The contents of chronic subdural hematoma and its CT density, with special reference to differentiation from subdural hygroma

Abstract

Diagnosis of posttraumatic subdural low density areas in computerized tomography (CT) has been quite controversial. Laboratory and in vivo CT analysis using the ACTA 0200FS were performed on 19 chronic subdural hematomas and five subdural hygromas. CT densities of the hematomas and of the ventricular CSF were obtained through regions of interest from preoperative plain CT scans. In six hematomas, both the supernatant and precipitant parts were separately sampled and the corresponding regions of interest were placed accordingly on the CT image. In these 25 hematoma samples, red blood cells (RBC), hematocrit, and hemoglobin (Hgb) varied greatly, though, these values correlated well with the CT densities. Plasma protein content was fairly constant with an average of 7.1±0.8 g/dl. Iron content was comparatively high, but still at levels negligible for CT density. There were four hematoma samples with RBC of less than 20×104 μl or Hgb of less than 2.0 g/dl. Their CT values ranged between 18 and 23 H.U., which were considered close to the in vivo serum level CT density. Five hygroma fluid showed no RBC and very little protein content of less than 0.4 g/dl. CT density ranged between -2 and 13 H.U.. One hygroma was so thin that placement of the region of interest was difficult. The edge effect of the skull was experimentally studied using a phantom skull filled with water. This revealed a remarkable overshoot of the CT values within ten pixels from the inner wall of the skull. Visual observation of the original CT pictures revealed four low density hematomas and seven mixed density ones. When compared to the density of the ventricular cavity, all of the low density hematomas and the supernatant part of the mixed density ones were clearly higher in density. All five hygromas appeared CSF dense or lower. Thus, visual observation offered clear distinction between hematoma and hygroma. In conclusion, because of the edge effect by the skull, thin subdural fluids could not be diagnosed by CT alone. Thick subdural fluids could be differentiated as either hematoma or hygroma by their CT densities. Subdural hematomas had in vivo CT densities of at least serum level or approximately 20 H.U., while subdural hygromas had densities close to CSF. These characteristics were best appreciated by visual observation of the CT scan films.