Multimodal Imaging of an Idiopathic Vascularized Epiretinal Membrane: A Case Report.

BACKGROUND Idiopathic epiretinal membranes (ERMs) are commonly associated with fibrovascular tissue, primarily observed in ischemic retinopathies. However, idiopathic vascularized ERMs (IVEM) are exceedingly rare, and their pathogenesis and clinical course remain poorly understood. This report aims to contribute to the limited literature on IVEM, shedding light on its characteristics and potential implications for patient management. CASE REPORT We present the case of a 70-year-old man diagnosed with idiopathic ERM in the left eye, revealing a neovascular complex within the membrane. Despite the absence of ocular symptoms and medical history, multimodal imaging using the Nidek Mirante, including spectral domain optical coherence tomography (SD-OCT) and optical coherence tomography angiography (OCT-A), revealed a thick pre-retinal hyper-reflective line with a partial posterior vitreous detachment and an abnormal vascular complex resembling a pruned-vascular-tree pattern. Notably, fluorescein angiography confirmed hyperfluorescence and leakage corresponding to the observed vessels. Despite the rarity of IVEM, the patient remained asymptomatic, and observation was deemed appropriate. CONCLUSIONS IVEM poses a rare challenge in clinical practice, necessitating a comprehensive understanding of its features and potential complications. While the etiopathogenesis remains unclear, hypertension has been proposed as a contributing factor. This case adds valuable insights to the growing literature on IVEM, emphasizing the importance of multimodal imaging in diagnosis and decision-making. Given the limited reports and varied treatment outcomes, managing IVEM requires careful consideration of observation and various therapeutic approaches, highlighting the need for further research to optimize patient care.

Interpretable detection of epiretinal membrane from optical coherence tomography with deep neural networks.

This study aimed to automatically detect epiretinal membranes (ERM) in various OCT-scans of the central and paracentral macula region and classify them by size using deep-neural-networks (DNNs). To this end, 11,061 OCT-images were included and graded according to the presence of an ERM and its size (small 100-1000 µm, large > 1000 µm). The data set was divided into training, validation and test sets (75%, 10%, 15% of the data, respectively). An ensemble of DNNs was trained and saliency maps were generated using Guided-Backprob. OCT-scans were also transformed into a one-dimensional-value using t-SNE analysis. The DNNs' receiver-operating-characteristics on the test set showed a high performance for no-ERM, small-ERM and large-ERM cases (AUC: 0.99, 0.92, 0.99, respectively; 3-way accuracy: 89%), with small-ERMs being the most difficult ones to detect. t-SNE analysis sorted cases by size and, in particular, revealed increased classification uncertainty at the transitions between groups. Saliency maps reliably highlighted ERM, regardless of the presence of other OCT features (i.e. retinal-thickening, intraretinal pseudo-cysts, epiretinal-proliferation) and entities such as ERM-retinoschisis, macular-pseudohole and lamellar-macular-hole. This study showed therefore that DNNs can reliably detect and grade ERMs according to their size not only in the fovea but also in the paracentral region. This is also achieved in cases of hard-to-detect, small-ERMs. In addition, the generated saliency maps can be used to highlight small-ERMs that might otherwise be missed. The proposed model could be used for screening-programs or decision-support-systems in the future.

Optical coherence tomography parameters as prognostic factors for stereopsis after vitrectomy for unilateral epiretinal membrane: a cohort study.

This retrospective cohort study explored the relationship between monocular and interocular optical coherence tomography (OCT) parameters and stereopsis in 56 patients undergoing pars plana vitrectomy (PPV) for unilateral idiopathic epiretinal membrane (IERM). IERM impairs visual functions, with symptoms ranging from asymptomatic to severe impairment. Despite established surgical interventions, including PPV with membrane peeling, the impact on advanced three-dimensional visual functions such as stereopsis remains inadequately investigated. All subjects were assessed for stereopsis, visual acuity, and metamorphopsia, alongside spectral domain OCT parameters. These visual functions significantly improved 3-month postoperatively. Central retinal thickness at the fovea, parafovea, and perifovea (CFT, CRT-3 mm, and CRT-6 mm), ectopic inner foveal layer thickness, and retinal layer thickness notably decreased 1 week to 3 months after surgery. The interocular difference in OCT parameters between bilateral eyes was included as a parameter. Baseline CRT-3 mm difference and inner nuclear layer (INL) thickness were independently correlated with postoperative stereopsis on the Titmus Stereo Test, while baseline CRT-6 mm difference and INL thickness were independently related to stereopsis on the TNO stereotest. This study highlights the substantial enhancement in stereopsis post-IERM surgery, with both interocular and monocular OCT parameters independently influencing postoperative stereopsis. These findings underscore the importance of retinal microstructures in assessing and predicting stereopsis in IERM patients after vitrectomy.

