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 Comparison of the Philips MicroDose Tomosynthesis System to 2D Digital Mammography (NCT02615509), a na na study. The primary outcome is To Compare the Clinical Performance of the Philips MicroDose Tomosynthesis System Images to Conventional 2D Mammography Images. Clinical Performance Will be Based on the Area Under the Receiver Operating Characteristics Curve..
| Feature | Registry value |
|---|---|
| ClinicalTrials.gov identifier | NCT02615509 |
| Phase | NA |
| Condition | Breast Cancer |
| Allocation | Not applicable (single group) |
| Design model | Single Group |
| Masking | None (open label) |
| Enrollment | 347 |
| Arms in the primary outcome | n/a |
| Lead sponsor | Philips Digital Mammography Sweden AB |
| Results posted | Yes |
2. Clinical Question
Population
The registry describes women undergoing breast imaging and records an enrollment of 347. The characteristics of the participants and the criteria used to select them are not reported in the registry record.
Intervention
The experimental imaging procedure was a bilateral two-view mammogram using the Philips MicroDose Tomosynthesis system. The study description also refers to four images from this system.
Comparator
The intended comparator was conventional 2D mammography using an FFDM system. The registry says conventional images were not taken before the study terminated, so the planned image-based comparison was not completed.
Primary question
The primary question was whether clinical performance based on the area under the receiver operating characteristics curve differed between Philips MicroDose Tomosynthesis images and conventional 2D mammography images. The registry says this comparison could not be calculated. The record instead reports biopsy results from the Philips system, but numerical biopsy findings are not reported in the registry record.
3. Trial Design
This was a single-group imaging study of the Philips MicroDose Tomosynthesis system. The registry lists the phase as not applicable, allocation as not applicable, and masking as none. The recorded enrollment is 347. The study status is terminated.
The study description planned for women to receive bilateral two-view mammograms using both the Philips system and an FFDM system. It says the order of these imaging procedures would be randomized. That planned order does not make this a randomized comparison of separate treatment arms: the registry identifies a single-group study, and its allocation field is not applicable. More importantly, the record states that conventional images were not taken because the study terminated. Therefore, the intended within-participant image comparison was not completed as described.
The stated primary objective was to compare safety and clinical performance of Philips tomosynthesis images with images from FFDM systems. The primary outcome was specified as clinical performance based on the area under the receiver operating characteristics curve. The registry says that this value could not be calculated. Instead, it reports biopsy results from the Philips system, without numerical results in the registry record.
Secondary outcomes concerned average glandular dose, non-cancer recall rates, sensitivity, and specificity for the imaging systems. The record describes sensitivity and specificity as reader-level measures based on BI-RADS categories. Numerical findings for these secondary outcomes are not reported in the registry record. The reported timeframe for the primary outcome is up to approximately 4 weeks; the registry describes the examination-to-biopsy interval as an estimate based on the relative urgency of detecting potential tumors.
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
| Stage | Imaging on Experimental Tomo Device |
|---|---|
| STARTED | 347 |
| COMPLETED | 201 |
| NOT COMPLETED | 146 |
Reasons for not completing this period:
| Reason | Imaging on Experimental Tomo Device |
|---|---|
| Lost to Follow-up | 134 |
| Death | 4 |
| Withdrawal by Subject | 8 |
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.
| Characteristic | Unit | Imaging on Experimental Tomo Device |
|---|---|---|
| Participants analysed | 347 | |
| Age, Categorical: <=18 years | Participants | 0 |
| Age, Categorical: Between 18 and 65 years | Participants | 243 |
| Age, Categorical: >=65 years | Participants | 104 |
| Sex: Female, Male: Female | Participants | 347 |
| Sex: Female, Male: Male | Participants | 0 |
| Ethnicity (NIH/OMB): Hispanic or Latino | Participants | 0 |
| Ethnicity (NIH/OMB): Not Hispanic or Latino | Participants | 347 |
| Ethnicity (NIH/OMB): Unknown or Not Reported | Participants | 0 |
| Race/Ethnicity, Customized: American indian or Alaskan native | Participants | 4 |
| Race/Ethnicity, Customized: Asian | Participants | 5 |
| Race/Ethnicity, Customized: African American | Participants | 105 |
| Race/Ethnicity, Customized: White | Participants | 233 |
| Region of Enrollment: United States | Participants | 347 |
6. Primary Endpoint
The registry lists the primary outcome as To Compare the Clinical Performance of the Philips MicroDose Tomosynthesis System Images to Conventional 2D Mammography Images. Clinical Performance Will be Based on the Area Under the Receiver Operating Characteristics Curve., measured over Up to approximately 4 weeks..
| Element | Registry entry |
|---|---|
| Result type | Count Of Participants |
| Measure of spread | not applicable |
| Unit | Participants |
| Endpoint classification | Binary Endpoint |
The stated primary endpoint was clinical performance measured by the area under the receiver operating characteristics curve, or AUC, for Philips MicroDose Tomosynthesis images compared with conventional 2D mammography images. In general, an ROC curve describes how a test’s sensitivity and specificity vary across possible classification thresholds. Its area provides a summary of discrimination across those thresholds rather than describing performance at only one threshold.
