Fig. 19

Zebra retreat and hindsight bias. A 75-year-old male patient with known tracheoesophageal fistula (green arrows, d, e) suddenly lost consciousness. Brain CT and CT angiography revealed edematous brain tissue, diffuse air within brain vasculature (b), and contrast filling limited to just proximal segments of the left middle cerebral artery and anterior cerebral artery (red arrows, a). Recent studies of the patient were re-evaluated to detect the etiology of cerebral air embolism. On thorax CT taken 3 days previously for hemoptysis, there was seen to be parenchymal ground glass opacities suggesting hemorrhage (c). In the mediastinal window, next to the fistula, there was an outpouching of the pulmonary vein (blue arrow, d) that had not been present on the previous study (e). The on-call radiologist had also noticed this newly developed outpouching but did not feel confident enough to mention this finding in the report due to the absence of apparent contrast extravasation into the tracheal lumen. Moreover, the radiologist had thought that hemoptysis originating from a pulmonary vein was quite an unexpected situation. This phenomenon is called a “zebra retreat.” It seems to be impossible to overlook or underrate that outpouching with “a priori knowledge of diffuse cerebral air embolism” due to the hindsight bias during retrospective evaluation