ABSTRACT
BACKGROUND: Computed tomography (CT) is the standard for grading blunt splenic injuries, but the true accuracy, especially for grade IV or V injuries as compared to pathological findings, is unknown. STUDY DESIGN: A retrospective study from 2005 to 2011 was undertaken. RESULTS: There were 214 adults admitted with blunt splenic injury and 170 (79%) were managed nonoperatively. The remaining 44 patients (21%) required surgical intervention. There was a significant difference in the Injury Severity Score (ISS) between those who did and those who did not require splenectomy: median 31 (interquartile [IQ] range 11-51) versus 22 (IQ range 9-35, p = 0.0002). Ten patients presented in shock, had a positive ultrasound, and went to surgery. The remaining 34 had CT scans prior to surgery. Twenty-five (73%) had injury grades IV or V. The CT scan correctly graded the injury in 14 (41%) and was incorrect in 20 (59%). The assigned grade by the CT scan underestimated the true injury grade by one grade in six cases (30%), by two or more grades in nine (45%), and the CT images were obscured by blood and deemed "ungradeable" in five (25%). The CT scan was more accurate for grades I and II (100%) than for grades III-V (25-43%). The reasons for inaccuracy were either inability to visualize that the laceration involved the hilar vessels or excessive perisplenic blood which obscured the injury and/or the hilum. CONCLUSIONS: CT for splenic injury is accurate for grades I and II, but underestimates the true extent of injury for grades III-V. The reasons for the lack of correlation are the inability to determine hilar involvement and excessive perisplenic blood obscuring the injury. Patients with these image characteristics by CT scan should undergo splenectomy earlier if there are any signs of hemodynamic instability.
