PGMI-BENCH dataset: Automated PGMI scores and breast imaging parameters from digital breast tomosynthesis screening

This dataset contains image-level data from the PGMI-BENCH study, a retrospective analysis of digital breast tomosynthesis (DBT) screening examinations evaluating automated image quality assessment systems. The dataset includes 778 recalled screening examinations with paired cranio-caudal and mediolateral oblique views processed by two automated PGMI-based software tools (SW1 and SW2). For each view, the dataset reports PGMI scores assigned by both systems, along with quantitative imaging parameters describing breast anatomy and acquisition characteristics. Variables include breast thickness, compression pressure, compression force, volumetric breast density (VBD%), breast volume, and categorical breast density estimates (SW1-derived). The dataset enables analysis of agreement between automated image quality assessment systems and their associations with anatomical and technical parameters. This dataset supports reproducibility of the PGMI-BENCH study and may be used for benchmarking, methodological research in image quality assessment, and development of automated quality control algorithms in breast imaging.