Machine Learning and Optical-Coherence-Tomography-Derived Radiomics Analysis to Predict the Postoperative Anatomical Outcome of Full-Thickness Macular Hole.

Full-thickness macular hole (FTMH) leads to central vision loss. It is essential to identify patients with FTMH at high risk of postoperative failure accurately to achieve anatomical closure. This study aimed to construct a predictive model for the anatomical outcome of FTMH after surgery. A retrospective study was performed, analyzing 200 eyes from 197 patients diagnosed with FTMH. Radiomics features were extracted from optical coherence tomography (OCT) images. Logistic regression, support vector machine (SVM), and backpropagation neural network (BPNN) classifiers were trained and evaluated. Decision curve analysis and survival analysis were performed to assess the clinical implications. Sensitivity, specificity, F1 score, and area under the receiver operating characteristic curve (AUC) were calculated to assess the model effectiveness. In the training set, the AUC values were 0.998, 0.988, and 0.995, respectively. In the test set, the AUC values were 0.941, 0.943, and 0.968, respectively. The OCT-omics scores were significantly higher in the "Open" group than in the "Closed" group and were positively correlated with the minimum diameter (MIN) and base diameter (BASE) of FTMH. Therefore, in this study, we developed a model with robust discriminative ability to predict the postoperative anatomical outcome of FTMH.

Machine learning and optical coherence tomography-derived radiomics analysis to predict persistent diabetic macular edema in patients undergoing anti-VEGF intravitreal therapy.

BACKGROUND: Diabetic macular edema (DME) is a leading cause of vision loss in patients with diabetes. This study aimed to develop and evaluate an OCT-omics prediction model for assessing anti-vascular endothelial growth factor (VEGF) treatment response in patients with DME. METHODS: A retrospective analysis of 113 eyes from 82 patients with DME was conducted. Comprehensive feature engineering was applied to clinical and optical coherence tomography (OCT) data. Logistic regression, support vector machine (SVM), and backpropagation neural network (BPNN) classifiers were trained using a training set of 79 eyes, and evaluated on a test set of 34 eyes. Clinical implications of the OCT-omics prediction model were assessed by decision curve analysis. Performance metrics (sensitivity, specificity, F1 score, and AUC) were calculated. RESULTS: The logistic, SVM, and BPNN classifiers demonstrated robust discriminative abilities in both the training and test sets. In the training set, the logistic classifier achieved a sensitivity of 0.904, specificity of 0.741, F1 score of 0.887, and AUC of 0.910. The SVM classifier showed a sensitivity of 0.923, specificity of 0.667, F1 score of 0.881, and AUC of 0.897. The BPNN classifier exhibited a sensitivity of 0.962, specificity of 0.926, F1 score of 0.962, and AUC of 0.982. Similar discriminative capabilities were maintained in the test set. The OCT-omics scores were significantly higher in the non-persistent DME group than in the persistent DME group (p < 0.001). OCT-omics scores were also positively correlated with the rate of decline in central subfield thickness after treatment (Pearson's R = 0.44, p < 0.001). CONCLUSION: The developed OCT-omics model accurately assesses anti-VEGF treatment response in DME patients. The model's robust performance and clinical implications highlight its utility as a non-invasive tool for personalized treatment prediction and retinal pathology assessment.

Proportion and risk factors of diabetic retinopathy by stage in less-developed rural areas of Hunan province of China: A multi-site cross-sectional study.

AIMS: To investigate the proportion and risk factors of diabetic retinopathy (DR) by stages in less-developed rural areas in Hunan Province of China. BACKGROUND: DR is common among people with diabetes but not well recognized in less-developed rural areas. There is insufficient evidence on the risk factors of DR by stages, making it challenging to develop targeted prevention and intervention programs for DR in primary care settings. METHODS: A multi-site cross-sectional survey was conducted among people with type 2 diabetes mellitus (T2DM) from four less-developed counties in Hunan Province of China. All participants underwent the screening of DR via digital fundus photography and completed self-reported questionnaires on their socio-demographic and clinical characteristics, diabetes self-efficacy, diabetes self-care behaviors, social support, family function, and health service accessibility. The multinomial logistic regression models were employed to explore the risk factors of DR by stage, which were selected based on the socio-ecological model, literature, and clinical experience. RESULTS: A total of 196 participants were included in this study with an average age of 57.43 ± 10.26. 59.6% (117/196) of the participants were identified as DR, including 37.2% (73/196) non-proliferative DR and 22.4% (44/196) proliferative DR. Compared to the non-DR group, the risk factors of non-proliferative DR and proliferative DR were diabetes duration (OR: 1.10, 95 CI%: 1.04-1.17; OR: 1.14, 95 CI% 1.06-1.22) and self-monitoring of blood glucose (OR: 1.09, 95 CI% 1.01-1.17; OR: 1.11, 95 CI%: 1.02-1.20); the protective factors of non-proliferative DR was accessible complication prevention and management education (OR: 0.37, 95 CI% 0.14-0.94) while the protective factors of proliferative DR were physical activities (OR: 0.89, 95 CI%: 0.80-0.98). Compared to the non-proliferative DR group, the protective factors of proliferative DR were physical activities (OR: 0.89, 95 CI% 0.02-0.89) and family function (OR: 0.84, 95 CI%: 0.04-0.84). CONCLUSION: DR was prevalent among people with T2DM in less-developed rural areas, indicating the need of strengthening DR screening. Risk factors of DR varied by stage while sharing some common factors. Future DR prevention and intervention programs may benefit from improving these factors to reduce the risk of DR by stage.

