Bermudez Camilo, Blaber Justin, Remedios Samuel W, Reynolds Jess E, Lebel Catherine, McHugo Maureen, Heckers Stephan, Huo Yuankai, Landman Bennett A
Department of Biomedical Engineering, Vanderbilt University, 2201 West End Ave, Nashville, TN, USA 37235.
Department of Electrical Engineering, Vanderbilt University, 2201 West End Ave, Nashville, TN, USA 37235.
Proc SPIE Int Soc Opt Eng. 2020;11313. Epub 2020 Mar 10.
Generalizability is an important problem in deep neural networks, especially in the context of the variability of data acquisition in clinical magnetic resonance imaging (MRI). Recently, the Spatially Localized Atlas Network Tiles (SLANT) approach has been shown to effectively segment whole brain non-contrast T1w MRI with 132 volumetric labels. Enhancing generalizability of SLANT would enable broader application of volumetric assessment in multi-site studies. Transfer learning (TL) is commonly to update neural network weights for local factors; yet, it is commonly recognized to risk degradation of performance on the original validation/test cohorts. Here, we explore TL by data augmentation to address these concerns in the context of adapting SLANT to anatomical variation (e.g., adults versus children) and scanning protocol (e.g., non-contrast research T1w MRI versus contrast-enhanced clinical T1w MRI). We consider two datasets: First, 30 T1w MRI of young children with manually corrected volumetric labels, and accuracy of automated segmentation defined relative to the manually provided truth. Second, 36 paired datasets of pre- and post-contrast clinically acquired T1w MRI, and accuracy of the post-contrast segmentations assessed relative to the pre-contrast automated assessment. For both studies, we augment the original TL step of SLANT with either only the new data or with both original and new data. Over baseline SLANT, both approaches yielded significantly improved performance (pediatric: 0.89 vs. 0.82 DSC, p<0.001; contrast: 0.80 vs 0.76, p<0.001). The performance on the original test set decreased with the new-data only transfer learning approach, so data augmentation was superior to strict transfer learning.
可推广性是深度神经网络中的一个重要问题，尤其是在临床磁共振成像（MRI）数据采集存在变异性的背景下。最近，空间局部化图谱网络切片（SLANT）方法已被证明能够有效地分割具有132个体积标签的全脑非对比T1加权MRI。提高SLANT的可推广性将使体积评估在多中心研究中得到更广泛的应用。迁移学习（TL）通常用于根据局部因素更新神经网络权重；然而，人们普遍认识到这存在使原始验证/测试队列性能下降的风险。在这里，我们通过数据增强来探索迁移学习，以在使SLANT适应解剖变异（例如成人与儿童）和扫描协议（例如非对比研究T1加权MRI与对比增强临床T1加权MRI）的背景下解决这些问题。我们考虑两个数据集：第一，30例幼儿的T1加权MRI，带有手动校正的体积标签，并根据手动提供的真值定义自动分割的准确性。第二，36对临床采集的对比前和对比后T1加权MRI数据集，并根据对比前的自动评估来评估对比后分割的准确性。对于这两项研究，我们用新数据或同时用原始数据和新数据来增强SLANT的原始迁移学习步骤。与基线SLANT相比，两种方法均产生了显著提高的性能（儿科：DSC为0.89对0.82，p<0.001；对比：0.80对0.76，p<0.001）。仅采用新数据的迁移学习方法会使原始测试集的性能下降，因此数据增强优于严格的迁移学习。
