Work

Revisiting artifact spread function evaluation in digital breast tomosynthesis systems

Year: 2025

Type: article

Abstract: Background Digital Breast Tomosynthesis (DBT) has proved to be a powerful tool in breast cancer detection. However, the presence of zipper artifacts poses challenges in accurate diagnosis and there is not global methodology to evaluate its effects. Purpose The purpose of this study is to determine the most suitable method to characterize zipper artifact using the Full Width at Half Maximum (FWHM) of the Artifact Spread Function (ASF) as figure of merit and the adequacy Al sphere phantom as zipper artifact generators. Methods The study compares the zipper artifact on General Electric Senographe Pristine (SP-GE) and Hologic Selenia Dimensions (SD-HO) by calculating ASF in three different ways: maximum value method, mean value method and a hybrid method proposed by the authors. Besides, the ASF generated by 1 mm aluminum sphere phantom was compared to those generated by 1 mm lead and tungsten markers, 2.3 mm ceramic markers and Mammo 156 TM Phantom from Gammex. Results For the SP-GE system, the ASF calculated with the hybrid method resulted the most suitable approach in terms of reliability. In contrast, for the SD-HO system, the mean value method aligns better with system characteristics. The study underscores the importance of considering system-specific nuances in methodology selection for artifact evaluation. The data indicates that, for most of the methods, ceramic and lead (Pb) markers replicates better the artifacts observed in clinical practice. Conclusions The article emphasizes the need for standardized methodologies that can accommodate system-specific differences while ensuring accurate artifact assessment.

Source: Physica Medica

DOI: 10.1016/j.ejmp.2025.104982