Effect of High Myopia on Delayed Absorption of Subretinal Fluid after Scleral Buckling Surgery.

This study compared the absorption of subretinal fluid (SRF) in patients with rhegmatogenous retinal detachment (RRD) with and without high myopia after scleral buckling (SB) and investigated the effect of high myopia on SRF absorption. This retrospective study included patients with primary macula-off RRD grouped according to myopia and age. The optical coherence tomography (OCT) and OCT angiography indicators included subretinal fluid height (SRFH), subfoveal choroidal thickness (SFCT), and choroidal capillary blood flow density (CCFD) measured regularly. The presence of SRF 3 months after surgery was defined as delayed absorption. Overall, 90 eyes of 89 patients were enrolled, and 46 eyes (51.11%) had high myopia. In 43 eyes (47.78%), SRF absorption was delayed. There was no significant difference in SRF absorption after SB between the high and non-high myopia groups; younger patients (<35 years) had a higher probability of delayed absorption (p < 0.05). The SFCT in high myopia was significantly thinner than that in the non-high myopia group (p < 0.05); SFCT and SRFH were positively correlated (rs = 0.275, p = 0.002), and there was a significant difference between the average CCFD with and without SRF (p < 0.05). High myopia had no significant effect on SRF absorption after SB.

Exploring the Immune Infiltration Landscape and M2 Macrophage-Related Biomarkers of Proliferative Diabetic Retinopathy.

Backgrounds: Diabetic retinopathy (DR), especially proliferative diabetic retinopathy (PDR), is the major cause of irreversible blindness in the working-age population. Increasing evidence indicates that immune cells and the inflammatory microenvironment play an important role during PDR development. Herein, we aim to explore the immune landscape of PDR and then identify potential biomarkers correlated with specific infiltrating immune cells. Methods: We mined and re-analyzed PDR-related datasets from the Gene Expression Omnibus (GEO) database. Using the cell-type identification by estimating relative subsets of RNA transcripts (CIBERSORT) algorithm, we investigated the infiltration of 22 types of immune cells in all selected samples; analyses of differences and correlations between infiltrating cells were used to reveal the immune landscape of PDR. Thereafter, weighted gene co-expression network analysis (WGCNA) and differential expression analysis were applied to identify the hub genes on M2 macrophages that may affect PDR progression. Results: Significant differences were found between infiltration levels of immune cells in fibrovascular membranes (FVMs) from PDR and normal retinas. The percentages of follicular helper T cells, M1 macrophages, and M2 macrophages were increased significantly in FVMs. Integrative analysis combining the differential expression and co-expression revealed the M2 macrophage-related hub genes in PDR. Among these, COL5A2, CALD1, COL6A3, CORO1C, and CALU showed increased expression in FVM and may be potential biomarkers for PDR. Conclusions: Our findings provide novel insights into the immune mechanisms involved in PDR. COL5A2, CALD1, COL6A3, CORO1C, and CALU are M2 macrophage-related biomarkers, further study of these genes could inform novel ideas and basis for the understanding of disease progression and targeted treatment of PDR.

Application of Improved U-Net Convolutional Neural Network for Automatic Quantification of the Foveal Avascular Zone in Diabetic Macular Ischemia.

