator
The comparator was Imagio Ultrasound alone. In the first read, readers assessed mammography as available together with Imagio Ultrasound. The registry record states that both imaging modalities were read for each subject, with the subject serving as their own control.
Primary question
The primary question was whether the specificity associated with Imagio Ultrasound alone differed from that associated with Imagio Ultrasound plus optoacoustic imaging when sensitivity was fixed at 95-99%. Specificity concerns correctly identifying benign masses. The registry record defines this comparison but does not report its numerical estimate.
The endpoint was based on interpretations by 10 independent readers and used biopsy diagnosis or a 12-month follow-up ruling of benign, determined by a truth panel, as the ground truth. The comparison therefore concerns imaging classifications against that reference, rather than a direct comparison of patient outcomes. Its registry timeframe is baseline to 12 month follow-up.
3. Trial Design
The registry record identifies NCT03708393 as the Imagio Feasibility Multi-Reader, Multi-Case Study. It describes a controlled, blinded, multi-reader, multi-case study with randomized allocation, single masking, and a sequential model. The record does not say which party was masked or describe the sequence in further detail; those aspects are not reported in the registry record.
The study evaluated device-based imaging conditions: Imagio (IUS+OA), Imagio Ultrasound, and mammography. The arm descriptions distinguish a first read of mammography as available plus Imagio Ultrasound from a second read of mammography as available plus Imagio Ultrasound with optoacoustic imaging. The endpoint description says both imaging modalities were read for each subject, making each subject their own control. The specific mechanism of randomization and the exact order or implementation of reads are not reported in the registry record.
The record lists 155 enrolled. It gives the study start as 2018-11-13, primary completion as 2018-12-09, and status as completed. The phase is recorded as NA. The stated primary analysis population was Cohort 1, described as the intent-to-diagnose population used to run the analysis to set the sample size for the future pivotal Reader-02 Study. Separate descriptions identify Cohort 2 for assessing false negative rate and Cohort 3 for assessing how readers scored selected masses.
The registry record does not report the allocation ratio, a detailed protocol for the reading sessions, or the rationale for the design features. It also does not report a sample-size justification for this feasibility study. The primary endpoint is defined, but the registry contains no analysis entries. Accordingly, the record supports a description of the planned comparison and study structure, not a numerical account of the primary result.
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: Breast mass images were acquired during PIONEER-01 study. The Reader-01 study-single arm, sequentially read (IUS vs. Imagio \[IUS+OA\]), controlled, blinded, multi-reader, multi-case (MRMC) feasibility study. New readers recruited; NO new subjects were recruited. Subject baseline, demographic and safety data comes from PIONEER-01 subjects who were randomly selected for Reader-01 study. Outcome measures compared IUS vs. IUS+OA. Truth based on biopsy diagnosis or 12 Mo F/U truth panel decision.
Overall Study
| Stage | Overall (Subjects Randomly Selected From PIONEER-01 Study) |
|---|---|
| STARTED | 155 |
| COMPLETED | 155 |
| NOT COMPLETED | 0 |
Analysis population for the primary outcome, as defined in the registry: 120 patients/masses were included in Cohort 1-the Intent-to-diagnose (ITD) population used to run the analysis to set the sample size for the future pivotal Reader-02 Study, 30 masses Cohort 2 to assess False negative rate, and 5 Cohort 3 DCIS, lymphoma, etc. masses to assess how readers scored. 155 masses in total across all cohorts analyzed for the Safety population analysis
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: Intent-to-diagnose (ITD)
| Characteristic | Unit | Overall (Subjects Randomly Selected From PIONEER-01 Study) |
|---|---|---|
| Participants analysed | 155 | |
| Age, Continuous | Years | 51.2 (SD 14.5) |
| Sex: Female, Male: Female | Participants | 155 |
| Sex: Female, Male: Male | Participants | 0 |
| Ethnicity (NIH/OMB): Hispanic or Latino | Participants | 24 |
| Ethnicity (NIH/OMB): Not Hispanic or Latino | Participants | 130 |
| Ethnicity (NIH/OMB): Unknown or Not Reported | Participants | 1 |
| Race (NIH/OMB): American Indian or Alaska Native | Participants | 0 |
| Race (NIH/OMB): Asian | Participants | 7 |
| Race (NIH/OMB): Native Hawaiian or Other Pacific Islander | Participants | 0 |
| Race (NIH/OMB): Black or African American | Participants | 24 |
| Race (NIH/OMB): White | Participants | 117 |
| Race (NIH/OMB): More than one race | Participants | 1 |
| Race (NIH/OMB): Unknown or Not Reported | Participants | 6 |
| Region of Enrollment: United States | Participants | 155 |
| Mass Diagnosis: Subjects: Benign (includes benign + Truth Panel Benign [TPB]) (Cohort 1) | Participants | 75 |
| Mass Diagnosis: Subjects: Malignant (Cohort 1) | Participants | 45 |
| Mass Diagnosis: Subjects: Malignant false negative in PIONEER-01 (Cohort 2) | Participants | 30 |
| Mass Diagnosis: Subjects: Malignant atypical (Cohort 3) | Participants | 5 |
6. Primary Endpoint
The registry lists the primary outcome as The Gain in Imagio IUS Alone Specificity vs Imagio (IUS+OA) Specificity at Fixed 95-99% Sensitivity (fSp), Cohort 1, measured over Baseline to 12 month follow-up.
