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 SonoVue®-Enhanced Ultrasound Versus Unenhanced US for Focal Liver Lesion Characterization (NCT00788697), a na phase 3 study. The primary outcome is Sensitivity: Percentage of True Positive Lesions Among All Malignant Lesions Per Truth Standard.
| Feature | Registry value |
|---|---|
| ClinicalTrials.gov identifier | NCT00788697 |
| Phase | Phase 3 |
| Condition | Liver Neoplasms |
| Allocation | Not applicable (single group) |
| Design model | Single Group |
| Masking | None (open label) |
| Enrollment | 349 |
| Arms in the primary outcome | n/a |
| Lead sponsor | Bracco Diagnostics, Inc |
| Results posted | Yes |
2. Clinical Question
Population
Participants had at least one target focal liver lesion requiring diagnostic work-up. The primary endpoint used the Intent-to-Diagnose population: 240 participants. Of those, 124 had a lesion classified as malignant by the truth standard and were included in the sensitivity analysis. The registry describes each participant as having a single lesion for characterization, so the lesion and participant were equivalent units of analysis for this endpoint.
Intervention
SonoVue-enhanced ultrasound was assessed for characterization of focal liver lesions. The study concerns the imaging assessment and its ability to classify a lesion as malignant or not malignant against a defined truth standard.
Comparator
The comparator was unenhanced ultrasound of the target lesion. The study is described as an intrapatient comparison, not as a randomized comparison between separate groups. Allocation is recorded as not applicable, and masking is recorded as none.
Primary question
Among lesions classified as malignant by the truth standard, what proportion were correctly characterized as malignant by each ultrasound method? The registry planned to assess sensitivity using diagnoses from each of three off-site assessors, who were blinded to patient data. The posted registry record does not report the resulting sensitivity estimates or their confidence intervals, so the comparative size and precision of the observed sensitivity results cannot be described.
The truth standard could be contrast-enhanced computed tomography, contrast-enhanced magnetic resonance imaging, tissue pathology or histology from resection or biopsy, or six-month follow-up. The primary endpoint timeframe is recorded as 24 hours to 6 months. This question concerns diagnostic classification against that reference, rather than a treatment effect on a clinical outcome.
3. Trial Design
This was a Phase III, completed, intrapatient comparative study of SonoVue-enhanced ultrasound and unenhanced ultrasound for focal liver lesion characterization. The registry lists 349 participants as enrolled. The study began in 2009-09, and primary completion is recorded as 2013-07. The condition is liver neoplasms, and the sponsor is Bracco Diagnostics, Inc.
An intrapatient comparison evaluates both imaging approaches within the same participant, rather than assigning separate participants to competing treatment groups. This design can reduce the influence of differences between participants when comparing diagnostic readings, because the assessments concern the same target lesion. The registry identifies a single-group model and records allocation as not applicable. It also records masking as none. The study is therefore not a randomized comparison, and its arms should not be read as like-for-like randomized control groups.
The primary endpoint was sensitivity for malignant focal liver lesions, evaluated using diagnoses from each of three off-site assessors who were blinded to patient data. The unit of analysis was the lesion, equivalent to the subject because each subject had a single lesion to characterize. The truth standard could draw on contrast-enhanced computed tomography or magnetic resonance imaging, pathology or histology, or six-month follow-up. The registry records an endpoint timeframe of 24 hours to 6 months.
The primary analysis population included 240 Intent-to-Diagnose participants; the sensitivity denominator was restricted to the 124 participants whose lesions were malignant according to the truth standard. The record lists eight outcome measures and 15 analyses. It also names secondary outcomes covering accuracy, predictive values, specific diagnoses of malignant and benign lesions, and inter-reader agreement. The registry does not report a detailed statistical analysis plan, the method used to compare paired readings, or the results for the primary endpoint.
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: A total of 74 patients received SonoVue in the training phase and were included only in safety population. A total of 263 patients received SonoVue in the efficacy phase. The 74 patients, included within the "Not Completed" category below, are the training patients.
Overall Study
| Stage | Safety Population |
|---|---|
| STARTED | 349 |
| COMPLETED | 240 |
| NOT COMPLETED | 109 |
Reasons for not completing this period:
| Reason | Safety Population |
|---|---|
| No study drug administered | 12 |
| Training Patients | 74 |
| no truth standard procedure available | 5 |
| Technically inadequate truth standard | 1 |
| Indeterminate diagnosis from truth stand | 17 |
Analysis population for the primary outcome, as defined in the registry: Among the 240 ITD participants, only 124 participants (lesions) were malignant based on the truth standard and were included for sensitivity.
