Detecting lung nodules with low-dose computed tomography (CT) can predict the future risk suffering from lung cancers. However, there are a few studies on lung nodules with low-dose CT and detecting rate is very low at present. In order to accurately detect lung nodules with low-dose CT, this project proposes a solution based on an integrated deep learning algorithm. Firstly, CT images are preprocessed via image clipping, normalization and segmentation, and the positive samples are expanded to balance the number of positive and negative samples. The features of candidate lung nodule samples are learned by using convolutional neural network and residual network, and then import into long short-term memory network, respectively. We then fuse these features, continuously optimize the network parameters during the training process, and finally obtain the model with an optimal performance. Compared to other algorithms, all metrics in the proposed algorithm are improved. This model has an obvious anti-interference ability. It is stable and can identify lung nodules effectively, which is expected to provide auxiliary diagnostic for early screening of lung cancers. However, with the development of 6G and Internet of Medical Things (IoMT), patients will raise higher requirements for the convenience and effectiveness of medical diagnosis. We will plan to incorporate the integrated deep learning algorithm into 6G-enabled IoMT to provide and share medical records and monitoring results online and realize an online diagnosis.