【摘要】 目的 分析利用人工智能算法对肺结节CT图像的特征提取效果及肺结节良恶性的影像学表现。方法 回顾性选取瑞安市人民医院2021年12月—2023年12月的298例患者(330个肺结节)的CT图像数据，采用基于期望最大化(EM)算法的肺结节特征提取模型对原始肺结节图像进行图像特征提取，并对图像特征提取效果进行比较分析，分析肺结节良恶性的影像学表现。结果 该模型的敏感度为0.955，提取的图像较好地保留了肺结节和血管。影像学分析显示，恶性组针状征(73.09%)、分叶征(69.96%)、空泡征象(59.19%)、血管收敛征象(74.89%)、胸膜牵引征象(17.49%)检出率均高于良性组(8.41%、0%、3.74%、4.67%、0%)，差异均有统计学意义(P<0.05)。结论 基于EM算法的检测模型的CT图像特征在肺结节检测中具有较好的效能，影像学表现中的针状征、分叶征、空泡征、血管收敛征、胸膜牵拉征可作为鉴别结节良恶性的指标。更多还原

【Abstract】 Objective To analyze the effect of CT image feature extraction of pulmonary nodules based on an artificial intelligence algorithm and the imaging manifestations of benign and malignant pulmonary nodules. Methods The CT image data of 298 patients(330 pulmonary nodules) in Rui’an People’s Hospital from December 2021 to December 2023 were retrospectively selected as the research objects. The pulmonary nodule feature extraction model based on the expectation maximization(EM) algorithm was used to extract the image features of the original pulmonary nodule images, and the effect of image feature extraction was compared and analyzed. The imaging manifestations of benign and malignant pulmonary nodules were analyzed. Results The results showed that the detection sensitivity of pulmonary nodules in this model was 0.955, and the pulmonary nodules and blood vessels were well preserved in the images. Imaging analysis showed that the detection rates of spiculation sign(73.09%), lobulation sign(69.96%), vacuole sign(59.19%), vascular convergence sign(74.89%), and pleural traction sign(17.49%) in the malignant group were higher than those in the benign group(8.41%, 0%, 3.74%, 4.67%, 0%, respectively), and the differences were statistically significant(all P < 0.05). Conclusions The CT image features of the detection model based on the EM algorithm have good performance in the detection of pulmonary nodules. The needle-like sign, lobulation sign, air bubble sign, vascular convergence sign, and pleural traction sign in the imaging manifestations can be used as indicators for differentiating benign and malignant nodules.更多还原