Abstract
Abstract: :
Purpose: This study is aimed to find out how closely the preoperative diagnosis of retinal detachment(RD) by ultrasonography(U/S) matched the intraoperative findings in cases of vitreous hemorrhage(V–HEMO). Methods: A chart review of 81 patients (86 eyes) who have undergone pars plana vitrectomy in treatment of V–HEMO between June 1996 and June 2004. False–positive, false–negative, sensitivity and specificity for U/S were calculated from cases which U/S findings lead to diagnosis of RD and from cases which intraoperative findings actually proved RD. Results: The causes of V–HEMO were proliferative diabetic retinopathy(39 eyes, 45%), ocular trauma(22 eyes, 26%), posterior vitreous detachment with or without tear(12 eyes, 14%), retinal vascular occlusion(7eyes, 8%), and post–intraocular surgery(5 eyes, 6%). The false–positive rate for RD was 33% (8 of 24 eyes). The false–negative rate for RD was 10% (6 of 62 eyes). The sensitivity of U/S was calculated 73% and the specificity was calculated 90%. U/S within the PDR group has false–positive rate of 38%, false–negative rate of 19%, sensitivity of 62% and specificity of 81%. Such findings from the group excluded from the PDR group showed false–positive rate of 27%, false–negative rate of 3%, sensitivity of 89% and specificity of 92%. Conclusions: U/S for patients with PDR were proven to be rather inaccurate, with 62% sensitivity and 81% specificity ; while 89%, 92% respectively in non PDR group. It has been reported that U/S is an effective diagnostic tool in patients with media opacity but physician should bear in mind that U/S can be unreliable in some patients with PDR and thus the physician should have a care in making diagnosis.
Keywords: imaging methods (CT, FA, ICG, MRI, OCT, RTA, SLO, ultrasound) • retinal detachment • diabetic retinopathy