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 | ITD Population |
|---|---|---|
| Participants analysed | 240 | |
| Age, Categorical: <=18 years | Participants | 0 |
| Age, Categorical: Between 18 and 65 years | Participants | 190 |
| Age, Categorical: >=65 years | Participants | 50 |
| Sex: Female, Male: Female | Participants | 117 |
| Sex: Female, Male: Male | Participants | 123 |
| Race/Ethnicity, Customized: White | participants | 161 |
| Race/Ethnicity, Customized: Black | participants | 32 |
| Race/Ethnicity, Customized: Asian | participants | 15 |
| Race/Ethnicity, Customized: Other | participants | 32 |
| Region of Enrollment: United States | participants | 225 |
| Region of Enrollment: Europe | participants | 15 |
6. Primary Endpoint
The registry lists the primary outcome as Sensitivity: Percentage of True Positive Lesions Among All Malignant Lesions Per Truth Standard, measured over 24 hours to 6 months.
| Element | Registry entry |
|---|---|
| Result type | Number |
| Measure of spread | 95% Confidence Interval |
| Unit | Percentage of true positive lesions |
| Endpoint classification | not reduced to a single comparison |
Sensitivity is the proportion of lesions classified as malignant by the truth standard that an ultrasound assessment also classifies as malignant. A true positive is defined in the registry as a target lesion characterized as malignant by both ultrasonography and the truth standard. The endpoint therefore focuses on detection of malignancy among lesions known to be malignant by the reference classification; it is not a measure of performance across all lesions.
The registry specifies sensitivity for SonoVue-enhanced ultrasound versus unenhanced ultrasound, using the diagnosis from each of three off-site assessors blinded to patient data. The lesion was the unit of analysis and was equivalent to the participant because each participant had a single lesion for characterization. Among the 240 Intent-to-Diagnose participants, 124 lesions were malignant by the truth standard and included in the sensitivity analysis.
The truth standard could be contrast-enhanced computed tomography, contrast-enhanced magnetic resonance imaging, tissue pathology or histology from surgical resection or biopsy, or six-month follow-up. These sources establish the reference classification against which ultrasound readings are judged. Because the truth standard can comprise different sources, interpretation depends on how those sources were applied and combined; the registry description does not provide further operational details.
Sensitivity is useful for describing how often a diagnostic method identifies malignancy in reference-positive cases. It does not, by itself, describe false-positive classifications, overall accuracy, or performance in other populations. Although the endpoint specifies a 95% confidence interval, the registry record does not report the sensitivity estimates or interval limits.
7. Why No Re-analysis Is Shown
8. Statistical Interpretation
The registry identifies the primary question and analysis population but does not report the numerical primary results. In particular, no sensitivity estimate for either ultrasound approach, no between-method contrast, and no confidence-interval limits are reported in the registry record. It is therefore not possible to describe the observed magnitude of the comparison, its direction, or its statistical precision from the posted results. The study’s stated purpose was to demonstrate superiority of enhanced ultrasound, but a stated objective is not evidence that the objective was met.
The sensitivity endpoint concerns only lesions classified as malignant by the truth standard. The registry identifies 124 such participants within the Intent-to-Diagnose population of 240. This subset defines the population relevant to sensitivity; it should not be confused with all enrolled participants or with the safety population. The registry’s description states that each participant had a single lesion for characterization, so lesion-level and participant-level units coincide for this endpoint.
The comparison was intrapatient, meaning the intended diagnostic comparison concerns assessments of the same target lesion by the two ultrasound approaches. As a general statistical principle, paired diagnostic readings should be interpreted with attention to the within-participant relationship. Treating the readings as if they came from unrelated groups could fail to reflect that pairing. The registry record does not report the statistical method used to account for the paired assessments, nor does it describe a formal superiority test or a decision rule.
Three off-site assessors contributed diagnoses, and the registry says they were blinded to patient data. Results may depend on how readings from assessors were handled, including whether each assessor’s performance was reported separately or combined. The registry record does not report that aggregation method or the assessor-specific numerical results. It also does not report the primary confidence intervals, so the degree of uncertainty around sensitivity cannot be judged.
The truth standard included imaging, pathology or histology, or follow-up. These are distinct sources of diagnostic information, and the record does not explain whether each participant received the same reference procedures or how discordant sources were resolved. That matters because sensitivity is defined relative to the truth standard: a change in reference classification can change which lesions enter the malignant denominator and which ultrasound classifications count as true positives.
