胶质母细胞瘤扫描-重复扫描MRI研究中的影像组学重复性陷阱
Radiomics Repeatability Pitfalls in a Scan-Rescan MRI Study of Glioblastoma.
作者信息
Hoebel Katharina V, Patel Jay B, Beers Andrew L, Chang Ken, Singh Praveer, Brown James M, Pinho Marco C, Batchelor Tracy T, Gerstner Elizabeth R, Rosen Bruce R, Kalpathy-Cramer Jayashree
机构信息
Athinoula A. Martinos Center for Biomedical Imaging, Department of Radiology (K.V.H., J.B.P., A.L.B., K.C., P.S., J.M.B., M.C.P., B.R.R., J.K.C.), and Stephen E. and Catherine Pappas Center for Neuro-Oncology (T.T.B., E.R.G.), Massachusetts General Hospital, 149 13th St, Charlestown, MA 02129; and Harvard-MIT Division of Health Sciences and Technology, Cambridge, Mass (K.V.H., J.B.P., K.C.).
出版信息
Radiol Artif Intell. 2020 Dec 16;3(1):e190199. doi: 10.1148/ryai.2020190199. eCollection 2021 Jan.
PURPOSE
To determine the influence of preprocessing on the repeatability and redundancy of radiomics features extracted using a popular open-source radiomics software package in a scan-rescan glioblastoma MRI study.
MATERIALS AND METHODS
In this study, a secondary analysis of T2-weighted fluid-attenuated inversion recovery (FLAIR) and T1-weighted postcontrast images from 48 patients (mean age, 56 years [range, 22-77 years]) diagnosed with glioblastoma were included from two prospective studies (ClinicalTrials.gov NCT00662506 [2009-2011] and NCT00756106 [2008-2011]). All patients underwent two baseline scans 2-6 days apart using identical imaging protocols on 3-T MRI systems. No treatment occurred between scan and rescan, and tumors were essentially unchanged visually. Radiomic features were extracted by using PyRadiomics https://pyradiomics.readthedocs.io/ under varying conditions, including normalization strategies and intensity quantization. Subsequently, intraclass correlation coefficients were determined between feature values of the scan and rescan.
RESULTS
Shape features showed a higher repeatability than intensity (adjusted < .001) and texture features (adjusted < .001) for both T2-weighted FLAIR and T1-weighted postcontrast images. Normalization improved the overlap between the region of interest intensity histograms of scan and rescan (adjusted < .001 for both T2-weighted FLAIR and T1-weighted postcontrast images), except in scans where brain extraction fails. As such, normalization significantly improves the repeatability of intensity features from T2-weighted FLAIR scans (adjusted = .003 [ score normalization] and adjusted = .002 [histogram matching]). The use of a relative intensity binning strategy as opposed to default absolute intensity binning reduces correlation between gray-level co-occurrence matrix features after normalization.
CONCLUSION
Both normalization and intensity quantization have an effect on the level of repeatability and redundancy of features, emphasizing the importance of both accurate reporting of methodology in radiomics articles and understanding the limitations of choices made in pipeline design. © RSNA, 2020See also the commentary by Tiwari and Verma in this issue.
目的
在一项胶质母细胞瘤MRI扫描-再扫描研究中，确定预处理对使用流行的开源放射组学软件包提取的放射组学特征的可重复性和冗余性的影响。
材料与方法
在本研究中，纳入了来自两项前瞻性研究（ClinicalTrials.gov NCT00662506[2009 - 2011]和NCT00756106[2008 - 2011]）的48例（平均年龄56岁[范围22 - 77岁]）诊断为胶质母细胞瘤患者的T2加权液体衰减反转恢复（FLAIR）和T1加权增强后图像进行二次分析。所有患者在3-T MRI系统上使用相同的成像方案，在相隔2 - 6天的时间内进行了两次基线扫描。扫描和再扫描之间未进行治疗，肿瘤在视觉上基本无变化。在不同条件下，包括归一化策略和强度量化，使用PyRadiomics（https://pyradiomics.readthedocs.io/）提取放射组学特征。随后，确定扫描和再扫描的特征值之间的组内相关系数。
结果
对于T2加权FLAIR图像和T1加权增强后图像，形状特征的可重复性高于强度特征（校正<0.001）和纹理特征（校正<0.001）。归一化改善了扫描和再扫描的感兴趣区域强度直方图之间的重叠（T2加权FLAIR图像和T1加权增强后图像的校正均<0.001），除非在脑提取失败的扫描中。因此，归一化显著提高了T2加权FLAIR扫描强度特征的可重复性（校正=0.003[分数归一化]和校正=0.002[直方图匹配]）。与默认的绝对强度分箱策略相比，使用相对强度分箱策略可降低归一化后灰度共生矩阵特征之间的相关性。
结论
归一化和强度量化均对特征的可重复性和冗余水平有影响，强调了在放射组学文章中准确报告方法以及理解管道设计中所做选择的局限性的重要性。©RSNA，2020另见本期Tiwari和Verma的评论。
Zotero 插件