Choroidal macrovascular and capillary alterations in eyes with idiopathic epiretinal membranes.

PURPOSE: To evaluate the choroidal vascular alterations and effect of surgical treatment in the setting of idiopathic epiretinal membranes. METHODS: The structure of the choroid was studied in 33 patients with unilateral idiopathic epiretinal membrane using optical coherence tomography with enhanced depth imaging and optical coherence tomography angiography. Eyes with epiretinal membrane underwent 25-gauge vitrectomy with epiretinal membrane and internal limiting membrane peeling. The choroidal vascularity index, Haller layer/choroidal thickness ratio, and choriocapillaris flow density were used to evaluate changes in choroidal structure after surgery and compare with the healthy fellow eyes. RESULTS: The choroidal vascularity index and Haller layer/choroidal thickness ratio of the eyes with epiretinal membrane were higher than those of the fellow eyes at baseline (p=0.009 and p=0.04, respectively) and decreased postoperatively compared with preoperative values (p=0.009 and p=0.001, respectively). The choriocapillaris flow of eyes with epiretinal membrane was lower than that of the fellow eyes at baseline (p=0.001) and increased after surgery compared with the preoperative value (p=0.04). The choroidal vascularity index, Haller layer/choroidal thickness ratio, and choriocapillaris flow values of the healthy fellow eyes were comparable at baseline and final visit. In eyes with epiretinal membrane, the final choroidal vascularity index correlated with the final choriocapillaris flow (r=-0.749, p=0.008) in the multivariate analysis. CONCLUSION: Idiopathic epiretinal membrane appears to affect the choroidal structure with increased choroidal vascularity index and Haller layer/ choroidal thickness ratio and decreased choriocapillaris flow. These macrovascular (choroidal vascularity index and Haller layer/choroidal thickness) and microvascular (choriocapillaris flow) alterations appear to be relieved by surgical treatment of the epiretinal membranes.

A classification of idiopathic epiretinal membrane based on foveal avascular zone area using optical coherence tomography angiography.

OBJECTIVE: To evaluate the characteristics and prognoses of idiopathic macular epiretinal membrane (iERM) using a classification based on the foveal avascular zone (FAZ) area. METHOD: IERMs were classified into four stages based on the FAZ area. Baseline FAZ-related parameters, pre-and postoperative central macular thickness (CMT), and best corrected visual acuity (BCVA) were observed and compared between different stages. The correlations of structural parameters with pre-and postoperative logMAR BCVA were analyzed. RESULTS: 162 iERM eyes were enrolled, including 105 eyes followed up for 12 months after surgery. The preoperative BCVA was better at the early stage. Postoperative BCVA at Stages 2 and 3 were better compared to Stage 4. The early stage was associated with thinner CMT pre-and postoperatively. However, there was no significant difference in CMT between postoperative Stages 1 and 2 or Stages 3 and 4. Preoperative logMAR BCVA was negatively correlated with FAZ area, perimeter, and FD-300 and was positively correlated with CMT and acircularity index (AI). CMT correlated positively with BCVA for each stage, except Stage 4; FAZ area, perimeter, and FD-300 had a negative correlation at Stage 1. Baseline BCVA and CMT positively correlated with BCVA at the last follow-up, while FAZ area and FD-300 were negatively correlated. Baseline BCVA had a positive correlation for each stage, except Stage 1; FD-300 had a negative correlation at Stages 2 and 3; CMT had a positive correlation at Stage 3. CONCLUSION: A classification based on the FAZ area was established innovatively. This classification can reflect the progression of iERM and is helpful to the postoperative prognosis.

(1) Classification based on FAZ area facilitate automation and consistency compared to the previous OCT-based qualitative grading.(2) With baseline FAZ stage advanced, thickened CMT and worsened BCVA was observed at baseline and 1-year post-operation. (3) Baseline FAZ area and FD-300 negatively correlated with logMAR BCVA at the last follow-up, reflecting the nonnegligible prognostic impact of macular vascular changes.

