Assessment of task-based performance from five clinical DBT systems using an anthropomorphic breast phantom

Area of Science:

• Medical Imaging
• Radiology
• Biomedical Engineering

Background:

• Digital breast tomosynthesis (DBT) is a crucial breast cancer screening tool, often used with full-field digital mammography (FFDM) or synthetic mammography (SM).
• Existing FDA-approved DBT systems vary significantly in design and imaging protocols, necessitating a standardized quantitative assessment method.
• The technical differences among DBT systems hinder direct comparison and performance evaluation.

Purpose of the Study:

• To introduce and validate a novel quantitative methodology for assessing the performance of commercial digital breast tomosynthesis (DBT) systems.
• To compare the efficacy of DBT against FFDM and synthetic mammography (SM) using a standardized phantom and reader study.
• To establish a benchmark for evaluating diverse DBT system designs and imaging protocols.

Main Methods:

• Fabrication of an anthropomorphic breast phantom using inkjet printing with realistic simulated masses (spiculated, KI-doped ink) and microcalcifications (calcium hydroxyapatite).
• Acquisition of phantom images using five commercial DBT systems, FFDM, and SM modalities under automatic exposure control.
• Conducting a four-alternative forced choice (4AFC) reader study to assess performance in detecting simulated masses and microcalcifications, followed by statistical analysis of proportion correct (PC).

Main Results:

• Digital breast tomosynthesis (DBT) exhibited the highest overall detection rates for simulated masses, with statistically significant improvements over SM for most systems.
• A positive correlation was observed between increased reader performance (PC) and greater gantry span across DBT systems.
• DBT and FFDM showed superior performance in detecting microcalcifications compared to SM, with statistically significant differences across all manufacturers.

Conclusions:

• The developed methodology, utilizing an inkjet-printed anthropomorphic phantom, offers a novel and effective approach for quantitatively evaluating clinical DBT imaging systems.
• This study is the first to employ such a realistic phantom for assessing the performance of multiple commercial DBT systems.
• The findings highlight DBT's advantages in detecting both masses and microcalcifications, providing valuable insights for system selection and development.