The paper introduces a novel deep learning approach, the Residual Encoder-Decoder Convolutional Neural Network (RED-CNN), for low-dose CT imaging. The RED-CNN combines autoencoders, deconvolution networks, and shortcut connections to effectively reduce noise, preserve structural details, and enhance lesion detection. The method is evaluated using both simulated and clinical datasets, demonstrating superior performance compared to state-of-the-art methods in terms of noise suppression, structural preservation, and lesion detection. The RED-CNN's effectiveness is attributed to its ability to handle raw data without the need for vendor-specific sinogram domain filtration or iterative reconstruction algorithms, making it a promising solution for improving the quality of low-dose CT images.The paper introduces a novel deep learning approach, the Residual Encoder-Decoder Convolutional Neural Network (RED-CNN), for low-dose CT imaging. The RED-CNN combines autoencoders, deconvolution networks, and shortcut connections to effectively reduce noise, preserve structural details, and enhance lesion detection. The method is evaluated using both simulated and clinical datasets, demonstrating superior performance compared to state-of-the-art methods in terms of noise suppression, structural preservation, and lesion detection. The RED-CNN's effectiveness is attributed to its ability to handle raw data without the need for vendor-specific sinogram domain filtration or iterative reconstruction algorithms, making it a promising solution for improving the quality of low-dose CT images.