The registry record says the conventional images were not taken because the study terminated. As a result, the planned AUC comparison could not be calculated. The record instead reports biopsy results from the Philips system, but it does not provide numerical biopsy findings in the registry record. Those biopsy results are not the planned comparative AUC endpoint, and they cannot establish the relative performance of the two imaging systems by themselves.
The stated timeframe is up to approximately 4 weeks. The registry says the examination-to-biopsy timeframe is an estimate informed by the relative urgency of detecting potential tumors. It does not report a calculated AUC, a comparison estimate, or its precision. The endpoint’s intended role was to compare clinical performance across imaging modalities; the available record does not supply the comparative data needed to answer that question.
7. Why No Re-analysis Is Shown
8. Statistical Interpretation
The central interpretive point is that the registry describes an intended comparison but also explains why its primary comparison could not be performed. Conventional images were not taken following study termination, so there is no reported AUC contrast between the Philips MicroDose Tomosynthesis system and conventional 2D mammography. The record’s stated substitute is biopsy results from the Philips system, but numerical biopsy findings are not reported in the registry record.
This means the primary question remains unanswered by the posted results. The record does not show whether one imaging approach had higher or lower discrimination, nor does it provide an estimate of the size or precision of any difference. It would be inappropriate to treat the change in reported outcome as if it were a completed comparison. Biopsy results from one imaging system do not, on their own, provide a comparative estimate against a system for which images were not taken.
The secondary outcomes are also not numerically reported in the registry record. These include average glandular dose, non-cancer recall rates, sensitivity, and specificity. The record describes the intended reader-level definitions for sensitivity and specificity, but does not give estimates, uncertainty intervals, or analysis results. No analyses are listed in the registry record. Consequently, the available information does not support a quantitative conclusion about the planned comparative clinical performance or secondary outcomes.
The study status is terminated, and the record identifies this as the reason the conventional images were not taken. Termination is important context for understanding why the planned endpoint is unavailable; it is not itself evidence about the accuracy, safety, or comparative performance of either imaging system. The registry does not report a reason for termination beyond the stated link to the missing conventional images.
Safety counts should be read separately from the efficacy question. For the experimental tomosynthesis group, the record lists 347 participants at risk, zero participants with serious adverse events, zero with other adverse events, and four deaths. It does not give event descriptions, timing, causes, or an assessment of whether the deaths were related to imaging. These counts therefore do not establish causation or provide a comparison with FFDM.
The record is best interpreted as an incomplete report of a planned modality comparison, with no reported primary comparative estimate and no numerical secondary outcome results. Any sample-size figures presented as planning illustrations are not observed power and do not repair the absence of the planned comparative data.
9. Statistical Methods Explained
The intended primary measure was the area under the receiver operating characteristics curve. In general, an ROC curve represents the trade-off between sensitivity and specificity as a classification threshold changes. Sensitivity describes the identification of cancer cases under a stated threshold; specificity describes the classification of non-cancer cases. The area under that curve is commonly used to summarize discrimination across thresholds, which is why it can serve as a broad measure of diagnostic performance. It does not, by itself, identify performance at a particular threshold or express the consequences of a particular classification decision.
For a comparison of imaging systems, the same underlying cases and an appropriate reference standard are needed to estimate and compare performance. The study description planned imaging with both the Philips system and an FFDM system, with the order randomized. However, the registry states that conventional images were not taken. Without those comparator images, the planned comparative AUC could not be calculated. The record explicitly says that it instead reports biopsy results from the Philips system. Numerical biopsy results are not reported in the registry record, so the statistical magnitude and precision of that alternative result cannot be assessed from the record.
What do the secondary measures represent?
The registry lists average glandular dose, non-cancer recall rates, sensitivity, and specificity as secondary outcomes. In general, dose is a measure of radiation exposure, while a recall rate describes how often additional assessment is requested after an initial imaging result. Sensitivity and specificity describe classification against a reference status: sensitivity concerns cancer cases and specificity concerns non-cancer cases. The registry specifies reader-level definitions using BI-RADS categories, with category 4 or higher for sensitivity and less than 4 for specificity. These general definitions explain what the measures are intended to capture; they do not supply the missing study estimates. Numerical results for these outcomes are not reported in the registry record.
