Uncertainty Contributors In Ctdi Benchmarking
Abstract
Purpose
Computed Tomography Dose Index (CTDI) is widely used for scanner benchmarking, quality assurance, and protocol optimization. Although CTDI methodology is standardized, practical implementation choices may introduce uncertainty. This study evaluates two often-overlooked contributors to CTDI uncertainty: phantom placement geometry and peripheral sampling strategy.
Methods
CTDI measurements were performed using a 16-cm head phantom and a 32-cm body phantom on multiple clinical CT systems. Adult head protocols were measured with the phantom positioned directly on the patient table and within the manufacturer-supplied head holder. Peripheral CTDI measurements were obtained at the 12, 3, 6, and 9 o’clock positions, along with center measurements, to calculate CTDIw and CTDIvol. Additional datasets assessed the impact of using individual peripheral positions versus the average of all four peripheral positions for CTDIw determination in head and abdomen protocols. Calculated CTDIvol values were compared with scanner-reported CTDIvol.
Results
Phantom placement significantly affected measured dose. Compared with head-holder positioning, on-table placement underestimated CTDIw/CTDIvol and DLP by approximately 5%, primarily due to reduced peripheral dose at the 6 o’clock position. On-table placement also increased angular dose non-uniformity, with the coefficient of variation across peripheral positions increasing from approximately 3% to nearly 7%. Evaluation of peripheral sampling strategy demonstrated systematic angular bias: the 6 o’clock position consistently underestimated average peripheral CTDI, while the 12 o’clock position tended to overestimate it. No single peripheral position reliably represented the average across scanners or protocols.
Conclusion
CTDI benchmarking is sensitive to both phantom placement geometry and peripheral sampling strategy. Simplified measurement approaches can introduce combined uncertainties exceeding 10%, even under otherwise standardized conditions. Standardized phantom positioning and full peripheral averaging are essential for accurate and reproducible CTDI measurements. Currently, we are investigating the uncertainty related to variability in x-ray tube start and end angular positions during repeated CTDI measurements.