International Delphi Study to Develop Standardized Quality Criteria and a Rating Tool for Liver CEUS Image Quality Assessment (ICON LiverCEUS)

Contrast-enhanced ultrasound (CEUS) is an established and widely used imaging modality for the evaluation of focal liver lesions and other hepatic abnormalities. It plays an increasingly important role in clinical diagnostics due to its real-time imaging capability, safety profile, and broad availability. Despite its clinical relevance, there is currently no internationally standardised definition of adequate CEUS image quality, and no validated criteria exist for its objective assessment. As a result, interpretation and perceived diagnostic quality of CEUS examinations may vary considerably between operators, institutions, and healthcare systems.

The aim of this study is to develop internationally agreed, consensus-based quality criteria for adequate CEUS image quality in liver examinations and to translate these criteria into a standardised rating tool. The overall objective is to improve the reproducibility, comparability, and standardisation of CEUS examinations in both clinical practice and research settings.

To achieve this, the study is structured into two work packages (WP). Work Package 1 (WP1) focuses on identifying and refining relevant quality criteria for CEUS of the adult liver through an international Delphi consensus process.

Work Package 2 (WP2) focuses on clustering the agreed criteria into overarching domains and developing a structured rating scale for practical application.

In WP1, a preliminary set of potential quality indicators will be generated based on a structured literature review and expert discussion within the study team. These indicators will reflect technical, procedural, and image-related aspects of liver CEUS examinations. The most commonly applied CEUS context, adult liver imaging, will serve as the clinical focus of the study. An international panel of CEUS experts will be recruited based on defined expertise criteria, including extensive clinical experience in ultrasound and CEUS, training involvement, and scientific contribution to the field. Approximately 35 experts will be invited to participate in the first Delphi round, with an expected final panel size of at least 20 participants across all rounds.

The Delphi process will consist of three iterative survey rounds. In each round, participants will rate their agreement with predefined statements describing potential quality criteria using a 7-point Likert scale ranging from "completely disagree" to "completely agree," with an additional option for "outside of my expertise." Participants may also suggest additional criteria in the first round. Free-text responses will be systematically analysed, categorised, and included in subsequent rounds if supported by at least 10% of participants.

After each round, anonymised aggregated feedback will be provided to participants, including summary statistics such as median scores, interquartile ranges, and agreement rates. Participants will then be asked to reconsider and, if appropriate, revise their ratings in light of group responses. Consensus will be defined a priori as at least 70% of panellists rating a criterion as "agree" or "completely agree." Criteria not reaching consensus after repeated rounds will be revised or excluded according to predefined rules.

WP2 will translate the final consensus-based criteria into a structured rating scale for assessing CEUS image quality. The agreed criteria will be organised into overarching domains reflecting key dimensions of image quality. Within each domain, individual criteria will serve as structured prompts to guide evaluation. The resulting scale will be designed to be practical and reproducible and will use a simplified scoring system for domain-level assessment. Depending on the Delphi findings, non-imaging or technical aspects may be included as a separate domain.

The draft rating scale will be reviewed by the Delphi panel participants who completed all rounds. They will be asked to evaluate agreement with each domain and with the overall structure of the scale using a structured binary response format, with optional qualitative feedback. If predefined acceptance thresholds are not met, an additional revision round will be conducted.

Data analysis will include descriptive statistics for each criterion, including median and interquartile range, as well as calculation of agreement proportions. Responses indicating "outside of my expertise" will be treated as missing data and excluded from statistical calculations. Criteria will be evaluated against predefined consensus thresholds to determine inclusion, revision, or exclusion across Delphi rounds.

The final outcome of this study will be an internationally derived consensus set of quality criteria for liver CEUS examinations and a structured, standardised rating tool for assessing CEUS image quality. This framework is intended to support improved quality assurance, enhance reproducibility, and facilitate harmonisation of CEUS practice across institutions and countries in both clinical and research contexts.