Author: James Ward, Anzi Zhao 👨🔬
Affiliation: Northwestern Medicine 🌍
Purpose: This study investigates the impact of tissue-equivalent attenuator materials, the region of interest (ROI) size, and the orientation of the attenuator positioning on the measured signal-to-noise ratio (SNR) when evaluating the performance of automatic exposure control (AEC) in digital mammography.
Methods: This study evaluated nine digital mammography systems with comparable output from the same manufacturer, using BR12 and PMMA attenuators at thicknesses of 2 cm, 4 cm, 6 cm, and 8 cm in 2D, magnification, and digital breast tomosynthesis (DBT) imaging modes. BR12 attenuators were positioned in portrait and landscape orientations. The evaluation followed the 2018 ACR Digital Mammography Quality Control (DMQC) manual. SNR was analyzed on for processing images using three square ROI sizes: 64, 128, and 256 pixels, with baseline SNR values established using a 1.0 cm² ROI on portrait-orientated BR12 attenuator images.
Results: The SNR varied significantly by the attenuator material in 2D magnification mode at 4cm and in DBT mode at 6 cm and 8 cm, with deviation from baseline ranging 22%-38%, 28%-81%, 3%-72%, respectively. In DBT mode, setting the ROI size to 256² increased average noise by 59%, 20%, 38%, and 80% across four thicknesses, leading to SNR decreased by 37%, 16%, 27%, and 44%, respectively. The BR12 attenuator, particularly at 8 cm in DBT mode and when positioned in landscape orientation, demonstrated an increase in SNR by 20%-43% across nine systems. These SNR deviations may lead to non-compliance with the AEC performance criteria outlined in the ACR DMQC manual, despite all units meeting the manufacturer's standard.
Conclusion: The selection of attenuators materials, the size of ROI, and the orientation of attenuator positioning could significantly impact the measured SNR, leading to erroneous failure. It is essential for medical physicists to maintain awareness and consistency during the longitudinal evaluation of AEC performance in digital mammography.