Department of Radiology, Brigham and Women's Hospital, Harvard Medical School, 75 Francis Street, Boston, MA, 02115, USA.
Department of Imaging, Dana-Farber Cancer Institute, Harvard Medical School, 450 Brookline Avenue, Boston, MA, 02215, USA.
Abdom Radiol (NY). 2021 Jan;46(1):311-318. doi: 10.1007/s00261-020-02629-w. Epub 2020 Jul 1.
To create a CT texture-based machine learning algorithm that distinguishes benign from potentially malignant cystic renal masses as defined by the Bosniak Classification version 2019.
In this IRB-approved, HIPAA-compliant study, 4,454 adult patients underwent renal mass protocol CT or CT urography from January 2011 to June 2018. Of these, 257 cystic renal masses were included in the final study cohort. Each mass was independently classified using Bosniak version 2019 by three radiologists, resulting in 185 benign (Bosniak I or II) and 72 potentially malignant (Bosniak IIF, III or IV) masses. Six texture features: mean, standard deviation, mean of positive pixels, entropy, skewness, kurtosis were extracted using commercial software TexRAD (Feedback PLC, Cambridge, UK). Random forest (RF), logistic regression (LR), and support vector machine (SVM) machine learning algorithms were implemented to classify cystic renal masses into the two groups and tested with tenfold cross validations.
Higher mean, standard deviation, mean of positive pixels, entropy, skewness were statistically associated with the potentially malignant group (P ≤ 0.0015 each). Sensitivity, specificity, positive predictive value, negative predictive value, and area under curve of RF model was 0.67, 0.91, 0.75, 0.88, 0.88; of LR model was 0.63, 0.93, 0.78, 0.86, 0.90, and of SVM model was 0.56, 0.91, 0.71, 0.84, 0.89, respectively.
Three CT texture-based machine learning algorithms demonstrated high discriminatory capability in distinguishing benign from potentially malignant cystic renal masses as defined by the Bosniak Classification version 2019. If validated, CT texture-based machine learning algorithms may help reduce interreader variability when applying the Bosniak classification.
创建一种基于 CT 纹理的机器学习算法，以区分 2019 年 Bosniak 分类定义的良性和潜在恶性囊性肾肿瘤。
这项经过机构审查委员会批准、符合 HIPAA 标准的研究纳入了 2011 年 1 月至 2018 年 6 月期间接受肾脏肿块协议 CT 或 CT 尿路造影的 4454 名成年患者。其中，257 个囊性肾肿瘤纳入最终研究队列。由 3 名放射科医生独立使用 Bosniak 2019 版对每个肿块进行分类，结果为 185 个良性（Bosniak I 或 II）和 72 个潜在恶性（Bosniak IIF、III 或 IV）肿块。使用商业软件 TexRAD（英国剑桥的 Feedback PLC）提取 6 个纹理特征：均值、标准差、阳性像素均值、熵、偏度、峰度。实施随机森林（RF）、逻辑回归（LR）和支持向量机（SVM）机器学习算法将囊性肾肿瘤分为两组，并进行十折交叉验证测试。
较高的均值、标准差、阳性像素均值、熵、偏度与潜在恶性组相关（P≤0.0015）。RF 模型的敏感性、特异性、阳性预测值、阴性预测值和曲线下面积分别为 0.67、0.91、0.75、0.88、0.88；LR 模型分别为 0.63、0.93、0.78、0.86、0.90；SVM 模型分别为 0.56、0.91、0.71、0.84、0.89。
三种基于 CT 纹理的机器学习算法在区分 2019 年 Bosniak 分类定义的良性和潜在恶性囊性肾肿瘤方面具有较高的鉴别能力。如果得到验证，基于 CT 纹理的机器学习算法可能有助于减少应用 Bosniak 分类时的读者间变异性。
