In recent times, computational studies have emerged as a viable alternative for complementing the efforts of experienced radiologists in disease diagnosis. Computed tomography (CT) studies are a common way of predicting the lung nodule malignancy for the early diagnosis and treatment of lung cancer in patients. Early detection of the type of nodule is the key to determining the appropriate treatment, thus increasing patient survival. Feature extraction is an important stage in classifying benign and malignant nodules in chest CT scans. However, determining the type of nodule in CT scans is a challeging task in medical imaging, since CT images cannot be evaluated as an average or generic image. Hence, this study is based on the application of machine learning techniques, specifically convolutional neural networks (CNNs) and transfer learning, for the feature extraction and identification of tomograms with pulmonary nodules on a public, Lung TIME dataset. Pretrained CNNs architectures (VGG, ResNet, MobileNet, Xception, NASNet and DenseNet) and a proposed CNN architecture were used, thereby obtaining a minimum training accuracy in pretrained architectures of 70.11% and minimum test accuracy of 33.91%. In contrast, in the proposed CNN, 95.25% and 94.21% respectively were obtained. These results show that the transfer learning is not always feasible in medical applications and architectures focused on the problem to be solved are usually most effective.