Department of General Medicine, The First Affiliated Hospital of Suzhou University, Suzhou 215006, Jiangsu, China.
Contrast Media Mol Imaging. 2022 Jun 6;2022:2574451. doi: 10.1155/2022/2574451. eCollection 2022.
This study was to evaluate the clinical efficacy of pemetrexed combined with platinum-based chemotherapy in the treatment of elderly lung cancer using electronic computed tomography (CT) images based on artificial intelligence algorithms. In this study, 80 elderly patients with lung cancer treated were selected and randomly divided into two groups: patients treated with pemetrexed combined with cisplatin were included in the pemetrexed group and patients treated with docetaxel combined with cisplatin were included in the docetaxel group, with 40 cases in each group. The DenseNet network was compared with the Let Net-5 and ResNet model and applied to the CT images of 80 elderly patients with lung cancer. The diagnosis accuracy of the DenseNet network (97.4%) was higher than that of the Let Net-5 network (80.1%) and ResNet model (95.5%). Carcinoembryonic antigen (CEA), cytokeratin fragment antigen 21-1 (CYFRA 21-1), and squamous cell-associated antigen (SCC) after chemotherapy in the pemetrexed group and docetaxel group were all lower than those before chemotherapy, showing statistically obvious differences ( < 0.05). The satisfaction degree of nursing care in the pemetrexed group (92.67%) was significantly higher than that in the docetaxel group (85.62%), and the difference was statistically significant ( < 0.05). Adverse reactions such as fatigue, diarrhea, and neutrophils in the pemetrexed group were lower than those in the docetaxel group, and the difference was statistically great ( < 0.05). The DenseNet convolutional neural network has high diagnostic accuracy; methotrexate combined with platinum chemotherapy can improve the chemotherapy effect in elderly patients with lung cancer, with low degree of adverse reactions and good overall tolerance, which can be used as the first-line treatment for elderly patients with lung cancer.
本研究旨在通过人工智能算法基于电子计算机断层扫描（CT）图像评估培美曲塞联合铂类化疗治疗老年肺癌的临床疗效。本研究选取 80 例老年肺癌患者，随机分为培美曲塞联合顺铂组（培美组）和多西他赛联合顺铂组（多西组），每组 40 例。比较 DenseNet 网络与 Let Net-5 和 ResNet 模型，并应用于 80 例老年肺癌患者的 CT 图像。DenseNet 网络的诊断准确率（97.4%）高于 Let Net-5 网络（80.1%）和 ResNet 模型（95.5%）。培美组和多西组患者化疗后癌胚抗原（CEA）、细胞角蛋白 19 片段抗原 21-1（CYFRA 21-1）和鳞状细胞相关抗原（SCC）均低于化疗前，差异有统计学意义（ < 0.05）。培美组的护理满意度（92.67%）明显高于多西组（85.62%），差异有统计学意义（ < 0.05）。培美组患者的不良反应发生率（如疲劳、腹泻和中性粒细胞减少症）低于多西组，差异有统计学意义（ < 0.05）。DenseNet 卷积神经网络具有较高的诊断准确率；培美曲塞联合铂类化疗能提高老年肺癌患者的化疗效果，不良反应程度低，整体耐受性好，可作为老年肺癌患者的一线治疗药物。
