Recovery of lung perfusion after pulmonary embolism (PE) is conditioned by several factors. The aim of this study was to analyze the differences of reperfusion after PE between patients with and without preexisting cardiopulmonary disease, and to develop a mathematical model to predict, at diagnosis, the size of defects in lung perfusion scan at 6 months after treatment. We included 83 patients with diagnosis of PE in a non-concurrent cohort study (cohort I: 26 with preexisting cardiopulmonary disease, and cohort II: 57 without previous cardiopulmonary disease). Lung perfusion scan was performed at diagnosis, 7-10 days after treatment and at 6 months. The defect size was quantified following a numerical score. The recovery of perfusion after 7-10 days was 33% in cohort I and 45% in cohort II, and 50% and 72%, respectively, at the last control at 6 months. A multiple-regression analysis was performed using the final size of defects at 6 months (y) as the dependent variable, and the defect size at diagnosis (x) and the presence of preexisting cardiopulmonary disease (z) as independent variables. The regression equation was y = 1.29 + 0.15x + 2.98z. We conclude that: (1) in patients with cardiopulmonary diseases, the remaining defects at 6 months were larger; (2) at diagnosis, using a mathematical model, it is possible to predict the size of the defects in lung perfusion scan at 6 months in patients with or without previous cardiopulmonary disease.