Development of a generative deep learning model to improve epiretinal membrane detection in fundus photography.

BACKGROUND: The epiretinal membrane (ERM) is a common retinal disorder characterized by abnormal fibrocellular tissue at the vitreomacular interface. Most patients with ERM are asymptomatic at early stages. Therefore, screening for ERM will become increasingly important. Despite the high prevalence of ERM, few deep learning studies have investigated ERM detection in the color fundus photography (CFP) domain. In this study, we built a generative model to enhance ERM detection performance in the CFP. METHODS: This deep learning study retrospectively collected 302 ERM and 1,250 healthy CFP data points from a healthcare center. The generative model using StyleGAN2 was trained using single-center data. EfficientNetB0 with StyleGAN2-based augmentation was validated using independent internal single-center data and external datasets. We randomly assigned healthcare center data to the development (80%) and internal validation (20%) datasets. Data from two publicly accessible sources were used as external validation datasets. RESULTS: StyleGAN2 facilitated realistic CFP synthesis with the characteristic cellophane reflex features of the ERM. The proposed method with StyleGAN2-based augmentation outperformed the typical transfer learning without a generative adversarial network. The proposed model achieved an area under the receiver operating characteristic (AUC) curve of 0.926 for internal validation. AUCs of 0.951 and 0.914 were obtained for the two external validation datasets. Compared with the deep learning model without augmentation, StyleGAN2-based augmentation improved the detection performance and contributed to the focus on the location of the ERM. CONCLUSIONS: We proposed an ERM detection model by synthesizing realistic CFP images with the pathological features of ERM through generative deep learning. We believe that our deep learning framework will help achieve a more accurate detection of ERM in a limited data setting.

Accuracy of biomicroscopy, ultrasonography and spectral-domain OCT in detection of complete posterior vitreous detachment.

BACKGROUND: To evaluate the accuracy of preoperative biomicroscopy (BM), ultrasonography (US), and spectral domain optical coherence tomography (SD-OCT) to determine complete posterior vitreous detachment (PVD) confirmed by intraoperative findings of triamcinolone acetonide-assisted pars plana vitrectomy (PPV). METHODS: This prospective study included all consecutive patients admitted for surgical treatment of the epiretinal membrane (ERM) and macular hole (MH). The presence of complete PVD was determined one day before PPV using BM, US, SD-OCT. The preoperative findings were compared to the PVD status determined during PPV. RESULTS: A total of 123 eyes from 123 patients were included in the study. Indications for PPV included ERM in 57 (46.3%), full thickness macular hole in 57 (46.3%) and lamellar macular hole in 9 (7.3%) patients. Complete PVD during PPV was observed in 18 (31.6%; 95%CI:18.7-49.9) patients with ERM and 13 (19.7%; 95%CI:10.4-33.7) patients with MH. The sensitivity of preoperative BM, US, SD-OCT was 48.4% (95%CI:30.2-66.9), 61.3% (95%CI:42.2-78.2) and 54.8% (95%CI:36.0-72.7) respectively. The specificity of preoperative BM, US, SD-OCT was 81.5% (95%CI:72.1-88.9), 90.2% (95%CI:82.2-95.4) and 85.9% (95%CI:77.0-92.3) respectively. With a prevalence of 25.2% of PVD in our sample the positive predictive value of preoperative BM, US, SD-OCT was 46.9% (95%CI:29.1-65.3), 67.9% (95%CI:47.6-84.1) and 56.7% (95%CI:37.4-74.5) respectively. CONCLUSION: Preoperative BM, US, and SD-OCT showed relatively low sensitivity but also good specificity in assessing complete PVD. A combination of all three diagnostic methods can provide a good assessment of the vitreoretinal interface state.

Deep learning-based prediction of the retinal structural alterations after epiretinal membrane surgery.

To generate and evaluate synthesized postoperative OCT images of epiretinal membrane (ERM) based on preoperative OCT images using deep learning methodology. This study included a total 500 pairs of preoperative and postoperative optical coherence tomography (OCT) images for training a neural network. 60 preoperative OCT images were used to test the neural networks performance, and the corresponding postoperative OCT images were used to evaluate the synthesized images in terms of structural similarity index measure (SSIM). The SSIM was used to quantify how similar the synthesized postoperative OCT image was to the actual postoperative OCT image. The Pix2Pix GAN model was used to generate synthesized postoperative OCT images. Total 60 synthesized OCT images were generated with training values at 800 epochs. The mean SSIM of synthesized postoperative OCT to the actual postoperative OCT was 0.913. Pix2Pix GAN model has a possibility to generate predictive postoperative OCT images following ERM removal surgery.