| Element | Registry entry |
|---|---|
| Result type | Mean |
| Measure of spread | 95% Confidence Interval |
| Unit | Percent of correct benign masses |
| Endpoint classification | Continuous Endpoint |
The primary endpoint is the gain in specificity for Imagio Ultrasound alone versus Imagio (IUS+OA) at fixed 95-99% sensitivity, for Cohort 1. Specificity is the ability to classify benign masses correctly among masses that are benign according to the reference standard. The endpoint’s stated unit is percent of correct benign masses.
The registry description says specificity at the fixed sensitivity range was interpolated from the area under receiver operating characteristic curves and averaged across all 10 independent readers. The curves represent the trade-off between sensitivity and specificity as classification thresholds vary. Fixing sensitivity makes the specificity comparison relevant to a setting where the sensitivity target is held within the stated range, rather than comparing performance at potentially different sensitivity levels.
The ground truth was biopsy diagnosis or a 12-month follow-up ruling of benign, as determined by a truth panel. This reference provides the basis for classifying an imaging result as correct or incorrect for the endpoint. The timeframe is baseline to 12 month follow-up. Although the registry record defines the measure and its construction, a numerical mean, confidence-interval limits, and direction or magnitude of the gain are not reported in the registry record.
7. Why No Re-analysis Is Shown
8. Statistical Interpretation
The registry record specifies the primary endpoint as a mean, with a 95% Confidence Interval, for a continuous measure. The measure is the contrast in specificity between the two imaging conditions at fixed 95-99% sensitivity, averaged across readers in Cohort 1. The endpoint definition gives a clear statistical target, but the record does not report a numerical mean or confidence-interval bounds. It therefore does not permit an assessment of the observed size or precision of the primary contrast.
The paired nature of the design matters to interpretation. The record says both modalities were read for each subject and describes the subject as their own control. This means the intended comparison uses results from the same subject under each imaging condition, rather than treating the two conditions as wholly separate sets of subjects. The registry record does not provide reader-level or subject-level data, nor enough detail to reconstruct an unadjusted estimate.
Specificity at a fixed sensitivity is not a general summary of every aspect of diagnostic performance. It focuses on correctly identifying benign masses while holding sensitivity within the stated range. Other operating points on a receiver operating characteristic curve may have different specificities. The record also lists secondary outcomes that address other performance or interpretation topics, including partial ROC area, BI-RADS downgrades and upgrades, negative likelihood ratio, and SenoGram measures. Their numerical results are not reported in the registry record.
The record lists seven outcome measures but no analysis entries. No p value, test statistic, reader-specific result, or numerical primary estimate is reported. Without those results, the endpoint’s stated confidence interval cannot be used to judge how narrowly or broadly the effect is estimated, and the direction of any observed difference cannot be established from the registry record.
