Shenouda Mena, Shaikh Abbas, Deutsch Ilana, Mitchell Owen, Kindler Hedy L, Armato Samuel G
The University of Chicago, Department of Radiology, Chicago, Illinois, United States.
Rice University, Houston, Texas, United States.
J Med Imaging (Bellingham). 2024 Nov;11(6):064501. doi: 10.1117/1.JMI.11.6.064501. Epub 2024 Dec 11.
The BRCA1-associated protein 1 () gene is of great interest because somatic () mutations are the most common alteration associated with pleural mesothelioma (PM). Further, germline mutation of the gene has been linked to the development of PM. This study aimed to explore the potential of radiomics on computed tomography scans to identify somatic gene mutations and assess the feasibility of radiomics in future research in identifying germline mutations.
A cohort of 149 patients with PM and known somatic mutation status was collected, and a previously published deep learning model was used to first automatically segment the tumor, followed by radiologist modifications. Image preprocessing was performed, and texture features were extracted from the segmented tumor regions. The top features were selected and used to train 18 separate machine learning models using leave-one-out cross-validation (LOOCV). The performance of the models in distinguishing between -mutated () and wild-type () tumors was evaluated using the receiver operating characteristic area under the curve (ROC AUC).
A decision tree classifier achieved the highest overall AUC value of 0.69 (95% confidence interval: 0.60 and 0.77). The features selected most frequently through the LOOCV were all second-order (gray-level co-occurrence or gray-level size zone matrices) and were extracted from images with an applied transformation.
This proof-of-concept work demonstrated the potential of radiomics to differentiate among in patients with PM. Future work will extend these methods to the assessment of germline mutation status through image analysis for improved patient prognostication.
乳腺癌1相关蛋白1（ ）基因备受关注，因为体细胞（ ）突变是与胸膜间皮瘤（PM）相关的最常见改变。此外，该基因的种系突变与PM的发生有关。本研究旨在探讨计算机断层扫描上的放射组学在识别体细胞 基因突变方面的潜力，并评估放射组学在未来识别种系突变研究中的可行性。
收集了149例已知体细胞 突变状态的PM患者队列，使用先前发表的深度学习模型首先自动分割肿瘤，随后由放射科医生进行修改。进行图像预处理，并从分割后的肿瘤区域提取纹理特征。选择顶级特征并使用留一法交叉验证（LOOCV）训练18个独立的机器学习模型。使用曲线下面积的受试者工作特征（ROC AUC）评估模型区分 -突变（ ）和 野生型（ ）肿瘤的性能。
决策树分类器的总体AUC值最高，为0.69（95%置信区间：0.60和0.77）。通过LOOCV最频繁选择的特征均为二阶特征（灰度共生矩阵或灰度大小区域矩阵），并从应用了变换的图像中提取。
这项概念验证工作证明了放射组学在区分PM患者的 方面的潜力。未来的工作将通过图像分析将这些方法扩展到种系 突变状态的评估，以改善患者的预后。