Morphologic analysis of the foveal avascular zone for prediction of postoperative visual acuity in advanced idiopathic epiretinal membrane.

To investigate the preoperative morphology of the foveal avascular zone (FAZ) for prediction of the postoperative visual acuity in advanced idiopathic epiretinal membrane (ERM). 28 patients (28 eyes) with unilateral idiopathic ERM who underwent pars plana vitrectomy with internal limiting membrane peeling were included. Superficial FAZ was measured preoperatively in both eyes using optical coherence tomography angiography. Area, perimeter, and circularity of FAZ were achieved, and the differences between the ERM eyes and the contralateral eyes were evaluated to analyze the degree of FAZ distortion in diseased eyes. The best-corrected visual acuity (BCVA) and central foveal thickness (CFT) were measured at baseline and more than 6 months after surgery. The correlations of the preoperative FAZ with BCVA and CFT were assessed. The FAZ in the eyes with ERM was significantly reduced, and the BCVA was significantly correlated with the FAZ area (FAZa) (P = 0.001) and the FAZ perimeter (FAZp) (P < 0.001) before surgery. LogMAR BCVA and CFT were significantly improved from 0.550 ± 0.221 to 0.354 ± 0.229 (P = 0.008), and from 524.393 ± 93.575 µm to 400.071 ± 75.979 µm (P < 0.001) after surgery. The preoperative FAZa and FAZp were significantly associated with letter score gain (P < 0.001, P < 0.001) and the postoperative final BCVA (P = 0.026, P = 0.006). The preoperative FAZp had correlation with ratio of postoperative to preoperative CFT (P = 0.016). The preoperative FAZp is a predictor of visual acuity and morphological prognosis after surgery in advanced idiopathic ERM.

The classification of stages of epiretinal membrane using convolutional neural network on optical coherence tomography image.

BACKGROUND AND OBJECTIVE: The gold standard for diagnosing epiretinal membranes is to observe the surface of the internal limiting membrane on optical coherence tomography images. The stages of the epiretinal membrane are used to decide the condition of the health of the membrane. The stages are not detected because some of them are similar. To accurately classify the stages, a deep-learning technology can be used to improve the classification accuracy. METHODS: A combinatorial fusion with multiple convolutional neural networks (CNN) algorithms are proposed to enhance the accuracy of a single image classification model. The proposed method was trained using a dataset of 1947 optical coherence tomography images diagnosed with the epiretinal membrane at the Taichung Veterans General Hospital in Taiwan. The images consisted of 4 stages; stages 1, 2, 3, and 4. RESULTS: The overall accuracy of the classification was 84%. The combination of five and six CNN models achieves the highest testing accuracy (85%) among other combinations, respectively. Any combination with a different number of CNN models outperforms any single CNN algorithm working alone. Meanwhile, the accuracy of the proposed method is better than ophthalmologists with years of clinical experience. CONCLUSIONS: We have developed an efficient epiretinal membrane classification method by using combinatorial fusion with CNN models on optical coherence tomography images. The proposed method can be used for screening purposes to facilitate ophthalmologists making the correct diagnoses in general medical practice.

Retinal vessel geometry in patients with idiopathic epiretinal membrane.

We investigated the associations between retinal vascular geometric measurements and idiopathic epiretinal membrane (ERM). Whether changes in retinal vascular geometry are independent of systemic cardiovascular risk factors was also evaluated. This retrospective, cross sectional study included 98 patients with idiopathic ERM, and 99 healthy age-matched controls. Quantitative retinal vascular parameters were measured from digital retinal fundus photographs using a semi-automated computer-assisted program. Multivariate logistic regression analyses were performed to evaluate associations between retinal vascular geometric parameters and the presence of idiopathic ERM after adjusting for systemic cardiovascular risk factors. There was no significant difference in the baseline characteristics of the two groups, except that the ERM group had a higher proportion of females than the control group. In multivariate regression analyses, female sex (odds ratio [OR] 0.402; 95% CI 0.196-0.802; P = 0.011), wider retinal venular caliber (OR 16.852; 95% CI 5.384-58.997; P < 0.001) and decreased total fractal dimension (OR 0.156; 95% CI 0.052-0.440; P = 0.001) were associated with idiopathic ERM. Idiopathic ERM was associated with alterations in global retinal microvascular geometric parameters, wider retinal venules, and less complex vascular branching patterns, independent of cardiovascular risk factors.

