文献速递乳腺MR

DCE-MRI影像组学分析鉴别乳腺luminal A型与luminal B型分子亚型

DCE-MRI Radiomics Analysis in Differentiating Luminal A and Luminal B Breast Cancer Molecular Subtypes.
PubMed 文献 · Acad Radiol · 2022-05-17

本研究纳入73例经病理证实的浸润性导管癌患者、共81个肿瘤病灶,提取DCE-MRI晚期增强图像的影像组学特征,比较luminal A型与luminal B型分子亚型间的差异。研究发现直方图偏度、形状体积、GLCM熵与能量、GLRLM及GLZLM多项特征在两组间有统计学显著差异,联合形状-体积与GLZLM-HGZE的模型鉴别两型的AUC为0.746。

要点

  • 形状-体积(ml)联合GLZLM-HGZE的模型鉴别luminal A/B型:AUC 0.746、敏感度0.744、特异度0.643、准确率0.694。
  • 体积(ml与体素数)、直方图偏度、GLCM熵/能量、GLRLM的GLNU、RLNU、HGRE及GLZLM多项特征在两型间差异有统计学意义。
  • 作者认为这些影像组学特征可反映分子亚型的瘤内异质性,提示可无创地表征luminal A与luminal B肿瘤。

核心结论

  • 研究识别出可能区分浸润性导管癌luminal A与luminal B分子亚型的影像组学特征,作者认为这些特征反映分子亚型的瘤内异质性,并提示影像组学分析可无创地表征luminal A与luminal B肿瘤。

研究设计与样本

  • 研究为影像组学分析,纳入73例经组织病理学证实的浸润性导管癌患者,共评估81个肿瘤病灶,于T1加权脂肪抑制矢状位增强晚期图像上完成预处理与三维分割,使用LIFEX软件提取43个影像组学特征(1个常规、4个形状、6个直方图、7个GLCM、11个GLRLM、3个NGLDM、11个GLZLM),按ER/PR、HER2及Ki-67状态将肿瘤分为luminal A型与luminal B型。

诊断标准与关键数值

  • 分子亚型定义:luminal A为ER阳性和/或PR阳性、HER2阴性、Ki-67<20;luminal B为ER阳性和/或PR阳性、HER2阳性,或HER2阴性且Ki-67≥20。在两组分子亚型间差异有统计学意义的影像组学特征包括直方图偏度,形状的volume-ml与volume-voxel,GLCM的entropy.log10、entropy.log2与energy,GLRLM的GLNU、RLNU与HGRE,NGLDM的busyness,以及GLZLM的GLNU、HGZE、ZLNU与SZE。形状-体积(ml)联合GLZLM-HGZE的模型AUC为0.746、敏感度0.744、特异度0.643、准确率0.694。

局限性

  • 原文未报告局限性。

以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。

英文原文(摘要)· 点击展开

RATIONALE AND OBJECTIVES: The aim of the present study was to investigate the association between Luminal A and Luminal B molecular subtypes and radiomic features of dynamic contrast‑enhanced magnetic resonance imaging in patients with invasive breast cancer.

MATERIALS AND METHODS: Seventy-three patients with histopathologically proven invasive ductal cancer (IDC) were selected. Tumors were classified into molecular subtypes: Luminal A (estrogen receptor (ER)-positive and/or progesterone receptor (PR)-positive, human epidermal growth factor receptor type 2 (HER2) -negative, proliferation marker Ki-67<20) and Luminal B (ER-positive and/or PR-positive, HER2-positive or HER2-negative with high Ki-67 ≥20). A total of 81 tumoral lesions were evaluated on T1-weighted fat-suppressed sagittal post-contrast late-phase MRI images after the required "pre-process" steps and 3D segmentations were made. Forty-three radiomic features including: 1 conventional, 4 shape, 6 histogram, 7 Grey-Level Co-occurrence Matrix (GLCM), 11 Grey-Level Run-Length Matrix (GLRLM), 3 Neighborhood Grey-Level Difference Matrix (NGLDM), 11 Grey-Level Zone-Length Matrix (GLZLM) were extracted by using the software LIFEX.

RESULTS: A statistically significant difference was found in radiomic features including; a) Histogram: "skewness", b) Shape: "volume-ml, volume-voxel," c) GLCM: "entropy.log10, entropy.log2, energy", d) GLRLM: "GLNU, RLNU, HGRE," e) NGLDM: "busyness," f) GLZLM: "GLNU, HGZE, ZLNU, SZE" between two different molecular subtypes. The model combining Shape-volume (ml) and GLZLM-HGZE yielded 0.746 area under the curve (AUC), 0.744 sensitivity, 0.643 specificity and 0.694 accuracy.

CONCLUSION: Radiomic properties that may distinguish Luminal A and Luminal B molecular subtypes of IDC were identified. The radiomic features were thought to reflect the intratumoral heterogeneity in molecular subtypes. This study demonstrated that the characterization of Luminal A and Luminal B tumors could be made non-invasively by radiomics analysis.

来源:Acad Radiol,查看原文。版权归原作者所有。

引用

  1. Lafcı, Celepli, Seher Öztekin, et al. DCE-MRI Radiomics Analysis in Differentiating Luminal A and Luminal B Breast Cancer Molecular Subtypes. Acad Radiol. 2022. PMID: 35595629. doi:10.1016/j.acra.2022.04.004

供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。