Videomics of the Upper Aero-Digestive Tract Cancer: Deep Learning Applied to White Light and Narrow Band Imaging for Automatic Segmentation of Endoscopic Images.
Videomics of the Upper Aero-Digestive Tract Cancer: Deep Learning Applied to White Light and Narrow Band Imaging for Automatic Segmentation of Endoscopic Images.
复制标题
上消化道肿瘤的视频组学：应用深度学习的白光和窄带成像用于内镜图像的自动分割。
DOI：
10.3389/fonc.2022.900451
复制
发表时间：
2022
影响因子：
4.7
通讯作者：
Peretti, Giorgio
中科院分区：
文献类型：
--
作者：
Azam, Muhammad Adeel;Sampieri, Claudio;Ioppi, Alessandro;Benzi, Pietro;Giordano, Giorgio Gregory;De Vecchi, Marta;Campagnari, Valentina;Li, Shunlei;Guastini, Luca;Paderno, Alberto;Moccia, Sara;Piazza, Cesare;Mattos, Leonardo S.;Peretti, Giorgio
关键词：
窄带成像（NBI）是一种内窥镜可视化技术，可用于上呼吸消化道（UADT）癌症检测和边缘评估。然而，NBI分析强烈依赖于运营商，需要高度的专业知识，从而限制了其更广泛的实施。最近，人工智能（AI）已经证明了在UADT视频内窥镜中的应用潜力。在人工智能方法中，深度学习算法，特别是卷积神经网络（CNN），特别适合在视频内窥镜上描绘癌症。本研究旨在开发一种CNN，用于内窥镜图像上UADT癌症的自动语义分割。收集喉鳞状细胞癌（LSCC）的白色光和NBI视频帧的数据集并手动注释。设计了一种新的DL分割模型（SegMENT）。Segment依赖于DeepLabV3+ CNN架构，使用Xception作为主干进行修改，并结合其他CNN的集成功能。SegMENT的性能与最先进的CNN（UNet，ResUNet和DeepLabv3）进行了比较。然后在从先前发表的研究中获得的口咽（OPSCC）和口腔SCC（OSCC）的NBI图像的两个外部数据集上验证SegMENT。在外部数据集上评估了通过集成技术进行域内迁移学习的影响。219例喉鳞状细胞癌患者回顾性纳入研究。总共683个视频帧组成了LSCC数据集，而OPSCC和OCSCC的外部验证队列包含116和102个图像。在LSCC数据集上，SegMENT的表现优于其他DL模型，获得了以下中值：0.68交集大于并集（IoU），0.81骰子相似系数（DSC），0.95召回率，0.78精度，0.97准确度。对于OCSCC和OPSCC数据集，结果上级先前发表的数据：中位性能指标分别改善如下：DSC=10.3%和11.9%，召回率=15.0%和5.1%，精确度=17.0%和14.7%，准确度=4.1%和10.3%。SegMENT实现了有前途的性能，表明即使在高度异构和复杂的UADT环境中，内窥镜图像中的自动肿瘤分割也是可行的。SegMENT在外部验证队列中的表现优于先前发表的结果。该模型显示出改善早期肿瘤检测、更精确活检和更好选择切除边缘的潜力。
Narrow Band Imaging (NBI) is an endoscopic visualization technique useful for upper aero-digestive tract (UADT) cancer detection and margins evaluation. However, NBI analysis is strongly operator-dependent and requires high expertise, thus limiting its wider implementation. Recently, artificial intelligence (AI) has demonstrated potential for applications in UADT videoendoscopy. Among AI methods, deep learning algorithms, and especially convolutional neural networks (CNNs), are particularly suitable for delineating cancers on videoendoscopy. This study is aimed to develop a CNN for automatic semantic segmentation of UADT cancer on endoscopic images. A dataset of white light and NBI videoframes of laryngeal squamous cell carcinoma (LSCC) was collected and manually annotated. A novel DL segmentation model (SegMENT) was designed. SegMENT relies on DeepLabV3+ CNN architecture, modified using Xception as a backbone and incorporating ensemble features from other CNNs. The performance of SegMENT was compared to state-of-the-art CNNs (UNet, ResUNet, and DeepLabv3). SegMENT was then validated on two external datasets of NBI images of oropharyngeal (OPSCC) and oral cavity SCC (OSCC) obtained from a previously published study. The impact of in-domain transfer learning through an ensemble technique was evaluated on the external datasets. 219 LSCC patients were retrospectively included in the study. A total of 683 videoframes composed the LSCC dataset, while the external validation cohorts of OPSCC and OCSCC contained 116 and 102 images. On the LSCC dataset, SegMENT outperformed the other DL models, obtaining the following median values: 0.68 intersection over union (IoU), 0.81 dice similarity coefficient (DSC), 0.95 recall, 0.78 precision, 0.97 accuracy. For the OCSCC and OPSCC datasets, results were superior compared to previously published data: the median performance metrics were, respectively, improved as follows: DSC=10.3% and 11.9%, recall=15.0% and 5.1%, precision=17.0% and 14.7%, accuracy=4.1% and 10.3%. SegMENT achieved promising performances, showing that automatic tumor segmentation in endoscopic images is feasible even within the highly heterogeneous and complex UADT environment. SegMENT outperformed the previously published results on the external validation cohorts. The model demonstrated potential for improved detection of early tumors, more precise biopsies, and better selection of resection margins.
登录
查看更多内容
影响因子：
5.1
作者：
Ji, Bin;Ren, Jianjun;Liu, Kai
通讯作者：
Liu, Kai
DOI：
10.1002/hed.23705
发表时间：
2015-08-01
影响因子：
2.9
作者：
Madana, J.;Lim, Chwee Ming;Loh, Kwok Seng
通讯作者：
Loh, Kwok Seng
影响因子：
4.7
作者：
Fiz I;Mazzola F;Fiz F;Marchi F;Filauro M;Paderno A;Parrinello G;Piazza C;Peretti G
通讯作者：
Peretti G
影响因子：
2.6
作者：
Carta, Filippo;Sionis, Sara;Puxeddu, Roberto
通讯作者：
Puxeddu, Roberto
影响因子：
4.6
作者：
Gong J;Holsinger FC;Noel JE;Mitani S;Jopling J;Bedi N;Koh YW;Orloff LA;Cernea CR;Yeung S
通讯作者：
Yeung S