INTRODUCTION: Little is known about the natural history of clot resolution in the initial weeks of anticoagulant therapy in patients with acute pulmonary embolism (PE). Clot resolution of acute pulmonary embolism (PE) was assessed with either computed tomography pulmonary angiography (CT-scan) or perfusion scintigraphy (Q-scan) after 3 weeks of treatment. METHODS: This was a predefined safety analysis of the Einstein PE study, including PE patients, randomized to either enoxaparin with vitamin K antagonists or rivaroxaban. A similar scan as at baseline was repeated after 3 weeks. The percentage of pulmonary vascular obstruction (PVO) was calculated based on a weighted semi-quantitative estimation of obstruction. Clot resolution was assessed blindly by calculating the relative change after 3 weeks. RESULTS: PE was diagnosed in 264 patients with CT-scan and in 83 with Q-scan. Baseline characteristics were comparable. At baseline, the mean PVO assessed with CT-scan (PVO-CT) and with Q-scan (PVO-Q) were both 21% (Standard deviation (SD) 13%) (p = 0.9). The mean relative decrease in PVO was 71% (SD) 33%) for PVO-CT, compared to 62% (SD 36%) for PVO-Q (p=0.02), while complete resolution was observed in 44% (116/264; 95% CI 38-50%) and 31% (26/83; 95% CI 22-42%) with of CT-scan and Q-scan, respectively (p=0.04). No difference in clot resolution between enoxaparin/VKA and rivaroxaban was found. CONCLUSION: In patients with acute PE, only 3 weeks of anticoagulant treatment leads to complete clot resolution in a considerable proportion of patients and normalization is more often observed with CT-scan than with Q-scan. © 2013 International Society on Thrombosis and Haemostasis.