BLUE RIBBON POSTER IMAGING: A Clinical Analytics Platform for Radiographic Protocol Optimization
Abstract
Purpose
In digital radiography (DR), the Exposure Index (EI) and Deviation Index (DI) provide standardized feedback on whether image receptor exposure is appropriately matched to patient size and clinical task. However, systematic, context-aware analysis of EI and DI across exams may not be easily available to a practice or provided in a way that is not helpful for clinical practice quality improvement. This work describes the development and application of a clinical analytics platform to evaluate DR acquisition techniques and support Image Wisely–aligned protocol optimization using DICOM metadata.
Methods
Clinical DR DICOM metadata were extracted using a Flywheel-based analytics pipeline and analyzed within onboard Jupyter Notebooks. EI and DI distributions were evaluated relative to practice-defined target values using comparable acquisition contexts, including vendor/system, exam type, and projection view. Summary statistics, including median DI, DI standard deviation, median EI, and corresponding EI target values, were analyzed for each exam–view combination, with sample sizes recorded for each group.
Results
Context-stratified analysis demonstrated substantial variability in receptor exposure across common exams and views. Several exam–view combinations exhibited high-magnitude median DI values and wide DI distributions, indicating systematic deviation from exposure targets and a need for protocol review or exposure target adjustment. At a single clinical site, pelvic examinations were acquired predominantly using manual techniques (≈98%) and exhibited consistent positive DI bias (median DI ≈ +1.4 to +2.9 across protocols), further supporting the need for protocol intervention and technologist education.
Conclusion
This clinical informatics platform enables automated, context-aware analysis of EI and DI using routinely available DR metadata. The approach supports data-driven protocol optimization, targeted technologist feedback, and practice-wide quality improvement. Embedding this form of analytics into routine operations may improve protocol optimization and standardization, for patient-tailored dose optimization aligned with Image Wisely principles, and improved consistency in image quality.