Image processing and quantification analysis for optical coherence tomography angiography in epiretinal membrane.

BACKGROUND: To explore image processing methods for optical coherence tomography angiography (OCTA) of the epiretinal membrane (ERM), and to evaluate the impact of ERM on vision by analyzing the retinal vasculature. METHODS: Thirty eyes of 30 patients with idiopathic ERM who underwent OCTA were retrospectively evaluated. Image processing of OCTA, including the Mexican hat filter (MHF) and exclusion of the foveal avascular zone (FAZ), was attempted using Fiji. OCTA parameters, including vessel density (VD), fractal dimension (FD), and vessel tortuosity (VT), were measured for large vessels only, capillaries only, and the whole vasculature. Pearson correlation analysis was used to evaluate the correlations between best-corrected visual acuity (BCVA) and OCTA parameters. RESULTS: The correlations between BCVA and retinal vasculature were much increased when the capillaries only instead of the whole vasculature was used for analysis. Both higher VD and FD of capillaries were correlated with better BCVA, and MHF largely increased their correlations (P < 0.0001 for both). In contrast, both higher VD and FD of the large vessels were associated with poorer BCVA (P = 0.042 and 0.049, respectively). A higher VT of capillaries was correlated with better BCVA, and both MHF and exclusion of the FAZ were necessary to reveal their correlations (P = 0.028) CONCLUSIONS: Separation of large vessels and capillaries was necessary to reveal the correlations between retinal vasculature and BCVA in ERM. MHF was necessary to elucidate all microvascular parameters of capillaries, and exclusion of the FAZ was mandatory for evaluation of VT.

Effects of image-sharpening algorithm on surgical field visibility during 3D heads-up surgery for vitreoretinal diseases.

We conducted clinical and experimental studies to investigate the effects of image-sharpening algorithms and color adjustments, which enabled real-time processing of live surgical images with a delay of 0.004 s. The images were processed with image-sharpening intensities of 0%, 12.5%, 25%, and 50% during cataract surgery, vitrectomy, peeling of epiretinal membrane, and peeling of internal limiting membrane (ILM) with the Ngenuity 3D visualization system. In addition, the images obtained with a yellow filter during the ILM peeling were processed with color adjustments. Five vitreoretinal surgeons scored the clarity of the images on a 10-point scale. The images of a 1951 United States Air Force grating target placed in no fluid (control), saline, and 0.1% and 1% milk solution were evaluated. The results showed that the mean visibility score increased significantly from 5.0 ± 0.6 at 0% to 6.4 ± 0.6 at 12.5%, 7.3 ± 0.7 at 25%, and 7.5 ± 0.9 at 50% (P < 0.001). The visibility scores during ILM peeling improved significantly with color adjustments (P = 0.005). In the experimental study, the contrast of the grating targets blurred by the 0.1% and 1% milk solution increased significantly by the image-sharpening procedure. We conclude that the image-sharpening algorithms and color adjustments improved the intraoperative visibility of 3D heads-up surgery.

Geometric Morphometrics Can Predict Postoperative Visual Acuity Changes in Patients With Epiretinal Membrane: A Retrospective Study.

Purpose: To investigate the efficacy of the geometric morphometrics method for the evaluation of retinal deformation in patients with epiretinal membrane (ERM) and determine whether the degree of deformation can serve as a predictive factor for postoperative visual outcome. Methods: We retrospectively evaluated data from 29 eyes of 29 patients with primary ERM. Preoperative optical coherence tomography images were compared with images of their normal fellow eyes using the geometric morphometrics thin-plate spline technique. Conventional parameters such as retinal layer thickness and previously reported indices were also measured. The correlation between the preoperative parameters and visual acuity was evaluated. Statistical comparisons were performed using a paired t-test, and associations between the optical coherence tomography image parameters and visual acuity were determined using Spearman's rank correlation coefficient. Results: Bending energy, which was calculated using geometric morphometrics, was significantly associated with visual acuity as well as conventional optical coherence tomography parameters and previously reported indices. Multiple regression analysis showed that bending energy was an independent predictive factor for postoperative visual acuity changes. Conclusions: The geometric morphometrics method is an effective approach for evaluating the severity of ERM and predicting the efficacy of surgery. Translational Relevance: Geometric morphometrics can effectively evaluate retinal deformation in eyes with epiretinal membrane.

