Deep Learning Segmentation of Triple-Negative Breast Cancer (TNBC) Patient Derived Tumor Xenograft (PDX) and Sensitivity of Radiomic Pipeline to Tumor Probability Boundary.
Deep Learning Segmentation of Triple-Negative Breast Cancer (TNBC) Patient Derived Tumor Xenograft (PDX) and Sensitivity of Radiomic Pipeline to Tumor Probability Boundary.
复制标题
DOI：
10.3390/cancers13153795
复制
发表时间：
2021-07-28
期刊：
影响因子：
5.2
通讯作者：
Shoghi KI
中科院分区：
文献类型：
--
作者：
Dutta K;Roy S;Whitehead TD;Luo J;Jha AK;Li S;Quirk JD;Shoghi KI
联合临床试验是一个新兴的研究领域，其中临床试验与相应的临床前试验相结合，以告知相应的临床试验。临床前研究有助于评估治疗效果、患者分层和设计最佳成像策略。人们对协调临床前和临床定量成像管道非常感兴趣。放射组学在临床成像中被广泛探索以预测对治疗的反应。在临床前成像中，高通量放射组学分析受到手动描绘肿瘤边界的限制，这是劳动密集型的，再现性差。我们提出的基于深度学习的系统经过训练，可以从多对比MR图像中自动分割肿瘤并提取放射组学特征。所提出的方法是高度可重复的放射组学特征具有显着的相关性。在临床前分支中部署该管道将提供高通量和可再现的放射组学分析。临床前磁共振成像（MRI）是联合临床研究管道中的关键组成部分。重要的是，MRI中肿瘤的分割是肿瘤表型和评估治疗反应的必要步骤。然而，手动分割是时间密集型的，并且遭受观察者间和观察者内的可变性以及缺乏再现性。本研究旨在开发一种自动化管道，用于使用深度学习（DL）算法从临床前T1 w和T2 w MR图像中准确定位和描绘TNBC PDX肿瘤，并评估放射组学特征对肿瘤边界的敏感性。我们测试了五种网络架构，包括U-Net，密集U-Net，Res-Net，经常性剩余UNet（R2 UNet）和密集R2 U-Net（D-R2 UNet），并将其与专家的手动划定进行了比较。为了减轻多个专家之间的偏差，应用同步真值和性能水平估计（STAPLE）算法来创建共识图。性能指标（F1-Score，Recall，Precision和AUC）用于评估网络的性能。多对比度D-R2 UNet表现最好，F1评分= 0.948;然而，所有网络的评分均在1-3%范围内。从D-R2 UNet中提取的放射组学特征与STAPLE衍生特征高度相关，67.13%的T1 w和53.15%的T2 w显示相关性ρ ≥ 0.9（p ≤ 0.05）。D-R2 UNet提取的特征表现出比STAPLE更好的再现性，发现86.71%的T1 w和69.93%的T2 w特征具有高度再现性（CCC ≥ 0.9，p ≤ 0.05）。最后，39.16%的T1 w和13.9%的T2 w特征被确定为对肿瘤边界扰动不敏感（斯皮尔曼相关性（−0.4 ≤ ρ ≤ 0.4））。我们开发了一种高度可重复的DL算法，以规避T1 w和T2 w MR图像的手动分割，并确定放射组学特征对肿瘤边界的敏感性。
Co-clinical trials are an emerging area of investigation in which a clinical trial is coupled with a corresponding preclinical trial to inform the corresponding clinical trial. The preclinical arm aids in assessing therapeutic efficacy, patient stratification, and designing optimal imaging strategies. There is much interest in harmonizing preclinical and clinical quantitative imaging pipelines. Radiomics is widely explored in clinical imaging to predict response to therapy. In preclinical imaging, high-throughput radiomic analysis is limited by manual delineation of tumor boundaries, which is labor intensive with poor reproducibility. Our proposed deep-learning-based system was trained to automatically segment tumors from multi-contrast MR images and extract radiomic features. The proposed method is highly reproducible with significant correlation in radiomic features. The deployment of this pipeline in the preclinical arm would provide high throughput and reproducible radiomic analysis. Preclinical magnetic resonance imaging (MRI) is a critical component in a co-clinical research pipeline. Importantly, segmentation of tumors in MRI is a necessary step in tumor phenotyping and assessment of response to therapy. However, manual segmentation is