Dutta Kaushik, Roy Sudipta, Whitehead Timothy Daniel, Luo Jingqin, Jha Abhinav Kumar, Li Shunqiang, Quirk James Dennis, Shoghi Kooresh Isaac
Department of Radiology, Washington University School of Medicine, St. Louis, MO 63110, USA.
Department of Surgery, Washington University School of Medicine, St. Louis, MO 63110, USA.
Cancers (Basel). 2021 Jul 28;13(15):3795. doi: 10.3390/cancers13153795.
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 ( ≤ 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, ≤ 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.
临床前磁共振成像（MRI）是联合临床研究流程中的关键组成部分。重要的是，MRI 中的肿瘤分割是肿瘤表型分析和治疗反应评估的必要步骤。然而，手动分割耗时且存在观察者间和观察者内的变异性，缺乏可重复性。本研究旨在开发一种自动化流程，使用深度学习（DL）算法从临床前 T1w 和 T2w MR 图像中准确地定位和描绘三阴乳腺癌（TNBC）人源肿瘤异种移植（PDX）模型肿瘤，并评估影像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、循环残差 U-Net（R2UNet）和密集 R2U-Net（D-R2UNet），并将其与专家手动描绘结果进行比较。为了减轻多位专家之间的偏差，应用了同时真相和性能水平估计（STAPLE）算法来创建一致性图谱。使用性能指标（F1 分数、召回率、精确率和 AUC）来评估网络的性能。多对比度 D-R2UNet 表现最佳，F1 分数 = 0.948；然而，所有网络的得分相互之间相差 1-3%。从 D-R2UNet 提取的影像组学特征与 STAPLE 衍生特征高度相关，T1w 的 67.13%和 T2w 的 53.15%表现出相关性 ρ≥0.9（≤0.05）。相对于 STAPLE，D-R2UNet 提取的特征表现出更好的可重复性，T1w 的 86.71%和 T2w 的 69.93%的特征被发现具有高度可重复性（CCC≥0.9，≤0.05）。最后，39.16%的 T1w 和 13.9%的 T2w 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 算法来规避 T1w 和 T2w MR 图像的手动分割，并确定了影像组学特征对肿瘤边界的敏感性。