Secondary outcome names include accuracy, predictive values, specific diagnoses of malignant and benign focal liver lesions, and inter-reader agreement. Their inclusion indicates additional planned areas of assessment, but the registry record does not provide their numerical results in the registry record described for this analysis. No conclusions about those outcomes, or about superiority, should be inferred from the endpoint definitions alone.
9. Statistical Methods Explained
The registered primary endpoint is sensitivity, expressed as the percentage of true-positive lesions among lesions classified as malignant by the truth standard. The registry defines a true positive as a target lesion called malignant by both ultrasound and the truth standard. This is a conditional measure: its denominator is reference-positive lesions, not every lesion assessed. It describes detection within that specific group and does not, by itself, describe how often benign lesions are incorrectly called malignant.
What does a confidence interval add?
The endpoint specifies a 95% confidence interval. In general, a confidence interval expresses the statistical uncertainty around an estimate, given the data and method used. It is not a guarantee that the true value lies within the interval, and it does not describe all sources of uncertainty, such as possible limitations in the reference classification or study conduct. A narrower interval generally indicates greater statistical precision under the analysis assumptions; a wider interval indicates less precision. The registry record does not report the sensitivity estimates or interval limits, so precision cannot be assessed from the posted record.
Why does the intrapatient design matter?
In an intrapatient comparison, the two imaging approaches are assessed within the same participant. As a general statistical principle, readings from the same lesion are related rather than independent. A suitable comparison should account for that pairing. For example, the pattern of disagreements between the two methods can be important: cases where one method is correct and the other is not carry information about their relative performance. The registry record does not state which statistical procedure was used for the comparison, so it is not possible to evaluate whether or how pairing was handled.
The registry identifies the lesion as the unit of analysis and says that each participant had a single lesion to characterize. For this primary endpoint, the lesion and participant are consequently equivalent units. The record names three off-site assessors, blinded to patient data, but does not describe whether estimates were calculated separately for each assessor, combined across assessors, or otherwise summarized. Nor does it report whether variation between assessors was included in the analysis.
How is the truth standard used?
Sensitivity is calculated relative to a reference classification. The registry lists contrast-enhanced computed tomography, contrast-enhanced magnetic resonance imaging, tissue pathology or histology from surgery or biopsy, and six-month follow-up as possible truth-standard sources. The method used to establish malignancy affects the denominator of reference-positive lesions and the identification of true-positive ultrasound classifications. In general, variation in reference procedures or uncertainty in reference classifications can affect diagnostic-performance estimates. The registry record does not state how the listed sources were assigned, combined, or adjudicated when they differed.
What is the role of the analysis population?
The primary description identifies an Intent-to-Diagnose population of 240 participants. Sensitivity was restricted to the 124 participants whose lesions were malignant according to the truth standard. This restriction follows the endpoint’s definition: lesions not classified as malignant by the reference do not enter the sensitivity denominator. It does mean that the sensitivity result, if reported, would answer a narrower question than overall diagnostic accuracy in all participants. The record does not explain how incomplete ultrasound assessments, unavailable reference results, or other missing measurements were handled.
Why might variation exceed a simple model’s assumptions?
As a general concept, over-dispersion occurs when observed variation is greater than a statistical model assumes. In diagnostic studies, clustering of readings by assessor, relationships among repeated assessments, or heterogeneous participants can make observations more variable than an independence-based analysis anticipates. Accounting for such structure can affect uncertainty estimates. The registry record does not report whether an over-dispersion assessment was performed or whether a model that allows extra variation was used. It would be inappropriate to assume a particular model from the endpoint description alone.
What can be said about the analysis plan?
The registry lists eight outcome measures and 15 analyses, and the primary endpoint specifies a 95% confidence interval. However, the detailed statistical methods, formal comparison procedure, superiority criterion, and handling of assessor results are not reported in the registry record. General descriptions of sensitivity or paired-analysis principles explain how such data may be interpreted, but they do not establish which procedures were actually used in this study.
10. Multiplicity, Missing Data and Analysis Populations
The registry lists eight outcome measures and 15 analyses. The named secondary outcomes include accuracy, positive and negative predictive values, specific diagnosis of malignant and benign focal liver lesions, and inter-reader agreement. The record does not report whether statistical adjustment was used to account for multiple outcomes, multiple analyses, or results across the three assessors. Without that information, the risk of finding apparently notable results through repeated testing cannot be evaluated. No conclusion about statistical significance across the overall set of analyses can be drawn from the outcome list alone.