Objectives: The foveal avascular zone (FAZ) is a biomarker for quantifying diabetic macular ischemia (DMI), to automate the identification and quantification of the FAZ in DMI, using an improved U-Net convolutional neural network (CNN) and to establish a CNN model based on optical coherence tomography angiography (OCTA) images for the same purpose. Methods: The FAZ boundaries on the full-thickness retina of 6 × 6 mm en face OCTA images of DMI and normal eyes were manually marked. Seventy percent of OCTA images were used as the training set, and ten percent of these images were used as the validation set to train the improved U-Net CNN with two attention modules. Finally, twenty percent of the OCTA images were used as the test set to evaluate the accuracy of this model relative to that of the baseline U-Net model. This model was then applied to the public data set sFAZ to compare its effectiveness with existing models at identifying and quantifying the FAZ area. Results: This study included 110 OCTA images. The Dice score of the FAZ area predicted by the proposed method was 0.949, the Jaccard index was 0.912, and the area correlation coefficient was 0.996. The corresponding values for the baseline U-Net were 0.940, 0.898, and 0.995, respectively, and those based on the description data set sFAZ were 0.983, 0.968, and 0.950, respectively, which were better than those previously reported based on this data set. Conclusions: The improved U-Net CNN was more accurate at automatically measuring the FAZ area on the OCTA images than the traditional CNN. The present model may measure the DMI index more accurately, thereby assisting in the diagnosis and prognosis of retinal vascular diseases such as diabetic retinopathy.

PRRX1 Is a Novel Prognostic Biomarker and Facilitates Tumor Progression Through Epithelial-Mesenchymal Transition in Uveal Melanoma.

Uveal melanoma (UM) is the most common primary intraocular malignancy in adults. UM develops and is sustained by inflammation and immunosuppression from the tumor microenvironment (TME). This study sought to identify a reliable TME-related biomarker that could provide survival prediction and new insight into therapy for UM patients. Based on clinical characteristics and the RNA-seq transcriptome data of 80 samples from The Cancer Genome Atlas (TCGA) database, PRRX1 as a TME- and prognosis-related gene was identified using the ESTIMATE algorithm and the LASSO-Cox regression model. A prognostic model based on PRRX1 was constructed and validated with a Gene Expression Omnibus (GEO) dataset of 63 samples. High PRRX1 expression was associated with poorer overall survival (OS) and metastasis-free survival (MFS) in UM patients. Comprehensive results of the prognostic analysis showed that PRRX1 was an independent and reliable predictor of UM. Then the results of immunological characteristics demonstrated that higher expression of PRRX1 was accompanied by higher expression of immune checkpoint genes, lower tumor mutation burden (TMB), and greater tumor cell infiltration into the TME. Gene set enrichment analysis (GSEA) showed that high PRRX1 expression correlated with angiogenesis, epithelial-mesenchymal transition (EMT), and inflammation. Furthermore, downregulation of PRRX1 weakened the process of EMT, reduced cell invasion and migration of human UM cell line MuM-2B in vitro. Taken together, these findings indicated that increased PRRX1 expression is independently a prognostic factor of poorer OS and MFS in patients with UM, and that PRRX1 promotes malignant progression of UM by facilitating EMT, suggesting that PRRX1 may be a potential target for UM therapy.

Microvascular Structure Changes After Intravitreal Ranibizumab Injection in Retinal Vein Occlusion Patients With and Without Macular Ischemia.

Purpose: To investigate the changes in the macular microvascular structure after anti-vascular endothelial growth factor (anti-VEGF) treatment in retinal vein occlusion (RVO) patients with and without macular ischemia. Methods: A total of 39 patients were divided into the macular ischemia group (n = 22) and the nonischemia group (n = 17) at baseline. All the patients received an intravitreal injection of ranibizumab with a 3+ pro re nata (PRN) regimen. The foveal avascular zone (FAZ) areas, macular vessel density (VD), and macular ischemic index (ISI) were evaluated at baseline and 3 and 6 months after treatment. Results: After treatment, some patients in the macular ischemia group achieved obvious reperfusion in macular nonperfusion areas. The VD and macular ISI improved in RVO patients, but the changes in VD and macular ISI were different in the two groups. The improvement of best corrected visual acuity (BCVA) was positively correlated with the improvement of macular perfusion status. Macular perfusion remained stable in most patients in RVO and only one patient had macular ischemia aggravation. Conclusion: The macular microvascular structures were stable in most RVO patients after anti-VEGF treatment. At the same time, some patients with macular ischemia presented reperfusion in macular nonperfusion areas, and still a few patients presented aggravated macular ischemia. Macular ISI is a good way to evaluate macular perfusion status in RVO compared to VD.

Effect of anti-VEGF treatment on nonperfusion areas in ischemic retinopathy.

In recent years, retinal ischemia such as that which occurs in diabetic retinopathy (DR) and retinal vein occlusion (RVO) has become a hotspot of ischemic retinopathy research. High levels of vascular endothelial growth factor (VEGF) are recognized as a major cause of macular edema (ME) in DR and RVO. High concentrations of VEGF in the vitreous can lead to serious retinal ischemia and hypoxia and form retinal nonperfusion areas (NPAs). Different levels of retinal ischemia can represent disease severity and progression. Anti-VEGF therapy as the first-line treatment for ME has been found to be effective in improving vision, but there are still disputes about whether anti-VEGF therapy could improve retinal ischemia and achieve reperfusion of previously developed retinal NPAs. Here, we review and summarize studies of the effects of anti-VEGF drugs on retinal ischemia, especially NPAs.