Evaluation of the optical coherence tomography findings of patients with idiopathic epiretinal membrane.

BACKGROUND: This study aimed to determine the relationship between basal visual acuity (VA) and basal optical coherence tomography (OCT) parameters of patients with idiopathic epiretinal membrane (iERM). METHODS: The study included 200 eyes of 200 patients with iERM. In the evaluation of basal VAs, logMAR 0.3 (0.5 on Snellen chart) was accepted as the limit, and the patients were divided into two groups as Group 1 with VA values below logMAR 0.3 and Group 2 with those above logMAR 0.3. Using OCT, the ellipsoid zone (EZ), interdigitation zone (IZ), external limiting membrane (ELM) integrity, peaked appearance of EZ in the fovea, disorganization of the inner retinal layers (DRIL), central foveal thickness (CFT), maximum retinal thickness (MRT), total photoreceptor length (TPL), photoreceptor outer segment length (PROSL), photoreceptor deformity index (PDI), and inner retinal layer irregularity index (IRLII) were evaluated and compared between the groups. RESULTS: There were an equal number of patients in Group 1 and Group 2. The preservation of ELM, EZ, and IZ integrity and the peaked appearance of EZ in the fovea were observed at higher rates in Group 1, whereas the rate of DRIL was higher in Group 2 (p < 0.05 for all). In the multiple regression analysis performed to evaluate categorical OCT parameters, only the presence of DRIL was found to be significantly associated with VA (p < 0.003). CFT*, MRT*, and IRLII** were significantly lower in Group 1 (*p < 0.001, **p = 0.001). TPU, PROSL, and FDI were similar between the two groups (p > 0.05 for all). CONCLUSION: According to the multiple regression analysis, only the presence of DRIL was correlated with lower VA values. Among the OCT parameters, an increase in CFT and MRT, as well as the IRLII value moving away from 1 were related to lower VA values.

Prediction of Visual Impairment in Epiretinal Membrane and Feature Analysis: A Deep Learning Approach Using Optical Coherence Tomography.

PURPOSE: The aim was to develop a deep learning model for predicting the extent of visual impairment in epiretinal membrane (ERM) using optical coherence tomography (OCT) images, and to analyze the associated features. METHODS: Six hundred macular OCT images from eyes with ERM and no visually significant media opacity or other retinal diseases were obtained. Those with best-corrected visual acuity ≤20/50 were classified as "profound visual impairment," while those with best-corrected visual acuity >20/50 were classified as "less visual impairment." Ninety percent of images were used as the training data set and 10% were used for testing. Two convolutional neural network models (ResNet-50 and ResNet-18) were adopted for training. The t-distributed stochastic neighbor-embedding approach was used to compare their performances. The Grad-CAM technique was used in the heat map generative phase for feature analysis. RESULTS: During the model development, the training accuracy was 100% in both convolutional neural network models, while the testing accuracy was 70% and 80% for ResNet-18 and ResNet-50, respectively. The t-distributed stochastic neighbor-embedding approach found that the deeper structure (ResNet-50) had better discrimination on OCT characteristics for visual impairment than the shallower structure (ResNet-18). The heat maps indicated that the key features for visual impairment were located mostly in the inner retinal layers of the fovea and parafoveal regions. CONCLUSIONS: Deep learning algorithms could assess the extent of visual impairment from OCT images in patients with ERM. Changes in inner retinal layers were found to have a greater impact on visual acuity than the outer retinal changes.

A systematic evaluation of human expert agreement on optical coherence tomography biomarkers using multiple devices.

