Systematic Overexposure In Pediatric Portable Radiography: A Multi-Modality Analysis Using Deviation Index and Physics-Based Mas Correction for Department-Wide Dose Optimization
Abstract
Purpose
To establish evidence-based mAs targets for pediatric portable radiography dose optimization through reverse engineering from Deviation Index (DI) data, implement targeted technologist training, and benchmark against international Diagnostic Reference Levels (DRLs).
Methods
Retrospective analysis of 5,179 pediatric portable exams (chest: 4,772; abdomen: 407) from 15 X-ray units. Unit-level medians were computed by weight group. Ideal mAs targets achieving DI=0 were calculated using mAs (ideal) = mAs (current) x 10(-DI/10). A 7-item EI/DI knowledge survey was administered to 36 technologists. Two-track training was implemented with post-training directives to reduce mAs and monitor DI. Calculated AD/DRL values were compared against published European and UK DRLs.
Results
All 15 AETs showed systematic overexposure (DI range: chest 4.3-7.6, abdomen 4.8-10.4). Less than 4% of exams achieved target DI of 0 compared to expected 68%. Reverse-engineered mAs targets require 59-84% reduction: chest 0.21-0.86 mAs, abdomen 0.22-12.9 mAs. Technologist survey revealed mean score 3.9/7 (55%) with critical gaps in DI interpretation (42-47% correct). Training was implemented with directives to reduce mAs, monitor DI with alert threshold greater than 2, and standardize Target EI. Our DRLs were 1.5-5.8 times higher than published European benchmarks, with highest discrepancies in less than 15 kg patients (2-6 times above DRLs).
Conclusion
This study provides actionable mAs targets for 50-85% dose reduction, identifies technologist knowledge gaps requiring two-track training, and demonstrates substantial optimization opportunity through international benchmarking. Implementation of physics-based mAs corrections with real-time DI monitoring offers an immediate pathway to reduce pediatric radiation exposure while maintaining diagnostic quality.