The record does not list analyses. It does not report a statistical model, an adjustment strategy, a method for combining reader assessments, a confidence interval, or a hypothesis test for the primary comparison. Those details are not reported in the registry record. Therefore, no particular inferential method should be attributed to the investigators beyond the endpoint description itself.
How should uncertainty be understood?
In general, a confidence interval describes the range of values compatible with an estimate under a specified statistical procedure and its assumptions. Its width conveys precision: a wider interval indicates greater uncertainty, while a narrower interval indicates greater precision. A confidence interval is not reported for the intended AUC comparison, and there is no reported comparative estimate around which such an interval could be interpreted. The record therefore does not establish either the direction or precision of a modality difference.
For binary outcomes, general statistical methods often compare event proportions or model the probability of an event. The choice of method depends on the design, outcome definition, and data structure. In this record, the listed secondary outcomes include binary classifications, but their numerical results and analysis methods are not reported. Reader-specific sensitivity and specificity may also require methods that account for the fact that measurements can come from multiple readers evaluating the same cases. The registry does not report how such dependence was handled, or whether it was handled at all.
What do the planning illustrations mean?
Any sample-size figures associated with the results are planning illustrations based on an observed effect, not the original trial design and not observed power. The stated assumptions are a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation. The illustrated effect to detect is a stated fraction of the observed effect. For time-to-event tables, the stated quantity is the total number of events using the Schoenfeld method; for other outcomes, it is participants per arm.
These assumptions describe an illustrative calculation, not evidence that the study was designed with those parameters or achieved a particular probability of detecting an effect. The registry does not report a sample-size justification or its original planning assumptions. Because the planned AUC comparison could not be calculated and the numerical substitute findings are not reported in the registry record, the observed effect underlying any illustration cannot be evaluated from the record. A planning illustration should not be read as a result of the trial, as proof of adequate precision, or as a substitute for a completed comparative analysis.
10. Multiplicity, Missing Data and Analysis Populations
The registry lists a primary endpoint and several secondary outcomes, but it does not report a multiplicity strategy. In general, evaluating several outcomes or several statistical tests can increase the chance of observing an apparently notable result by chance. Researchers may address this by defining a hierarchy of outcomes, adjusting significance thresholds, or treating secondary analyses as exploratory. Which approach was intended for this study is not reported in the registry record.
Missing-data handling is also not reported in the registry record. The clearest missing element is the conventional imaging: the record says those images were not taken after termination, making the intended comparative AUC unavailable. The record does not describe imputation, exclusion rules, or other methods for missing participant-level or reader-level measurements. It also does not report analysis populations, such as rules for which participants or cases would be included in each endpoint analysis.
No analyses are listed. Thus, the record does not provide the number of cases contributing to each endpoint, how readers or repeated assessments were handled, or whether any sensitivity analyses were planned or performed. These omissions matter because analysis populations and missing-data procedures affect what an estimate represents. They also limit assessment of whether the intended outcomes and analyses were carried through consistently.
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 |
|---|---|---|---|
| Imaging on Experimental Tomo Device | 0 | 347 | 0.0% |
Deaths
| Arm | Deaths (participants affected of at risk) |
|---|---|
| Imaging on Experimental Tomo Device | 4 of 347 |
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 |
|---|---|
| Imaging on Experimental Tomo Device | 0 of 347 |
The registry lists one group, Imaging on Experimental Tomo Device. It records 347 participants at risk for serious adverse events and zero participants with a serious adverse event. It also records 347 at risk for other adverse events and zero with other adverse events. The same group lists four deaths.
These are counts in the registry record, not a comparative safety estimate. There is no FFDM image group with reported safety counts to serve as a comparator, and the single-group design does not support attributing differences between imaging systems. The record does not provide descriptions or timing for the deaths, their causes, or whether investigators considered them related to the imaging procedure. It also does not provide additional information about the scope or duration of adverse-event follow-up.
A recorded zero for reported adverse events does not establish that no event could occur; it describes what the registry lists for this group. Conversely, the four deaths should not be interpreted as caused by the imaging system without information on cause and attribution, which is not reported in the registry record. The safety counts do not answer the intended comparative clinical-performance question.
12. What the Result Does, and Does Not, Mean
The result’s main meaning is that the planned comparison of Philips MicroDose Tomosynthesis with conventional 2D mammography was not completed as described. The registry says the study terminated before conventional images were taken, so the primary AUC comparison could not be calculated. This is a limitation of the available evidence, not a finding that either system performed better or worse.
The registry says it reports biopsy results from the Philips system instead. However, numerical biopsy findings are not reported in the registry record. Without those values, readers cannot assess the number or pattern of biopsy outcomes, their uncertainty, or their relationship to the planned endpoint. And without conventional images, biopsy findings from the Philips system alone cannot establish a relative performance difference.