The record identifies an Intent-to-Diagnose population of 240 participants and a malignant-lesion sensitivity subset of 124 participants. It does not describe how missing ultrasound readings, incomplete truth-standard information, withdrawals, or other unavailable measurements were handled. It also does not report whether exclusions from the analysis populations occurred or the reasons for them. These details matter because missing or excluded cases can affect which lesions contribute to estimates and how representative those estimates are.
The registry states that the three off-site assessors were blinded to patient data. It does not report how assessor-level results were combined, whether each assessor was analyzed separately, or how disagreements between assessors were handled. The statistical analysis plan, any prespecified hierarchy among outcomes, and any multiplicity procedure are not reported in the registry record. These omissions limit assessment of how selectively reported results may relate to the full set of planned analyses.
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 |
|---|---|---|---|
| Safety Population | 5 | 337 | 1.5% |
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 |
|---|---|
| Safety Population | 54 of 337 |
Most frequent serious adverse event terms
| Event term | Organ system | Safety Population |
|---|---|---|
| Metabolic acidosis | Metabolism and nutrition disorders | 1 of 337 |
| Dehydration | Metabolism and nutrition disorders | 1 of 337 |
| Orthostatic hypotension | Vascular disorders | 1 of 337 |
| Presyncope | Nervous system disorders | 1 of 337 |
| Haemothorax | Respiratory, thoracic and mediastinal disorders | 1 of 337 |
| Pulmonary embolism | Respiratory, thoracic and mediastinal disorders | 1 of 337 |
Most frequent non-serious adverse event terms
| Event term | Organ system | Safety Population |
|---|---|---|
| Dysgeusia | Nervous system disorders | 10 of 337 |
| Headache | Nervous system disorders | 10 of 337 |
| Nausea | Gastrointestinal disorders | 9 of 337 |
| Abdominal pain | Gastrointestinal disorders | 4 of 337 |
| Injection site pain | General disorders | 4 of 337 |
| Diarrhoea | Gastrointestinal disorders | 3 of 337 |
| Vomiting | Gastrointestinal disorders | 3 of 337 |
| Abdominal discomfort | Gastrointestinal disorders | 2 of 337 |
| Feeling hot | General disorders | 2 of 337 |
| Pain in extremity | Musculoskeletal and connective tissue disorders | 2 of 337 |
Serious events with the highest arm-level rate: metabolic acidosis (0.3%), dehydration (0.3%), orthostatic hypotension (0.3%), presyncope (0.3%), haemothorax (0.3%).
The registry reports a safety population of 337 participants. Serious adverse events were reported for 5 participants, and other adverse events were reported for 54 participants. These are separate event categories in the record; the figures alone do not show whether participants with serious events also had other events, nor do they describe the number of events per participant.
The listed serious adverse event terms are metabolic acidosis, dehydration, orthostatic hypotension, presyncope, haemothorax, and pulmonary embolism. Each term is reported as 0.3%. The registry record does not give further detail about timing, severity beyond the serious-event classification, relatedness to the imaging procedures, or clinical course. It also does not report deaths; the death field is null, which should not be interpreted as evidence that no deaths occurred.
These safety figures are descriptive counts from the safety population. They do not establish that an imaging approach caused an event, and the record does not provide a randomized control comparison for attributing differences in adverse events. Because the study is an intrapatient diagnostic comparison rather than a randomized comparison between separate groups, the reported counts should be read as safety observations in the recorded population, not as comparative risk estimates.
12. What the Result Does, and Does Not, Mean
The registry record establishes that the study was designed to compare SonoVue-enhanced ultrasound with unenhanced ultrasound for focal liver lesion characterization, using sensitivity for malignant lesions as the primary endpoint. It also identifies the analysis population, the truth-standard options, the assessor arrangement, and the endpoint confidence-interval specification. These details clarify the intended diagnostic question, but they do not reveal the numerical answer.
The primary sensitivity estimates and their 95% confidence intervals are not reported in the registry record. Consequently, the posted information does not establish which approach had the higher observed sensitivity, how large any difference was, or how precisely sensitivity was estimated. The stated purpose of demonstrating superiority does not substitute for the missing outcome estimates and comparison results.
Even if sensitivity results were available, sensitivity alone would not describe all aspects of diagnostic performance. It addresses malignant lesions identified by the truth standard; it does not alone quantify false positives among non-malignant lesions or overall agreement with the reference. The registry names additional outcomes, including accuracy, predictive values, and inter-reader agreement, but their numerical results are not reported in the information.
