Detection of masses and architectural distortions in digital breast tomosynthesis

本数据集名为“数字乳腺断层合成图像中肿块和结构扭曲的检测”，由杜克大学创建，包含22,032个重建的DBT体积，属于5,060名患者。数据集内容丰富，包括正常、需进一步影像检查、良性活检和癌症四个组别，由两位经验丰富的放射科医生标注。创建过程中，通过回顾性分析和标注确保数据质量。该数据集主要应用于乳腺癌筛查，旨在通过AI算法提高检测准确性和效率。

This dataset is titled "Detection of Masses and Architectural Distortions in Digital Breast Tomosynthesis Images" and was developed by Duke University. It comprises 22,032 reconstructed DBT volumes sourced from 5,060 unique patients. The dataset includes four distinct cohorts: normal cases, cases requiring additional imaging evaluation, benign biopsy-confirmed cases, and cancer cases, and was annotated by two board-certified radiologists with rich experience in breast imaging. Data quality was validated through retrospective review and annotation throughout the dataset construction process. This dataset is primarily designed for breast cancer screening applications, with the objective of improving detection accuracy and efficiency using artificial intelligence algorithms.