Management of Post-traumatic Aniridia with Retinal Detachment

Purpose: To reconstruct the anatomic and functional impairment in patients with post-traumatic aniridia, aphakia, and retinal detachment. Methods: Four patients with unilateral aniridia and aphakia as well as retinal detachment as results of severe eye injuries underwent scleral buckling, vitrectomy, membrane peeling, endolaser photocoagulation, silicone oil or gas temponade, combined with iris diaphragm-IOL implantation. Results: All four patients achieved successfully anatomic and functional reconstruction after surgery. During five to 22 months postoperative follow-up, all retinas remained attached. The final visual acuity increased from finger counting to 0.1 ～ 0.3. Conclusions: The combination of vitreoretinal surgery and iris diaphrgm-IOL implantation is an effective method for post-traumatic aniridia, aphakia and traumatic retinal detachment. It could ameliorate photophobia and improve the biocular vision.Furthermore, artificial iris diaphragm implantation could prevent silicone oil-endothelia contact and salvage silicone keratopathy. Eye Science 2001; 17:35 ～ 38.

Clinical Observation on Preservation of Lens Anterior Capsule in Reduction of Complications Associated with Silicone Oil

Purpose: To observe the effect of preservation of anterior capsule on the incidence of complications associated with silicone oil. Methods: 82 patients(82 eyes) accepted trans pars plana vitrectomy(TPPV) combined with lensectomy of whom , 30 eyes with preservation of anterior capsule(PAC) and 52eyes with no preservation of anterior capsule (NPAC). Analysis of the incidence of complications associated with use of oil. Results:The incidence rate was 50.0 % in NPAC group, and 23.3 % in PAC group( P ＜ 0. 025 ). There were seconda'y glaucoma (21. 1% ), band keratopathy ( 13.5 % ) and corneal decompensation(9.6 % ) in NPAC group, while there was none of them in PAC group. Conclusions: Preservation of anterior capsule is an eftective method to reduce the complications associated with silicone oil. Eye Science 2001; 17:39 ～ 41.

基于三维限制区域生长的SD-OCT中浆NRD病变区域分割

中浆(CSC)病变区域的大小对于病变的诊断及研究有着关键的作用,而视网膜神经上皮层脱离(NRD)形态在中浆病变中最为普遍且病变程度最为严重,因此快速准确地分割出NRD病变区域十分重要。给出一种全自动的频域光学相干断层(SD-OCT)中浆NRD病变分割方法。首次在三维空间进行NRD病变分割,将二维图像上的病变区域分割问题转化为三维空间的体分割问题,充分利用了数据的三维结构信息,提高了分割精度。18组中浆NRD病变的SD-OCT图像的实验结果表明:该算法能够准确分割出中浆NRD病变,且平均覆盖率高达89.5%。与其他4种分割方法相比,所提方法精度最高且耗时最短,在临床应用与研究中具有极大的优势。

基于深度学习的眼底病变区域自动分割

基于光学相关层析成像(OCT),准确分割出视网膜液体区域相关异常和视网膜色素上皮分离对眼底疾病诊断具有重要意义。论文提出一种基于深度学习的分割方法,实现对视网膜色素上皮脱离(PED)、视网膜下液体(SRF)和视网膜水肿区域(REA)等病变类型区域的自动分割。首先使用迁移学习模型InceptionV4对所有病变类型进行分类,然后构建生成对抗网络对每种病变类型区域进行自动分割。实验结果表明,该方法Dice相似性系数约为0.75,灵敏度约为0.95,特异性约为0.96,召回率约0.98,整体优于其他方法。