Convolutional Neural Networks Applied to the Diagnosis of Cardiovascular Disease

Cardiovascular disease (CVD) is a major global health problem that poses a significant threat to public health worldwide. According to the World Heart Federation (WHF), deaths from CVD have increased by 60% globally over the past 30 years, from 12.1 million in 1990 to 20.5 million in 2021. This follows a deep learning approach to enable early prediction and reduce their incidence. Three different models are proposed and compared: Deep Neural Network (DNN), Convolutional Neural Network (CNN), and Multi-Layer Perceptron (MLP). Experiments are conducted in two scenarios using a dataset of 918 clinical records, and another in which the variables were reduced to the most significantly correlated with CVD. The CNN trained on the original dataset showed the best performance among the proposed models. Specifically, the evaluation metrics of the CNN model were as follows: an accuracy of 92.17%, a sensitivity of 94.51%, a specificity of 90.78%, an F1 score of 93.30%, and an area under the ROC curve of 90.03%. In addition, this model was implemented as an application programming interface (API) using FastAPI, which facilitated its integration into a prototype web application developed in TypeScript.