Department of Pathology, Beijing Chest Hospital, Capital Medical University/Beijing Tuberculosis and Thoracic Tumor Research Institute, Beijing, 101149, China.
Digital Manufacturing Laboratory, Beijing Institute of Technology, Beijing, 100081, China.
BMC Pulm Med. 2023 Jul 5;23(1):244. doi: 10.1186/s12890-023-02537-x.
The detection of epidermal growth factor receptor (EGFR) mutations in patients with non-small cell lung cancer is critical for tyrosine kinase inhibitor therapy. EGFR detection requires tissue samples, which are difficult to obtain in some patients, costing them the opportunity for further treatment. To realize EGFR mutation prediction without molecular detection, we aimed to build a high-accuracy deep learning model with only haematoxylin and eosin (H&E)-stained slides.
We collected 326 H&E-stained non-small cell lung cancer slides from Beijing Chest Hospital, China, and used 226 slides (88 with EGFR mutations) for model training. The remaining 100 images (50 with EGFR mutations) were used for testing. We trained a convolutional neural network based on ResNet-50 to classify EGFR mutation status on the slide level.
The sensitivity and specificity of the model were 76% and 74%, respectively, with an area under the curve of 0.82. When applying the double-threshold approach, 33% of the patients could be predicted by the deep learning model as EGFR positive or negative with a sensitivity and specificity of 100.0% and 87.5%. The remaining 67% of the patients got an uncertain result and will be recommenced to perform further examination. By incorporating adenocarcinoma subtype information, we achieved 100% sensitivity in predicting EGFR mutations in 37.3% of adenocarcinoma patients.
Our study demonstrates the potential of a deep learning-based EGFR mutation prediction model for rapid and cost-effective pre-screening. It could serve as a high-accuracy complement to current molecular detection methods and provide treatment opportunities for non-small cell lung cancer patients from whom limited samples are available.
在非小细胞肺癌患者中检测表皮生长因子受体（EGFR）突变对于酪氨酸激酶抑制剂治疗至关重要。EGFR 检测需要组织样本，但在某些患者中难以获得，使他们失去了进一步治疗的机会。为了在没有分子检测的情况下实现 EGFR 突变预测，我们旨在仅使用苏木精和伊红（H&E）染色幻灯片构建高精度的深度学习模型。
我们收集了来自中国北京胸科医院的 326 张 H&E 染色非小细胞肺癌幻灯片，其中 226 张（88 张有 EGFR 突变）用于模型训练。其余 100 张图像（50 张有 EGFR 突变）用于测试。我们基于 ResNet-50 训练了一个卷积神经网络，以在幻灯片水平上对 EGFR 突变状态进行分类。
该模型的灵敏度和特异性分别为 76%和 74%，曲线下面积为 0.82。当应用双阈值方法时，深度学习模型可以预测 33%的患者 EGFR 阳性或阴性，其灵敏度和特异性分别为 100.0%和 87.5%。其余 67%的患者得到不确定的结果，将被建议进行进一步检查。通过纳入腺癌亚型信息，我们在 37.3%的腺癌患者中实现了 100%的 EGFR 突变预测灵敏度。
我们的研究表明，基于深度学习的 EGFR 突变预测模型具有快速、经济有效的初步筛选潜力。它可以作为当前分子检测方法的高精度补充，为那些无法获得足够样本的非小细胞肺癌患者提供治疗机会。
