Author: Njood Alsaihati, Ehsan Samei, Justin B. Solomon 👨🔬
Affiliation: Duke University, Clinical Imaging Physics Group, Department of Radiology, Duke University Health System 🌍
Purpose: Inconsistent imaging procedure naming across and within institutions challenges clinical decision-making, quality assessment, and data analysis in radiology. For example, The American College of Radiology (ACR) CT Dose Index Registry (DIR) requires manual or semi-manual mapping of institutional study descriptions to RadLex Playbook IDs, a process prone to ambiguity and errors, impacting data consistency for analysis. The purpose of this study was to develop a comprehensive data ontology framework and provide concrete demonstrations of its application for efficient CT data profiling and characterization.
Methods: A systematic review of existing literature and methodologies related to standardization in CT imaging informed framework development. The review identified key principles and gaps in nomenclature consistency and metadata standardization. The framework defines tags across study levels, irradiation events (scans), and series. These tags capture essential aspects of CT imaging workflows, including patient demographics, ordered and performed protocols, resulting interpretation, imaging environment, pharmaceutical usage, acquisition parameters, and reconstruction settings. Key examples were identified where the ontology framework offers practical benefits and impact.
Results: The data ontology framework was developed using generalizable tags, demonstrating its significant potential through two key areas. First, it enhances systems like the ACR CT DIR by reducing ambiguity in mapping and improving data consistency for benchmarking and quality assessment. Second, it addresses challenges with The Centers for Medicare & Medicaid Services (CMS) new quality measure, “Excessive Radiation Dose or Inadequate Image Quality for Diagnostic Computed Tomography (CT) in Adults.” By enabling efficient CT data profiling, the framework ensures data relevance, simplifies the complexity of CT data, and protects institutions from penalties for clinically justified personalized care, while ensuring compliance.
Conclusion: The data ontology framework addresses critical challenges in CT imaging, enabling targeted data analysis. Its applications highlight its potential to improve quality assessment, compliance, and data reliability in radiology.