From the Departments of Diagnostic Radiology (Y.U., R.F., A.D., B.G., C.R.), Obstetrics and Gynecology (X.Z.Z., L.G.), and Pathology (J.A., L.F.), Royal Victoria Hospital, McGill University Health Centre, 1001 Decarie Blvd, Montreal, QC, Canada H4A 3J1; Department of Radiology, Kobe University Graduate School of Medicine, Kobe, Japan (Y.U.); Reza Forghani Medical Services, Montreal, Canada (B.F., R.F.); Department of Radiology, Jewish General Hospital, Montreal, Canada (R.F.); Department of Body and Interventional Imaging, Hôpital Lariboisière, Assistance Publique-Hôpitaux de Paris, Université Diderot-Paris 7 and INSERM U965, Paris, France (A.D.); and Department of Radiology, Université Paris Descartes Sorbonne Paris Cité, Assistance Publique-Hôpitaux de Paris, Hôpital Européen Georges Pompidou, Paris, France (F.C.).
Radiology. 2017 Sep;284(3):748-757. doi: 10.1148/radiol.2017161950. Epub 2017 May 10.
Purpose To evaluate the associations among mathematical modeling with the use of magnetic resonance (MR) imaging-based texture features and deep myometrial invasion (DMI), lymphovascular space invasion (LVSI), and histologic high-grade endometrial carcinoma. Materials and Methods Institutional review board approval was obtained for this retrospective study. This study included 137 women with endometrial carcinomas measuring greater than 1 cm in maximal diameter who underwent 1.5-T MR imaging before hysterectomy between January 2011 and December 2015. Texture analysis was performed with commercial research software with manual delineation of a region of interest around the tumor on MR images (T2-weighted, diffusion-weighted, and dynamic contrast material-enhanced images and apparent diffusion coefficient maps). Areas under the receiver operating characteristic curve and diagnostic performance of random forest models determined by using a subset of the most relevant texture features were estimated and compared with those of independent and blinded visual assessments by three subspecialty radiologists. Results A total of 180 texture features were extracted and ultimately limited to 11 features for DMI, 12 for LVSI, and 16 for high-grade tumor for random forest modeling. With random forest models, areas under the receiver operating characteristic curve, sensitivity, specificity, accuracy, positive predictive value, and negative predictive value were estimated at 0.84, 79.3%, 82.3%, 81.0%, 76.7%, and 84.4% for DMI; 0.80, 80.9%, 72.5%, 76.6%, 74.3%, and 79.4% for LVSI; and 0.83, 81.0%, 76.8%, 78.1%, 60.7%, and 90.1% for high-grade tumor, respectively. Sensitivity, specificity, accuracy, positive predictive value, and negative predictive value of visual assessment for DMI were 84.5%, 82.3%, 83.2%, 77.7%, and 87.8% (reader 3). Conclusion The mathematical models that incorporated MR imaging-based texture features were associated with the presence of DMI, LVSI, and high-grade tumor and achieved equivalent accuracy to that of subspecialty radiologists for assessment of DMI in endometrial cancers larger than 1 cm. However, these preliminary results must be interpreted with caution until they are validated with an independent data set, because the small sample size relative to the number of features extracted may have resulted in overfitting of the models. RSNA, 2017 Online supplemental material is available for this article.
目的 评估基于磁共振成像（MR）的纹理特征的数学模型与深肌层浸润（DMI）、脉管侵犯（LVSI）和组织学高级别子宫内膜癌之间的相关性。
材料与方法 本回顾性研究获得了机构审查委员会的批准。这项研究纳入了 137 名最大直径大于 1cm 的子宫内膜癌患者，这些患者在 2011 年 1 月至 2015 年 12 月期间在接受子宫切除术前行 1.5-T MR 成像检查。采用商用研究软件，在 MR 图像（T2 加权、弥散加权和动态对比增强图像及表观弥散系数图）上手动勾画肿瘤周围的感兴趣区进行纹理分析。采用随机森林模型确定的最相关纹理特征子集估计并比较了受试者工作特征曲线下面积和诊断性能，并与三位放射科亚专业医生的独立盲法评估结果进行了比较。
结果 共提取了 180 个纹理特征，最终将其限制在用于随机森林模型的 DMI 为 11 个、LVSI 为 12 个、高级别肿瘤为 16 个特征。在随机森林模型中，对于 DMI，曲线下面积、敏感性、特异性、准确性、阳性预测值和阴性预测值分别为 0.84、79.3%、82.3%、81.0%、76.7%和 84.4%；对于 LVSI，曲线下面积、敏感性、特异性、准确性、阳性预测值和阴性预测值分别为 0.80、80.9%、72.5%、76.6%、74.3%和 79.4%；对于高级别肿瘤，曲线下面积、敏感性、特异性、准确性、阳性预测值和阴性预测值分别为 0.83、81.0%、76.8%、78.1%、60.7%和 90.1%。DMI 视觉评估的敏感性、特异性、准确性、阳性预测值和阴性预测值分别为 84.5%、82.3%、83.2%、77.7%和 87.8%（读者 3）。
结论 纳入基于 MR 成像的纹理特征的数学模型与 DMI、LVSI 和高级别肿瘤的存在相关，并且在评估大于 1cm 的子宫内膜癌的 DMI 方面与放射科亚专业医生的准确性相当。然而，由于提取的特征数量相对于样本量较小，模型可能存在过拟合，因此这些初步结果的解释需要谨慎，直到它们在独立数据集上得到验证。RSNA，2017 在线补充材料可在本文中查看。