The study’s intrapatient structure is relevant to interpreting any comparison because readings from the same lesion are related. The record does not specify the statistical method used to account for that structure, how the three assessors’ readings were handled, or how the truth-standard sources were combined. The posted record also does not detail missing-data handling or multiplicity procedures. These gaps prevent a full evaluation of precision, analysis assumptions, and the robustness of any claimed comparison.
The safety counts describe reported events in the safety population but do not establish causation or comparative risk. Overall, the registry provides a design and endpoint framework, while leaving the primary comparative result and several key analytical details unreported.
13. Why This Trial Matters Statistically
This study raises several statistical issues that are important in diagnostic research. First, its intrapatient comparison means that the two imaging assessments concern the same lesion. As a general principle, such paired data contain information in the relationship between readings, not only in the separate performance of each method. The analysis should reflect that structure, but the registry record does not report the method used.
Second, the endpoint is defined relative to a truth standard assembled from possible imaging, pathology or histology, and follow-up sources. Diagnostic accuracy estimates depend on the reference classification: it determines which lesions count as malignant and whether an ultrasound result is considered correct. Understanding how reference information was assigned and reconciled is therefore important, but that operational detail is not reported in the registry record.
Third, sensitivity is only one part of diagnostic performance. It focuses on reference-positive lesions and does not describe false-positive classifications among reference-negative lesions. The registry names other diagnostic outcomes, including accuracy and predictive values, but their numerical findings are not reported in the information. This makes it especially important not to treat the primary endpoint definition as a complete account of classification performance.
Finally, three assessors and multiple listed outcomes create analytical questions about assessor variation and multiplicity. The record identifies blinded off-site assessors and lists eight outcome measures and 15 analyses, but does not state how assessor readings were summarized or whether multiplicity was addressed. The absence of the primary sensitivity estimates and confidence intervals also prevents assessment of the size and precision of the central result. These limitations illustrate why a clear endpoint and design description are necessary but not sufficient for interpreting a diagnostic comparison.
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 registry describes the study as intrapatient and comparative, while also recording a single-group model, allocation as not applicable, and masking as none. It is not a randomized comparison between separate groups. This design is appropriate to describe as a within-participant diagnostic comparison, but it should not be interpreted as a randomized trial with independent treatment arms.
The primary endpoint is measured against a truth standard that may consist of different sources: contrast-enhanced computed tomography, contrast-enhanced magnetic resonance imaging, pathology or histology, or six-month follow-up. The registry does not explain which source applied to each participant, how sources were combined, or how disagreements were resolved. Differences in reference classification could affect the sensitivity denominator and the designation of true positives.
The primary endpoint uses diagnoses from three off-site assessors blinded to patient data. The record does not explain whether each assessor’s performance was reported separately, how assessors’ readings were combined, or whether reader variability was incorporated into uncertainty estimates. The method for comparing the paired ultrasound assessments is also not reported.
The registry identifies 240 Intent-to-Diagnose participants and 124 malignant lesions in the sensitivity analysis, but does not describe missing-data handling, exclusions, or the completeness of reference-standard assessments. It lists eight outcome measures and 15 analyses without reporting multiplicity procedures. Most importantly, the primary sensitivity estimates and confidence intervals are not reported in the registry record. These gaps limit assessment of the result’s magnitude, precision, robustness, and reproducibility.
15. Related Tutorials
16. Related Statistical Calculators
17. Sources
- ClinicalTrials.gov: SonoVue®-Enhanced Ultrasound Versus Unenhanced US for Focal Liver Lesion Characterization — NCT00788697.
- Registry record last updated 2017-12-12; status last verified 2017-11.
- Results first posted 2017-06-14.
- 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
NCT00788697 is a completed Phase III intrapatient comparative study of SonoVue-enhanced versus unenhanced ultrasound for focal liver lesion characterization. The registry lists 349 enrolled participants, with the study beginning in 2009-09 and primary completion in 2013-07. It records a single-group model, allocation as not applicable, and masking as none.
The primary endpoint was sensitivity for malignant lesions against a truth standard based on imaging, pathology or histology, or follow-up. The Intent-to-Diagnose population included 240 participants, of whom 124 had lesions classified as malignant by the truth standard and entered the sensitivity analysis. Diagnoses from three off-site assessors blinded to patient data were specified.
The registry record does not report the primary sensitivity estimates or confidence intervals, the statistical method for comparing paired readings, or detailed procedures for missing data and multiplicity. In the safety population of 337 participants, 5 had serious adverse events and 54 had other adverse events. Deaths are not reported.