OBJECTIVES: To assess the agreement in evaluating optical coherence tomography (OCT) variables in the leading macular diseases such as neovascular age-related macular degeneration (nAMD), diabetic macular oedema (DMO) and retinal vein occlusion (RVO) among OCT-certified graders. METHODS: SD-OCT volume scans of 356 eyes were graded by seven graders. The grading included presence of intra- and subretinal fluid (IRF, SRF), pigment epithelial detachment (PED), epiretinal membrane (ERM), conditions of the vitreomacular interface (VMI), central retinal thickness (CRT) at the foveal centre-point (CP) and central millimetre (CMM), as well as height and location of IRF/SRF/PED. Kappa statistics (κ) and intraclass correlation coefficient (ICC) were used to report categorical grading and measurement agreement. RESULTS: The overall agreement on the presence of IRF/SRF/PED was κ = 0.82/0.85/0.81; κ of VMI condition was 0.77, that of ERM presence 0.37. ICC for CRT measurements at CP and CMM was excellent with an ICC of 1.00. Height measurements of IRF/SRF/PED showed robust consistency with ICC = 0.85-0.93. There was substantial to almost perfect agreement in locating IRF/SRF/PED with κ = 0.67-0.86. Between diseases, κ of IRF/SRF presence was 0.69/0.80 for nAMD, 0.64/0.83 for DMO and 0.86/0.89 for RVO. CONCLUSION: Even in the optimized setting, featuring certified graders, standardized image acquisition and the use of a professional reading platform, there is a disease dependent variability in biomarker evaluation that is most pronounced for IRF in nAMD as well as DMO. Our findings highlight the variability in the performance of human expert OCT grading and the need for AI-based automated feature analyses.

OPTICAL COHERENCE TOMOGRAPHY FEATURES IN IDIOPATHIC NONINFECTIOUS INTERMEDIATE, POSTERIOR, OR PANUVEITIS.

PURPOSE: To characterize optical coherence tomography features in patients with idiopathic intermediate, posterior, or panuveitis. METHODS: This is a retrospective case series of all consecutive cases of idiopathic intermediate, posterior, or panuveitis at four tertiary care centres between 2010 and 2021. RESULTS: A total of 94 eyes (55 patients) were followed for an average duration of 29.8 (SD 21) months. The median central macular thickness was 284 µ m at baseline and 267 µ m at last follow-up. At baseline and last follow-up, respectively, 24% and 20% of uveitic eyes had intraretinal fluid, 12% and 1% subretinal fluid, and 43% and 54% epiretinal membrane. In addition, ellipsoid zone abnormalities on en-face were noted in 34% and 19% of cases at baseline and last follow-up, respectively. The baseline median visual acuity was significantly lower among cases with ellipsoid zone en-face slab abnormalities compared with those without (0.2 logarithm of minimum angle of resolution [interquartile range: 0-0.6] vs. 0.1 logarithm of minimum angle of resolution [interquartile range: 0-0.3], P = 0.0051). CONCLUSION: With initiation of treatment, the central macular thickness, intraretinal fluid, subretinal fluid, and ellipsoid zone en-face abnormalities improved over time, whereas the number of cases with epiretinal membrane increased among eyes with idiopathic intermediate, posterior, and panuveitis. Presence of ellipsoid zone en-face abnormalities at presentation may be associated with worse visual acuity.

End-to-end multi-task learning approaches for the joint epiretinal membrane segmentation and screening in OCT images.

BACKGROUND AND OBJECTIVES: The Epiretinal Membrane (ERM) is an ocular disease that can cause visual distortions and irreversible vision loss. Patient sight preservation relies on an early diagnosis and on determining the location of the ERM in order to be treated and potentially removed. In this context, the visual inspection of the images in order to screen for ERM signs is a costly and subjective process. METHODS: In this work, we propose and study three end-to-end fully-automatic approaches for the simultaneous segmentation and screening of ERM signs in Optical Coherence Tomography images. These convolutional approaches exploit a multi-task learning context to leverage inter-task complementarity in order to guide the training process. The proposed architectures are combined with three different state of the art encoder architectures of reference in order to provide an exhaustive study of the suitability of each of the approaches for these tasks. Furthermore, these architectures work in an end-to-end manner, entailing a significant simplification of the development process since they are able to be trained directly from annotated images without the need for a series of purpose-specific steps. RESULTS: In terms of segmentation, the proposed models obtained a precision of 0.760 ± 0.050, a sensitivity of 0.768 ± 0.210 and a specificity of 0.945 ± 0.011. For the screening task, these models achieved a precision of 0.963 ± 0.068, a sensitivity of 0.816 ± 0.162 and a specificity of 0.983 ± 0.068. The obtained results show that these multi-task approaches are able to perform competitively with or even outperform single-task methods tailored for either the segmentation or the screening of the ERM. CONCLUSIONS: These results highlight the advantages of using complementary knowledge related to the segmentation and screening tasks in the diagnosis of this relevant pathology, constituting the first proposal to address the diagnosis of the ERM from a multi-task perspective.