The choroidal vascular index (CVI) is the ratio of the luminal area to the total choroidal area. It can not only reflect the changes in the vascular composition of the choroid, but also serve as an observation index for follow-up treatment effects. CVI is a new biometric tool, which is gradually applied to the observation of choroidal structure in various eye diseases. It has great application prospects in the study of pathophysiological mechanisms, disease process monitoring and efficacy evaluation such as central serous chorioretinopathy, polypoid choroidal vascular disease, age-related macular degeneration, diabetic retinopathy,etc. Understanding the research progress of CVI in various eye diseases can provide reference for clinical research of CVI.
ObjectiveTo observe and preliminarily discuss the distribution characteristics of the non-perfusion area (NP) of the retina in different stages of diabetic retinopathy (DR) and its changes with the progression of DR. MethodsA retrospective clinical study. From October 2018 to December 2020, 118 cases of 175 eyes of DR patients diagnosed in Eye Center of Renmin Hospital of Wuhan University were included in the study. Among them, there were 64 males with 93 eyes and 54 females with 82 eyes; the average age was 56.61±8.99 years old. There were 95 eyes of non-proliferative DR (NPDR), of which 25, 47, and 23 eyes were mild, moderate, and severe; 80 eyes were proliferative DR (PDR). Ultra-wide-angle fluorescein fundus angiography was performed with the British Optos 200Tx imaging system, and the fundus image was divided into posterior, middle, and distal parts with Image J software, and the ischemic index (ISI) was calculated. The difference of the retina in different DR staging groups and the difference of ISI were compared in the same area. The Kruskal-Wallis test was used to compare the ISI between the different DR staging groups and the Kruskal-Wallis one-way analysis of variance was used for the pairwise comparison between the groups. ResultsThe ISI of the posterior pole of the eyes in the moderate NPDR group, severe NPDR group, and PDR group were significantly greater than that in the distal periphery, and the difference was statistically significant (χ2=6.551, 3.540, 6.614; P=0.000, 0.002, 0.000). In severe NPDR group and PDR group, the ISI of the middle and peripheral parts of the eyes was significantly greater than that of the distal parts, and the difference was statistically significant (χ2=3.027, 3.429; P=0.015, 0.004). In the moderate NPDR group, there was no significant difference in ISI between the peripheral and distal parts of the eye (χ2=2.597, P=0.057). The ISI of the posterior pole of the eyes in the moderate NPDR group and the PDR group was significantly greater than that in the middle periphery, and the difference was statistically significant (χ2=3.955, 3.184; P=0.000, 0.009). In the severe NPDR group, there was no significant difference in ISI between the posterior pole and the middle periphery of the eye (χ2=0.514, P=1.000). Compared with the mild NPDR group and the moderate NPDR group, the ISI of the whole retina, posterior pole, middle and distal parts of the PDR group was larger, and the difference was statistically significant (χ2=－7.064, －6.349,－6.999, －5.869, －6.695, －6.723, －3.459, －4.098; P=0.000, 0.000, 0.000, 0.000, 0.000, 0.000, 0.003, 0.000). ConclusionThe NP of the eyes with different DR stages is mainly distributed in the posterior pole and the middle periphery. The higher the severity of DR, the greater the NP in the posterior and middle periphery.
ObjectiveTo build a small-sample ultra-widefield fundus images (UWFI) multi-disease classification artificial intelligence model, and initially explore the ability of artificial intelligence to classify UWFI multi-disease tasks. MethodsA retrospective study. From 2016 to 2021, 1 608 images from 1 123 patients who attended the Eye Center of the Renmin Hospital of Wuhan University and underwent UWFI examination were used for UWFI multi-disease classification artificial intelligence model construction. Among them, 320, 330, 319, 268, and 371 images were used for diabetic retinopathy (DR), retinal vein occlusion (RVO), pathological myopia (PM), retinal detachment (RD), and normal fundus images, respectively. 135 images from 106 patients at the Tianjin Medical University Eye Hospital were used as the external test set. EfficientNet-B7 was selected as the backbone network for classification analysis of the included UWFI images. The performance of the UWFI multi-task classification model was assessed using the receiver operating characteristic curve, area under the curve (AUC), sensitivity, specificity, and accuracy. All data were expressed using numerical values and 95% confidence intervals (CI). The datasets were trained on the network models ResNet50 and ResNet101 and tested on an external test set to compare and observe the performance of EfficientNet with the 2 models mentioned above. ResultsThe overall classification accuracy of the UWFI multi-disease classification artificial intelligence model on the internal and external test sets was 92.57% (95%CI 91.13%-92.92%) and 88.89% (95%CI 88.11%-90.02%), respectively. These were 96.62% and 92.59% for normal fundus, 95.95% and 95.56% for DR, 96.62% and 98.52% for RVO, 98.65% and 97.04% for PM, and 97.30% and 94.07% for RD, respectively. The mean AUC on the internal and external test sets was 0.993 and 0.983, respectively, with 0.994 and 0.939 for normal fundus, 0.999 and 0.995 for DR, 0.985 and 1.000 for RVO, 0.991 and 0.993 for PM and 0.995 and 0.990 for RD, respectively. EfficientNet performed better than the ResNet50 and ResNet101 models on both the internal and external test sets. ConclusionThe preliminary UWFI multi-disease classification artificial intelligence model using small samples constructed in this study is able to achieve a high accuracy rate, and the model may have some value in assisting clinical screening and diagnosis.