The Chinese Mammography Database (CMMD): An online mammography database with biopsy confirmed types for machine diagnosis of breast

Breast carcinoma is the second largest cancer in the world among women. Early detection of breast cancer has been shown to increase the survival rate, thereby significantly increasing patients' lifespans. Mammography, a noninvasive imaging tool with low cost, is widely used to diagnose breast disease at an early stage due to its high sensitivity. The recent popularization of artificial intelligence in computer-aided diagnosis creates opportunities for advances in areas such as (1) Computer-aided detection for locating suspect lesions such as mass and microcalcification, leaving the classification to the radiologist; and (2) Computer-aided diagnosis for characterizing the suspicious region of lesion and/or estimate its probability of onset; and (3) Findings of predictive image-based biomarkers by applying the computational methods to mine the potential relationships between image representation and molecular subtype, including Luminal A, Luminal B, HER2 positive, and Triple-negative.

However, existing publicly available mammography databases are limited by small sample size, lack of diversity in patient populations, missing biopsy confirmations and unknown molecular sub-types. To help fill the gap, we built a database conducted on 1,775 patients from China with benign or malignant breast disease who underwent mammography examination between July 2012 and January 2016. The database consists of 3,728 mammographies from these 1,775 patients, with biopsy confirmed type of benign or malignant tumors. For 749 of these patients (1,498 mammographies) we also include patients' molecular subtypes. Image data were acquired on a GE Senographe DS mammography system.

乳腺癌是全球女性群体中第二大高发恶性肿瘤。早期筛查乳腺癌已被证实可提升患者生存率，进而显著延长其生存期。乳腺钼靶摄影（Mammography）作为一种低成本无创成像手段，因具备较高的诊断灵敏度，被广泛应用于乳腺疾病的早期诊断。近年来人工智能在计算机辅助诊断领域的普及，为诸多方向的技术进步提供了契机，包括：(1) 计算机辅助检测（Computer-aided detection）：用于定位肿块、微钙化等可疑病灶，后续分类工作由放射科医师完成；(2) 计算机辅助诊断（Computer-aided diagnosis）：对可疑病变区域进行特征表征，并/或估算其发病概率；(3) 通过计算方法挖掘影像表征与分子亚型之间的潜在关联，从而获取基于影像的预测性生物标志物，其中分子亚型包括管腔A型（Luminal A）、管腔B型（Luminal B）、人表皮生长因子受体2阳性型（HER2 positive）以及三阴性（Triple-negative）。然而，当前公开可用的乳腺钼靶数据库普遍存在样本量偏小、患者群体多样性不足、缺乏活检验证结果以及分子亚型信息缺失等局限性。为填补这一研究空白，我们构建了一款乳腺钼靶数据库：该数据集纳入了2012年7月至2016年1月期间接受钼靶检查的1775名中国乳腺良恶性疾病患者。该数据库共包含这1775名患者的3728例乳腺钼靶影像，所有病例均经活检证实为良性或恶性肿瘤。其中749名患者（对应1498例钼靶影像）的分子亚型信息也已纳入数据集。所有影像数据均采用GE Senographe DS乳腺钼靶摄影系统采集。