The registry identifies Cohort 1 as the intent-to-diagnose population used for the primary analysis and describes the analysis as a basis for setting the sample size for a future pivotal study. That stated purpose is distinct from showing a definitive clinical result. The record does not report a conclusion about whether one imaging condition had higher specificity, nor does it provide information needed to assess how that comparison would generalize beyond the described study population and reading conditions.
9. Statistical Methods Explained
The registry description gives some features of the endpoint’s statistical construction, but it does not provide a complete analysis plan or numerical results. The explanations below distinguish those recorded features from general statistical concepts; they should not be read as additional methods used by the study.
What does specificity describe?
As a general statistical concept, specificity is the proportion of truly negative cases correctly classified as negative. In this study’s endpoint, the relevant negative cases are benign masses, and a correct benign classification contributes to specificity. The endpoint is expressed as percent of correct benign masses. The truth panel’s biopsy diagnosis or 12-month follow-up ruling of benign supplies the reference classification described in the record.
Specificity depends on the classification threshold used to call a finding positive or negative. Moving a threshold can change the balance between detecting positive cases and correctly classifying negative cases. The registry’s primary endpoint avoids comparing specificity at different sensitivity levels by fixing sensitivity at 95-99%. This makes the contrast specifically about how the imaging conditions compare in specificity under the stated sensitivity constraint. It does not imply that specificity is constant at other sensitivity levels.
Why use ROC curves and interpolation?
As a general statistical concept, a receiver operating characteristic curve displays sensitivity against the false-positive rate across classification thresholds; the false-positive rate is related to specificity. The area under a ROC curve provides an overall summary across thresholds, while a partial ROC area focuses on a portion of the curve. The primary endpoint description says the fixed-sensitivity specificity was interpolated from the ROC curves. Interpolation estimates a value between observed points on a curve; it is not the same as directly observing a classification at precisely that point.
The registry record says the primary specificity measure was averaged across all 10 independent readers. Averaging reader results gives a summary across the reader group, but the record does not explain the exact averaging procedure or how reader-to-reader variability was handled in uncertainty estimates. Those details are not reported in the registry record. The record also lists partial ROC area as a secondary outcome, which is conceptually distinct from the primary fixed-sensitivity specificity comparison.
What does the paired design contribute?
As a general statistical concept, a paired comparison accounts for measurements made under multiple conditions on the same experimental unit. The registry says both imaging modalities were read for each subject and explicitly describes the subject as their own control. A paired design can separate within-subject differences from some between-subject variation, because each subject contributes information under both conditions. The size and precision of that benefit depend on the data and analysis, neither of which can be reconstructed from the registry record.
The study is also described as multi-reader, multi-case. In general, results in such studies may reflect variation among cases and variation among readers. The endpoint definition states that results were averaged across independent readers, but it does not report a model for reader or case variability, a method for constructing the confidence interval, or whether any additional adjustment was used. Those methods are not reported in the registry record.
How should the confidence interval be understood?
As a general statistical concept, a confidence interval represents the uncertainty in an estimate under the assumptions of the method used to calculate it. A narrower interval indicates greater precision under those assumptions; a wider interval indicates less precision. The stated 95% Confidence Interval is part of the primary endpoint definition, but neither its bounds nor the method used to calculate it is reported in the registry record. No inference about precision can therefore be made from the endpoint definition alone.
The registry record identifies the endpoint parameter as a mean of a continuous measure. It does not report a primary estimate, a p value, or a statistical test. General explanations of means and confidence intervals do not fill that results gap: the record’s numerical estimate and its uncertainty would be needed to characterize the observed contrast.
What do the planning illustrations represent?
As a general statistical concept, a sample-size calculation estimates the number of participants or events that would be needed under stated assumptions to detect a specified effect with a chosen error rate and power. The registry’s planning assumptions describe an illustration based on the observed effect, not the trial’s own design. The assumptions state a two-sided significance level of 0.05, equal allocation between arms, a normal approximation, and an effect specified as a fraction of the observed effect. For time-to-event tables, the stated convention is total events; for other outcomes, it is participants per arm.
These calculations are planning illustrations, not evidence that the study had a particular probability of detecting its effect. The record does not report the calculated sample-size values or the numerical observed effect on which the illustrations are based. The actual rationale or sample-size justification for this feasibility study is not reported in the registry record. No observed-power claim follows from the planning assumptions.