The listed secondary measures could have addressed other aspects of imaging performance, including dose, recall, sensitivity, and specificity. Their numerical findings are not reported in the registry record, and no analyses are listed. The available record therefore does not provide quantitative estimates for those outcomes or a basis for judging their precision.
The safety information describes event counts for the experimental imaging group. It does not provide a comparison between imaging systems, and the record does not report whether the deaths were related to imaging. These counts should not be converted into claims about cause or comparative risk.
In general, a study can be informative about feasibility or about the reason a planned analysis was unavailable even when it cannot answer its original comparative question. Here, the record establishes that termination prevented the intended image comparison. It does not establish a treatment effect, diagnostic advantage, or equivalence. Any sample-size calculations presented as planning illustrations should be kept separate from the observed evidence: they are not observed power and cannot compensate for an unavailable primary estimate.
13. Why This Trial Matters Statistically
This study illustrates why a planned endpoint and the data ultimately available for analysis need to be distinguished. The registry’s primary question concerned comparative clinical performance using AUC, but the conventional images required for that comparison were not taken. Once the comparator data are absent, the intended contrast cannot be recovered simply by reporting another outcome from the experimental system.
It also illustrates the importance of distinguishing a study’s imaging sequence from its allocation design. The description says the order of the two planned imaging procedures would be randomized, while the registry records a single-group model and allocation as not applicable. Randomized order can address sequence-related issues in a paired imaging procedure, in general, but it does not create a comparison when one modality’s images were never obtained.
Statistically, the record shows that interpretation depends not only on the endpoint’s definition but also on whether the required observations exist and whether an analysis is reported. The stated sensitivity and specificity definitions are meaningful in principle, but without estimates, analysis details, and information about the cases and readers contributing data, their statistical implications cannot be judged.
The safety counts also demonstrate why event categories and attribution should remain distinct. The registry lists adverse-event counts and deaths, but does not describe causes or relatedness. Statistical reporting should preserve that distinction rather than treating a count as evidence of causation.
Finally, planning calculations and empirical results answer different questions. A planning illustration under stated assumptions can describe sample sizes associated with a hypothetical target effect. It does not show that an effect was observed, that an analysis had a particular power, or that a trial’s planned comparison was completed. This distinction is especially important when the registry reports that the primary comparative endpoint could not be calculated.
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 consequential trial-specific limitation is the absence of conventional images. The registry directly links this to study termination and states that the intended AUC could not be calculated. The alternative mentioned in the record—biopsy results from the Philips system—does not restore the missing comparison, and numerical biopsy findings are not reported in the registry record.
The study is recorded as single-group, with allocation not applicable and masking none. Although the description planned randomized order for imaging with both systems, it does not describe a completed two-system comparison because the conventional images were not taken. The record does not report whether any other part of the intended imaging procedure was completed for all participants.
Numerical results for the secondary outcomes are not reported in the registry record. The record also does not list analyses, confidence intervals, hypothesis tests, or details about reader-level analysis. It does not report participant characteristics, endpoint-specific denominators, missing-data methods, analysis populations, or a multiplicity strategy. These omissions restrict assessment of representativeness, precision, and the handling of repeated reader assessments.
The registry’s safety counts have limited context: it does not report event descriptions, timing, causes, or relatedness for the four deaths. The record also does not state the reason for termination beyond the fact that conventional images were not taken. These limitations prevent stronger conclusions about the intended comparison, the secondary outcomes, and the interpretation of safety events.
15. Related Tutorials
16. Related Statistical Calculators
17. Sources
- ClinicalTrials.gov: Comparison of the Philips MicroDose Tomosynthesis System to 2D Digital Mammography — NCT02615509.
- Registry record last updated 2020-05-08; status last verified 2020-04.
- Results first posted 2020-05-08.
- Data retrieved from ClinicalTrials.gov on 2026-11-09.
- Group sizes, means, event counts and sponsor-reported p-values are reproduced as posted. Effect estimates, confidence intervals and p-values not attributed to the sponsor are independently calculated from the posted summary statistics.
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
The registry describes a single-group imaging study comparing the intended clinical performance of Philips MicroDose Tomosynthesis images with conventional 2D mammography images. It records enrollment of 347 and a terminated status. The study description planned bilateral two-view imaging with both systems and randomized the order of imaging.
The primary endpoint was clinical performance based on AUC. The registry says conventional images were not taken because the study terminated, making the AUC comparison impossible to calculate. It instead reports biopsy results from the Philips system, but numerical biopsy findings are not reported in the registry record. Numerical results for the listed secondary outcomes are also not reported, and no analyses are listed.
For the experimental tomosynthesis group, the registry lists zero serious adverse events, zero other adverse events, and four deaths. It does not report causes or relatedness. The available record therefore documents why the intended comparison could not be completed, but does not provide a numerical comparative performance estimate.