论著摘要 |【MR】半自动肿瘤分割软件对多形性胶质母细胞瘤的影像组学特征质量的影响(双语版)

2017-08-29 11:48:32 admin 18

Quality of Radiomic Features in Glioblastoma Multiforme: Impact of Semi-Automated TumorSegmentation Software.

发表日期:2017.4.19    来源:Korean J Radiol. 

作者:Lee M1,2Woo B3Kuo MD4,5Jamshidi N5Kim JH1,2,3.

作者介绍

    1.Center for Medical-IT Convergence Technology Research, Advanced Institutes of Convergence Technology, Seoul National University, Suwon 16229, Korea.

    2.Department of Radiology, Seoul National University Hospital, Seoul 03080, Korea.

    3.Department of Transdisciplinary Studies, Graduate School of Convergence Science and Technology, Seoul National University, Suwon 16229, Korea.

    4.Department of Electronic and Computer Engineering, National Chiao Tung University, Hsinchu 300, Taiwan.

    5.Department of Radiological Sciences, University of California, Los Angeles, Los Angeles, CA 90095, USA.

目的

本研究的目的是评估多形性胶质母细胞瘤(GBM)中由半自动肿瘤分割软件获得的肿瘤体积得到的影像组学特征的可靠性和质量。

The purpose of this study was to evaluate the reliability and quality of radiomic features in glioblastoma multiforme (GBM) derived from tumor volumes obtained with semi-automated tumor segmentation software.

材料和方法

从癌症成像档案中心下载了45例使用增强T1加权成像和液体衰减反转恢复MR序列的多形性胶质母细胞瘤患者(29名男性,16名女性)的MR图像。两个评估者分别使用两个半自动分割工具(TumorPrism3D3D切片机)独立分割肿瘤。每个肿瘤中与增强病变坏死部分和非增强T2高信号强度部分相对应的目标区域进行分割。总共提取了180个成像特征,使用类内相关系数,集群共识和兰德统计,评估其在稳定性,归一化动态范围(NDR)和冗余度方面的质量。



MR images of 45 GBM patients (29 males, 16 females) were downloaded from The Cancer Imaging Archive, in which post-contrast T1-weighted imaging and fluid-attenuated inversion recovery MR sequences were used. Two raters independently segmented the tumors using two semi-automated segmentation tools (TumorPrism3D and 3D Slicer). Regions of interest corresponding to contrast-enhancing lesion, necrotic portions, and non-enhancing T2 high signal intensity component were segmented for each tumor. A total of 180 imaging features were extracted, and their quality was evaluated in terms of stability, normalized dynamic range (NDR), and redundancy, using intra-class correlation coefficients, cluster consensus, and Rand Statistic.

结果

我们的研究结果表明,多形性胶质母细胞瘤中大部分影像组学特征是高度稳定的。 180个特征中90%以上表现出良好的稳定性(类内相关系数[ICC]0.8),而只有7个特征稳定性差(ICC<0.5)。大多数一级统计和形态特征显示中度至高NDR4> NDR1),而超过35%的纹理特征显示差的NDR<1)。特征仅被分为5组,表明它们是高度冗余的。



Our study results showed that most of the radiomic features in GBM were highly stable. Over 90% of 180 features showed good stability (intra-class correlation coefficient [ICC] ≥ 0.8), whereas only 7 features were of poor stability (ICC < 0.5). Most first order statistics and morphometric features showed moderate-to-high NDR (4 > NDR ≥1), while above 35% of the texture features showed poor NDR (< 1). Features were shown to cluster into only 5 groups, indicating that they were highly redundant.

结论

使用半自动化软件工具提供了足够可靠的肿瘤分割和特征稳定性,从而有助于克服用户干预的固有的评估者和评估者之间的差异。然而,在进一步发展影像组学之前,需要评估包括归一化动态范围和冗余性在内的特征质量,以确定代表性的特征标记。

The use of semi-automated software tools provided sufficiently reliable tumor segmentation and feature stability; thus helping to overcome the inherent inter-rater and intra-rater variability of user intervention. However, certain aspects of feature quality, including NDR and redundancy, need to be assessed for determination of representative signature features before further development of radiomics.

关键词

特征质量,多形性胶质母细胞瘤,影像组学,半自动分割,癌症基因组图谱,癌症成像档案中心

Feature quality; Glioblastoma multiforme; Radiomics; Semi-automated segmentation; The Cancer Genome Atlas; The Cancer Imaging Archive


慧影医疗科技(北京)有限公司

地点:北京市海淀区中关村东升科技园B2-C103

电话:400-890-9020

邮箱:radcloud@huiyihuiying.com

关闭
图片
图片
  • 人工智能诊断云平台