10. Multiplicity, Missing Data and Analysis Populations
The registry record lists seven outcome measures. The secondary outcomes include partial ROC area and BI-RADS downgrades and upgrades for Cohort 1, the negative likelihood ratio comparison for Cohort 1, SenoGram usage for Imagio (IUS+OA) only, and SenoGram performance. SenoGram performance appears twice in the listed outcome names. The list indicates that several aspects of imaging performance or reader interpretation were of interest, but it does not describe how the analyses were prioritized or interpreted together.
Multiplicity refers, as a general statistical concept, to the increased chance of at least one apparently notable result when many statistical tests are conducted. Possible approaches include adjusting for multiple comparisons or defining a hierarchy of outcomes in advance. No multiplicity adjustment, testing hierarchy, or other strategy is reported in the registry record. It is therefore not possible to determine whether or how the listed outcomes were handled collectively.
The record lists no analysis entries. It does not provide numerical results for the primary endpoint or secondary outcomes, nor does it report a missing-data method. The extent of missing measurements, the handling of unreadable images or incomplete follow-up, and any exclusions from analyses are not reported in the registry record.
The primary population is described as 120 patients/masses in Cohort 1 and is labeled the intent-to-diagnose population used for the analysis. The record separately describes 30 masses in Cohort 2 and 5 masses in Cohort 3, for different stated purposes. It does not report detailed eligibility rules, participant flow, or the number of cases contributing to each reader’s analysis. The 155 enrollment figure alone does not establish which enrolled cases contributed to a specific endpoint analysis.
11. Safety Results
The registry reports serious adverse events as affected participants divided by participants at risk.
| Arm | Participants with serious adverse events | Participants at risk | Percentage |
|---|---|---|---|
| Overall (Masses Randomly Selected From PIONEER-01 Study) | 0 | 155 | 0.0% |
Deaths
| Arm | Deaths (participants affected of at risk) |
|---|---|
| Overall (Masses Randomly Selected From PIONEER-01 Study) | 0 of 155 |
Other (non-serious) adverse events
Participants with at least one non-serious event at or above the registry's 0% reporting threshold.
| Arm | Participants affected of at risk |
|---|---|
| Overall (Masses Randomly Selected From PIONEER-01 Study) | 0 of 155 |
The registry safety entry is labeled Overall (Masses Randomly Selected From PIONEER-01 Study). It records 155 at risk for serious adverse events and 0 affected. For other adverse events, it records 155 at risk and 0 affected. The death count is 0, and no serious adverse event terms are listed.
These figures describe the entries in the registry safety record. They should not be read as a detailed account of every possible safety outcome: the record does not provide event narratives, severity breakdowns, or additional context about ascertainment. It also does not explain how the safety group relates to the Cohort 1, Cohort 2, and Cohort 3 endpoint populations. The safety counts therefore do not establish that no events occurred outside the categories or reporting captured in this entry.
The record concerns an imaging-device study and describes the safety group as masses randomly selected from PIONEER-01. The listed counts support only the narrow statement that no serious or other adverse events and no deaths were recorded in this safety entry. Further detail on safety monitoring and reporting is not reported in the registry record.
12. What the Result Does, and Does Not, Mean
The registry record establishes what the primary endpoint was intended to measure: specificity for Imagio Ultrasound alone versus Imagio Ultrasound plus optoacoustic imaging at fixed 95-99% sensitivity, for Cohort 1. It also describes the reference standard, reader averaging, and paired reading arrangement. These details clarify the comparison the study was designed to make.
However, a defined endpoint is not itself a result. The registry record does not report the primary mean, the 95% Confidence Interval bounds, or a test result. It therefore does not show the observed magnitude, direction, or precision of the specificity contrast. The listed secondary outcomes also lack numerical results in the registry record. No conclusion about diagnostic performance can be drawn from the endpoint definition alone.
The record describes a feasibility multi-reader, multi-case study and states that the Cohort 1 analysis was used to set the sample size for a future pivotal Reader-02 Study. That stated role is relevant to interpreting the design: the record frames this analysis as informing later study planning, not as a complete report of confirmatory evidence. It does not report the sample-size value derived for that future study or the rationale for the feasibility study’s enrollment.
