From the Radiomics Group, Vall d'Hebron Institute of Oncology, Carrer de Natzaret 115-117, Barcelona 08035, Spain (O.P., C. Macarro, C. Monreal, M.L., A.G.R., F.G., K.B., R.P.L.); Oncology Data Science Group, Vall d'Hebron Institute of Oncology, Barcelona, Spain (V.N., R.D.); Molecular Pathology Group, Vall d'Hebron Institute of Oncology, Barcelona, Spain (G.S., S.S., P.N.); Department of Medical Oncology, Vall d'Hebron University Hospital, Barcelona, Spain (I.B., M.V., E.G., J.C.); Molecular Therapeutic Research Unit, Vall d'Hebron Institute of Oncology, Barcelona, Spain (I.B., M.V., E.G., J.C.); Department of Radiology, Vall d'Hebron University Hospital, Barcelona, Spain (M.E.); Biomakers and Clonal Dynamics Group, Vall d'Hebron Institute of Oncology (VHIO), Barcelona, Spain (R.T.); Department of Physiology and Medical Physics, Centre for Systems Medicine, Royal College of Surgeons in Ireland, Dublin, Ireland (A.T.B.); and National Pre-clinical Imaging Centre, Dublin, Ireland (A.T.B.).
Radiol Artif Intell. 2024 Mar;6(2):e230118. doi: 10.1148/ryai.230118.
Purpose To identify precise three-dimensional radiomics features in CT images that enable computation of stable and biologically meaningful habitats with machine learning for cancer heterogeneity assessment. Materials and Methods This retrospective study included 2436 liver or lung lesions from 605 CT scans (November 2010-December 2021) in 331 patients with cancer (mean age, 64.5 years ± 10.1 [SD]; 185 male patients). Three-dimensional radiomics were computed from original and perturbed (simulated retest) images with different combinations of feature computation kernel radius and bin size. The lower 95% confidence limit (LCL) of the intraclass correlation coefficient (ICC) was used to measure repeatability and reproducibility. Precise features were identified by combining repeatability and reproducibility results (LCL of ICC ≥ 0.50). Habitats were obtained with Gaussian mixture models in original and perturbed data using precise radiomics features and compared with habitats obtained using all features. The Dice similarity coefficient (DSC) was used to assess habitat stability. Biologic correlates of CT habitats were explored in a case study, with a cohort of 13 patients with CT, multiparametric MRI, and tumor biopsies. Results Three-dimensional radiomics showed poor repeatability (LCL of ICC: median [IQR], 0.442 [0.312-0.516]) and poor reproducibility against kernel radius (LCL of ICC: median [IQR], 0.440 [0.33-0.526]) but excellent reproducibility against bin size (LCL of ICC: median [IQR], 0.929 [0.853-0.988]). Twenty-six radiomics features were precise, differing in lung and liver lesions. Habitats obtained with precise features (DSC: median [IQR], 0.601 [0.494-0.712] and 0.651 [0.52-0.784] for lung and liver lesions, respectively) were more stable than those obtained with all features (DSC: median [IQR], 0.532 [0.424-0.637] and 0.587 [0.465-0.703] for lung and liver lesions, respectively; < .001). In the case study, CT habitats correlated quantitatively and qualitatively with heterogeneity observed in multiparametric MRI habitats and histology. Conclusion Precise three-dimensional radiomics features were identified on CT images that enabled tumor heterogeneity assessment through stable tumor habitat computation. CT, Diffusion-weighted Imaging, Dynamic Contrast-enhanced MRI, MRI, Radiomics, Unsupervised Learning, Oncology, Liver, Lung . © RSNA, 2024 See also the commentary by Sagreiya in this issue.
目的 确定 CT 图像中精确的三维放射组学特征，以便通过机器学习计算稳定且具有生物学意义的肿瘤异质性评估的栖息地。
材料与方法 本回顾性研究纳入了 331 例癌症患者（平均年龄，64.5 岁±10.1[标准差]；185 例男性患者）的 605 次 CT 扫描（2010 年 11 月至 2021 年 12 月）中 2436 个肝脏或肺部病灶。从原始和受扰（模拟复测）图像中计算了三维放射组学特征，这些图像的特征计算核半径和箱大小不同。使用组内相关系数（ICC）的下限 95%置信区间（LCL）来衡量重复性和再现性。通过结合重复性和再现性结果（ICC 的 LCL≥0.50）来确定精确特征。使用精确的放射组学特征和原始数据以及受扰数据中的高斯混合模型获得栖息地，并与使用所有特征获得的栖息地进行比较。使用 Dice 相似系数（DSC）评估栖息地的稳定性。在一项病例研究中，通过 13 例接受 CT、多参数 MRI 和肿瘤活检的患者，探索了 CT 栖息地的生物学相关性。
结果 三维放射组学表现出较差的重复性（ICC 的 LCL：中位数[四分位数范围]，0.442[0.312-0.516]）和核半径的较差再现性（ICC 的 LCL：中位数[四分位数范围]，0.440[0.33-0.526]），但箱大小的再现性极好（ICC 的 LCL：中位数[四分位数范围]，0.929[0.853-0.988]）。26 个放射组学特征是精确的，在肺部和肝脏病变中存在差异。使用精确特征获得的栖息地（DSC：中位数[四分位数范围]，0.601[0.494-0.712]和 0.651[0.52-0.784]，分别用于肺部和肝脏病变）比使用所有特征获得的栖息地更稳定（DSC：中位数[四分位数范围]，0.532[0.424-0.637]和 0.587[0.465-0.703]，分别用于肺部和肝脏病变；<.001）。在病例研究中，CT 栖息地与多参数 MRI 栖息地和组织学中观察到的异质性在数量和质量上均具有相关性。
结论 在 CT 图像上确定了精确的三维放射组学特征，通过稳定的肿瘤栖息地计算能够评估肿瘤异质性。
放射组学、未分组学习、肿瘤学、肺、肝脏、CT、扩散加权成像、动态对比增强 MRI、MRI
