Prediction of reader estimates of mammographic density using convolutional neural networks
Prediction of reader estimates of mammographic density using convolutional neural networks
复制标题
DOI：
10.1117/1.jmi.6.3.031405
复制
发表时间：
2019-07-01
影响因子：
2.4
通讯作者：
Astley, Susan M.
中科院分区：
文献类型：
--
作者：
Ionescu, Georgia, V;Fergie, Martin;Astley, Susan M.
乳房X线密度是乳腺癌的重要危险因素。在最近的研究中，使用视觉模拟量表（VAS）进行视觉评估的百分比密度显示出比现有的自动密度测量更强的风险预测，这表明读者可能会识别手工算法尚未捕获的相关图像特征。通过深度学习，可以将这些知识封装在自动方法中。我们建立了卷积神经网络 (CNN) 来预测全视野数字乳房 X 光检查的密度 VAS 分数。 CNN 使用全图像乳房 X 光照片进行训练，每个图像都标有两个独立读者的平均 VAS 分数。每个 CNN 都会学习乳房 X 线摄影外观和 VAS 评分之间的映射，以便在测试时，它们可以预测未见过的图像的 VAS 评分。使用来自 16,968 名女性的 67,520 张乳房 X 线摄影图像对网络进行训练，并使用包含 73,128 张图像的数据集进行模型选择。两组病例对照的筛查检测到癌症的对侧乳房 X 光检查和随后检测到的患有癌症的女性的先前图像，与年龄、绝经状态、产次、HRT 和 BMI 的对照相匹配，用于评估乳腺癌预测的表现。在病例对照组中，筛查检测出的癌症中最高与最低五分位的癌症奇数比为 2.49（95% CI：1.59 至 3.96），既往癌症的奇数比为 4.16（2.53 至 6.82），匹配的一致性指数为 0.587（0.542 至 0.627）和 0.616（0.578）至 0.655）。先前测试集的读者 VAS 和预测 VAS 之间没有显着差异（似然比卡方，p = 0.134）。我们的全自动方法在癌症风险预测方面显示出有希望的结果，并且与人类的表现相当。 (C) 作者。由 SPIE 根据 Creative Commons Attribution 4.0 Unported 许可证发布。
Mammographic density is an important risk factor for breast cancer. In recent research, percentage density assessed visually using visual analogue scales (VAS) showed stronger risk prediction than existing automated density measures, suggesting readers may recognize relevant image features not yet captured by hand-crafted algorithms. With deep learning, it may be possible to encapsulate this knowledge in an automatic method. We have built convolutional neural networks (CNN) to predict density VAS scores from full-field digital mammograms. The CNNs are trained using whole-image mammograms, each labeled with the average VAS score of two independent readers. Each CNN learns a mapping between mammographic appearance and VAS score so that at test time, they can predict VAS score for an unseen image. Networks were trained using 67,520 mammographic images from 16,968 women and for model selection we used a dataset of 73,128 images. Two case-control sets of contralateral mammograms of screen detected cancers and prior images of women with cancers detected subsequently, matched to controls on age, menopausal status, parity, HRT and BMI, were used for evaluating performance on breast cancer prediction. In the case-control sets, odd ratios of cancer in the highest versus lowest quintile of percentage density were 2.49 (95% CI: 1.59 to 3.96) for screen-detected cancers and 4.16 (2.53 to 6.82) for priors, with matched concordance indices of 0.587 (0.542 to 0.627) and 0.616 (0.578 to 0.655), respectively. There was no significant difference between reader VAS and predicted VAS for the prior test set (likelihood ratio chi square, p = 0.134). Our fully automated method shows promising results for cancer risk prediction and is comparable with human performance. (C) The Authors. Published by SPIE under a Creative Commons Attribution 4.0 Unported License.