The results cannot be extended from this record to claims about patient outcomes, clinical decision-making, or performance in settings beyond the described readers, cases, and imaging conditions. Nor can absence of posted numerical results be taken as evidence that the imaging conditions were equivalent or different. The registry record supports a description of the question and recorded design, while the central estimate and its uncertainty remain not reported in the registry record.
13. Why This Trial Matters Statistically
This study is statistically notable for its multi-reader, multi-case setup and its paired comparison. The registry record states that each subject was read under both imaging conditions and served as their own control. That structure is relevant because diagnostic imaging comparisons can be affected by differences among cases; pairing makes the within-subject comparison central to the question.
The primary endpoint also makes a specific operating-point comparison rather than relying only on a broad summary of performance. By evaluating specificity at fixed 95-99% sensitivity, it focuses on correct benign classifications while holding the sensitivity target in the stated range. This highlights why the operating point and reference standard matter when interpreting diagnostic measures. The endpoint uses a truth-panel determination based on biopsy diagnosis or a 12-month follow-up ruling of benign.
The study’s stated feasibility purpose also connects the analysis to later research planning. The registry says the Cohort 1 analysis was used to set the sample size for a future pivotal Reader-02 Study. Planning based on an observed effect requires care, and the registry’s separate sample-size illustrations are identified as planning rather than as the trial’s own design or observed power.
At the same time, the record illustrates the difference between a well-described statistical question and a reported result. It provides an endpoint definition and study structure, but no numerical primary estimate, confidence-interval bounds, or analysis entries. This limits what can be learned about the actual size and precision of the contrast from the registry record.
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 most important limitation for interpreting the primary question is the absence of a numerical endpoint result. The registry record specifies a mean and a 95% Confidence Interval but does not report the estimate or interval bounds. Without them, the size, direction, and precision of the comparison cannot be assessed.
The endpoint description gives some methodological detail, including interpolation from ROC curves, averaging across 10 independent readers, and a subject-as-own-control design. It does not report the precise interpolation method, the construction of the confidence interval, or how reader and case variability were represented. The masking identity, allocation implementation, and reading sequence are also not reported in the registry record.
The Cohort 1 population is described as 120 patients/masses and as an intent-to-diagnose population. The record does not provide participant flow, detailed analysis inclusion rules, the completeness of follow-up, or the extent and handling of missing data. It also does not explain the relationship between the safety group and the endpoint cohorts. Those omissions limit assessment of the analysis population and of possible selection or missing-data effects.
The registry lists seven outcome measures and no analysis entries. It does not report a multiplicity strategy, a detailed statistical analysis plan, or numerical results for the secondary outcomes. Although the record identifies Cohort 1 as supporting planning for a future pivotal study, the actual sample-size rationale for this feasibility study is not reported in the registry record. These limitations concern the information available in the record and do not establish that a particular bias or analytical problem occurred.
15. Related Tutorials
16. Related Statistical Calculators
17. Sources
- ClinicalTrials.gov: Imagio Feasibility Multi-Reader, Multi-Case Study — NCT03708393.
- Registry record last updated 2021-08-06; status last verified 2021-07.
- Results first posted 2021-08-06.
- 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.
Continue through the Clinical Biostats statistical pathway
Use this trial's endpoint as a starting point for deeper study of effect measures, confidence intervals, and sample-size planning.
18. Record Summary
NCT03708393 is a completed randomized, single-masked, sequential device study described as controlled, blinded, multi-reader, and multi-case. It enrolled 155 and compared Imagio Ultrasound alone with Imagio Ultrasound plus optoacoustic imaging, with mammography as available in the read descriptions. Both imaging conditions were read for each subject.
The primary endpoint was the gain in specificity at fixed 95-99% sensitivity for Cohort 1, using ROC-based interpolation and an average across 10 independent readers. Cohort 1 is described as 120 patients/masses in the intent-to-diagnose population. The record also describes Cohort 2 and Cohort 3 for separate purposes.
The primary endpoint’s numerical estimate and confidence-interval bounds are not reported in the registry record. The record lists seven outcome measures and no analysis entries. Its safety entry records 0 serious adverse events, 0 other adverse events, and 0 deaths.