PURPOSE: To evaluate the significance of multislice CT for the diagnosis of uncertain penetrating globe injuries. MATERIALS AND METHODS: Based on a retrospective chart review between 2002 and 2007, we identified 59 patients presenting with severe ocular trauma with uncertain rupture of the globe due to massive subconjunctival and/or anterior chamber hemorrhage. The IOP (intraocular pressure) was within normal range in all patients. High resolution multidetector CT (MD-CT) scans (16 slice scans) with axial and coronar reconstructions were performed in all patients. The affected eye was examined for signs of penetrating injury such as abnormal eye shape, scleral irregularities, lens dislocation or intravitreal hemorrhages. Four experienced radiologists read the CT scans independently. Beside the diagnosis, the relevant morphological criteria and the optimal plane orientation (axial or coronar) were specified. The sensitivity, specificity, and negative and positive predictive value were calculated. Additionally the interobserver variability was determined by applying the Cohen's kappa test. Surgical sclera inspections were performed in all cases as a standard of reference. The evaluations of the CT examination were compared with the surgery reports. RESULTS: 59 patients were evaluated (42 men, 17 women). The mean age was 29 years (range 7 - 91). In 17 patients a rupture of the globe was diagnosed during surgery. 12 of these 17 penetrating injuries (70.6 %) were classified correctly by MDCT, 5 of the 17 (29.4 %) were not detectable. 42 patients did not have an open globe injury. 41 of these patients were diagnosed correctly negative by MDCT, and one patient was classified false positive. This results in a sensitivity of 70 % with a specificity of 98 %. There was high inter-rater agreement with kappa values between 0.89 - 0.96. Most discrepancies were caused by wrong negative findings. The most frequent morphologic criteria for open globe injury were the deformation (n = 10) and the volume reduction (n = 7) of the globe. These changes were most certain and visible in the axial orientation. CONCLUSION: In about one third of all unclear cases with rupture of the globe due to severe trauma, CT evaluation failed to correctly diagnose the open globe injury, which might lead to a delay in necessary surgical intervention. Thus, surgical sclera inspection is always mandatory in such unclear trauma cases regardless of the MD-CT results.
PURPOSE: To evaluate the significance of multislice CT for the diagnosis of uncertain penetrating globe injuries. MATERIALS AND METHODS: Based on a retrospective chart review between 2002 and 2007, we identified 59 patients presenting with severe ocular trauma with uncertain rupture of the globe due to massive subconjunctival and/or anterior chamber hemorrhage. The IOP (intraocular pressure) was within normal range in all patients. High resolution multidetector CT (MD-CT) scans (16 slice scans) with axial and coronar reconstructions were performed in all patients. The affected eye was examined for signs of penetrating injury such as abnormal eye shape, scleral irregularities, lens dislocation or intravitreal hemorrhages. Four experienced radiologists read the CT scans independently. Beside the diagnosis, the relevant morphological criteria and the optimal plane orientation (axial or coronar) were specified. The sensitivity, specificity, and negative and positive predictive value were calculated. Additionally the interobserver variability was determined by applying the Cohen's kappa test. Surgical sclera inspections were performed in all cases as a standard of reference. The evaluations of the CT examination were compared with the surgery reports. RESULTS: 59 patients were evaluated (42 men, 17 women). The mean age was 29 years (range 7 - 91). In 17 patients a rupture of the globe was diagnosed during surgery. 12 of these 17 penetrating injuries (70.6 %) were classified correctly by MDCT, 5 of the 17 (29.4 %) were not detectable. 42 patients did not have an open globe injury. 41 of these patients were diagnosed correctly negative by MDCT, and one patient was classified false positive. This results in a sensitivity of 70 % with a specificity of 98 %. There was high inter-rater agreement with kappa values between 0.89 - 0.96. Most discrepancies were caused by wrong negative findings. The most frequent morphologic criteria for open globe injury were the deformation (n = 10) and the volume reduction (n = 7) of the globe. These changes were most certain and visible in the axial orientation. CONCLUSION: In about one third of all unclear cases with rupture of the globe due to severe trauma, CT evaluation failed to correctly diagnose the open globe injury, which might lead to a delay in necessary surgical intervention. Thus, surgical sclera inspection is always mandatory in such unclear trauma cases regardless of the MD-CT results.