JAMA Ophthalmology Published online May 25, 2023
這項研究旨在研究糖尿病黃斑缺血（diabetic macula ischemia, DMI）在預測糖尿病視網膜病變（DR）進展、糖尿病性黃斑水腫（DME）發展和視力（VA）惡化的價值。研究人員使用基於光學相干斷層掃描血管影像學（OCTA）圖像的自動二值化(automated binary) DMI算法。他們使用深度學習算法分析了表層毛細血管網（superficial capillary plexus, SCP）和深層毛細血管網（deep capillary plexus, DCP）的OCTA圖像。
The study aimed to investigate the prognostic value of diabetic macular ischemia (DMI) in diabetic retinopathy (DR) progression, diabetic macular edema (DME) development, and visual acuity (VA) deterioration in patients with diabetes. The researchers used an automated binary DMI algorithm based on optical coherence tomography angiography (OCTA) images. They analyzed OCTA images of the superficial capillary plexus (SCP) and deep capillary plexus (DCP) using a deep learning algorithm.
The study included a total of 321 eyes from 178 patients. The presence of SCP-DMI and DCP-DMI at baseline was significantly associated with DR progression. DCP-DMI was also associated with DME development and VA deterioration after adjusting for various factors. The researchers found that the presence of DMI in both SCP and DCP improved the predictive discrimination for DR progression.
The study concluded that the presence of DMI on OCTA images, particularly in the DCP, provided prognostic value for DR progression, DME development, and VA deterioration in patients with diabetes.