Pulmonary thromboembolism is considered one of the most relevant cardiovascular conditions and is also one of the main causes of mortality worldwide. This pathology stands out among cardiovascular medical emergencies and is one of the most frequent causes of death related to the cardiovascular system. Pulmonary thromboembolism diagnosis is normally performed using computed tomography angiography; however, interpreting such images by expert radiologists is time-consuming and may be vulnerable to interobserver variability and human errors. In addition, a late diagnosis can be critical for a patient. Computed tomography angiography images are interpreted by experts in this domain, which is subjective and based on professional experience. The current study is focused on developing and validating an artificial intelligence neural network model, using convolutional neural networks, to support general physicians, internists, and other related specialties in the early diagnosis of pulmonary thromboembolism, using tomography. The challenge is to enhance the accuracy and efficiency of diagnosis while reducing the demand for extensive manual interpretations in advanced deep learning techniques and uniformly classifying this pathology. In addition, we evaluated the effectiveness of treatment with a more accurate and early detection of pulmonary thromboembolism. The proposed future work aims to investigate and develop the convolutional neural network model for the automatic diagnosis of pulmonary embolism, furthering the advancement of artificial intelligence in medicine. It will hopefully eliminate several manual interpretations, thus giving the right attention and treatment in a timely and efficient way for patients.