time-intensive and suffers from inter- and intra- observer variability and lack of reproducibility. This study aimed to develop an automated pipeline for accurate localization and delineation of TNBC PDX tumors from preclinical T1w and T2w MR images using a deep learning (DL) algorithm and to assess the sensitivity of radiomic features to tumor boundaries. We tested five network architectures including U-Net, dense U-Net, Res-Net, recurrent residual UNet (R2UNet), and dense R2U-Net (D-R2UNet), which were compared against manual delineation by experts. To mitigate bias among multiple experts, the simultaneous truth and performance level estimation (STAPLE) algorithm was applied to create consensus maps. Performance metrics (F1-Score, recall, precision, and AUC) were used to assess the performance of the networks. Multi-contrast D-R2UNet performed best with F1-score = 0.948; however, all networks scored within 1–3% of each other. Radiomic features extracted from D-R2UNet were highly corelated to STAPLE-derived features with 67.13% of T1w and 53.15% of T2w exhibiting correlation ρ ≥ 0.9 (p ≤ 0.05). D-R2UNet-extracted features exhibited better reproducibility relative to STAPLE with 86.71% of T1w and 69.93% of T2w features found to be highly reproducible (CCC ≥ 0.9, p ≤ 0.05). Finally, 39.16% T1w and 13.9% T2w features were identified as insensitive to tumor boundary perturbations (Spearman correlation (−0.4 ≤ ρ ≤ 0.4). We developed a highly reproducible DL algorithm to circumvent manual segmentation of T1w and T2w MR images and identified sensitivity of radiomic features to tumor boundaries.
登录
查看更多内容
DOI：
10.1158/1078-0432.ccr-15-1762
发表时间：
2016-04-01
期刊：
Clinical cancer research : an official journal of the American Association for Cancer Research
影响因子：
--
作者：
Krepler C;Xiao M;Sproesser K;Brafford PA;Shannan B;Beqiri M;Liu Q;Xu W;Garman B;Nathanson KL;Xu X;Karakousis GC;Mills GB;Lu Y;Ahmed TA;Poulikakos PI;Caponigro G;Boehm M;Peters M;Schuchter LM;Weeraratna AT;Herlyn M
通讯作者：
Herlyn M
影响因子：
5
作者：
Balagurunathan, Yoganand;Gu, Yuhua;Gillies, Robert J.
通讯作者：
Gillies, Robert J.
影响因子：
4.4
作者：
Ashton, EA;Takahashi, C;Ekholm, S
通讯作者：
Ekholm, S
影响因子：
--
作者：
Hu P;Wang J;Zhong H;Zhou Z;Shen L;Hu W;Zhang Z
通讯作者：
Zhang Z
影响因子：
8.8
作者：
Li S;Shen D;Shao J;Crowder R;Liu W;Prat A;He X;Liu S;Hoog J;Lu C;Ding L;Griffith OL;Miller C;Larson D;Fulton RS;Harrison M;Mooney T;McMichael JF;Luo J;Tao Y;Goncalves R;Schlosberg C;Hiken JF;Saied L;Sanchez C;Giuntoli T;Bumb C;Cooper C;Kitchens RT;Lin A;Phommaly C;Davies SR;Zhang J;Kavuri MS;McEachern D;Dong YY;Ma C;Pluard T;Naughton M;Bose R;Suresh R;McDowell R;Michel L;Aft R;Gillanders W;DeSchryver K;Wilson RK;Wang S;Mills GB;Gonzalez-Angulo A;Edwards JR;Maher C;Perou CM;Mardis ER;Ellis MJ
通讯作者：
Ellis MJ