Comparison of Inflammatory and Angiogenic Factors in the Aqueous Humor of Vitrectomized and Non-Vitrectomized Eyes in Diabetic Macular Edema Patients.

Objectives: To compare the aqueous concentrations of inflammatory and angiogenetic factors in vitrectomized vs. non-vitrectomized eyes with diabetic macular edema (DME). Methods: Aqueous samples were obtained from 107 eyes with DME before intravitreal injection of anti-VEGF, 36 eyes with previous pars plana vitrectomy (PPV) combined with pan-retinal endolaser photocoagulation (PRP), and 71 treatment-naïve. Interleukin (IL)-6, IL-8, interferon-induced protein (IP)-10, monocyte chemoattractant protein (MCP)-1, and vascular endothelial growth factor (VEGF) were measured by cytometric bead array (CBA). Optical coherence tomography (OCT) was used for measuring central retinal thickness (CRT). Results: IL-6, IL-8, IP-10, and MCP-1 in aqueous humor of DME vitrectomized eyes were significantly higher than in non-vitrectomized DME eyes, while VEGF was lower than in non-vitrectomized DME eyes. VEGF in aqueous humor significantly correlated with CRT for DME in non-vitrectomized DME eyes. IL-6, IL-8, IP-10, and MCP-1 in aqueous humor were not significantly associated with VEGF for DME in vitrectomized eyes. Conclusions: Inflammation might play an important role in the pathogenesis of DME in vitrectomized eyes. Moreover, inflammation might play a central role in the development of DME via the VEGF-independent pathway. Thus, anti-inflammatory therapy might be a strategy for DME in vitrectomized eyes.

Multifactor analysis of delayed absorption of subretinal fluid after scleral buckling surgery.

BACKGROUND: The purpose of this study is to assess the absorption of subretinal fluid (SRF) after scleral buckling (SB) surgery for the treatment of rhegmatogenous retinal detachment (RRD). We also examined related factors that may affect the delayed absorption of SRF. METHODS: This retrospective study included patients who underwent successful SB surgery for the treatment of macula-off RRD and in which the retina was reattached after the surgery. The patients were categorized according to gender, duration, age, the number, and location of retinal breaks. Subfoveal choroidal thickness (SFCT), height of subretinal fluid (SRFH), and the choriocapillaris flow density (CCFD) within 3 × 3 mm macular fovea were included. Delayed absorption was determined by the SRF that remained unabsorbed for 3 months after the procedure. The endpoint was determined when the SRF could no longer be observed. RESULTS: A total of 62 patients (63 eyes) were enrolled. In 35 eyes (56.45%) SRF was completely absorbed and in 28 (43.55%) eyes delayed absorption of SRF in macular areas was observed at 3 months after surgery. A young age (< 35 years), inferior retinal breaks were associated with good outcomes by applying multivariable analysis on the rate of SRF absorption after SB instead of gender, the number of breaks, and duration (p < 0.05). CCFD was significantly different between the SRF group and the non-SRF group after SB (0.66 ± 0.04% vs 0.63 ± 0.05%, P < 0.05). SRFH showed a moderate positive correlation with SFCT (rs = 0.462, p = 0.000), however, using binary logistic regression analysis it was determined that SFCT was not related to the absorption of the SRF. CONCLUSIONS: The absorption of SRF after SB may be correlated with choriocapillaris flow density. Age and location of breaks are significant factors affecting the absorption of SRF. The duration of disease is an uncertain factor due to several subjective reasons.

The interplay of circulating tumor DNA and chromatin modification, therapeutic resistance, and metastasis.

Peripheral circulating free DNA (cfDNA) is DNA that is detected in plasma or serum fluid with a cell-free status. For cancer patients, cfDNA not only originates from apoptotic cells but also from necrotic tumor cells and disseminated tumor cells that have escaped into the blood during epithelial-mesenchymal transition. Additionally, cfDNA derived from tumors, also known as circulating tumor DNA (ctDNA), carries tumor-associated genetic and epigenetic changes in cancer patients, which makes ctDNA a potential biomarker for the early diagnosis of tumors, monitory and therapeutic evaluations, and prognostic assessments, among others, for various kinds of cancer. Moreover, analyses of cfDNA chromatin modifications can reflect the heterogeneity of tumors and have potential for predicting tumor drug resistance.