Computed tomography in the diagnosis and prognosis of open-globe injuries

OBJECTIVE: Determine sensitivity, specificity, and prognostic signs of orbital and ocular computed tomography (CT) in diagnosing patients with open globe injury. DESIGN: Randomized masked review of computed tomograms and retrospective clinical correlation of patients with ocular trauma. PARTICIPANTS: Two hundred patients who underwent CT evaluation for ocular trauma between 1989 and 1993. METHODS: CTs were read by three masked observers; findings were tabulated and compared for variability among observers; sensitivity and specificity were calculated and CT findings were grouped according to visual outcome retrieved from record review. RESULTS: In the absence of clinical information, sensitivity and specificity were 75% and 93%, respectively. The positive predictive value ranged from 88% to 97%, with a calculated overall positive predictive value of 95%. Patients who had a poor visual outcome (visual acuity <2/200) or who underwent enucleation had significantly more CT findings than patients with a good visual outcome. Vitreous hemorrhage, absence of lens, and severe distortion of vitreous space are among the most common CT findings associated with poor visual outcome. CONCLUSIONS: CT is not sensitive enough to be solely relied upon for diagnosis of all open globe injuries. CT findings only complement clinical findings, increasing the clinician's overall ability to make an accurate diagnosis of open globe injury, and may provide useful prognostic information regarding visual outcome.