Missed intracranial hemorrhage

Imaging description Traumatic head injuries may result in intraparenchymal, intraventricular, subarachnoid, subdural, or epidural hemorrhage. Acute hemorrhage is characterized by hyperattenuation on CT, and the classic appearances of the various types of hemorrhage are well known. However, certain types of hemorrhage may be overlooked, especially subdural and subarachnoid hemorrhages. Images from a head CT are routinely reviewed in the axial plane. However, important findings may be missed on axial images alone. In particular, hemorrhages oriented in a horizontal plane are prone to volume-averaging effects which may result in false-negative results. This is especially true of hemorrhages which occur adjacent to bone, such as the floor of the anterior and middle cranial fossae, where volume-averaging with adjacent bone leads to decreased detection (Figure 3.1). This issue is compounded by the fact that hemorrhages have a tendency to occur adjacent to bony structures in certain mechanisms of injury [1]. The addition of coronal and sagittal reformations may improve the diagnostic accuracy by reducing both false-negative and false-positive results (Figure 3.2). A study of 109 patients with intracranial hemorrhage found that the addition of coronal reformations resulted in a change in interpretation in approximately 25% of cases, compared with axial images alone [2]. Another cause of missed hemorrhage involves the use of inappropriate window and level values (Figure 3.3). If the window is too narrow, a small subdural hemorrhage may be difficult to distinguish from the adjacent bone. Optimal values for the window and level will vary among scanners, but a reasonable starting point may be a window of 200 and a level of 50.