An important technical issue in head and neck PET/CT is the identification of aneurysms, cancer, and tumor. In this paper, we propose the use of the Convolutional Neural Network (CNN) approach to this problem with an aim of automating the process of identifying the above-mentioned conditions. The following are the strategies used in our approach image preprocessing to improve image quality, data augmentation to increase the dataset volume, and the utilization of modern CNN’s architectures to improve the detection performance. Performance evaluation performed for the proposed model shows improvement in terms of precision, recall, and overall accuracy. Our model reached an accuracy of 94% thereby raising the diagnostic accuracy and affording an efficient solution to enhancing the actual utilization of medical images in the observation and treatment of diseases at large scale. Such a result should be useful in practical real-world applications.