A method for automatically identifying a potential pleural effusion in medical image data of a thorax of a patient from a scan by means of a medical scanner is provided. It includes at least the steps of accepting rib cage detection data of the rib cage of the patient from the image data, which rib cage detection data include rib cage extent data of a rib cage extent of the interior of the rib cage, accepting lung detection data of the lung of the patient from the image data, which lung detection data comprise lung extent data of a lung extent of the external boundary of the lung, accepting mediastinum detection data of all organs of the mediastinum in the thorax (Th) of the patient from the image data, which mediastinum detection data comprise mediastinum extent data of a mediastinum extent of the external boundary of the mediastinum, and subtracting the lung extent and the mediastinum extent from the rib cage extent while forming pleural effusion identification data.