Search Results (3)

Background: The vitreomacular interface (VMI) encompasses a group of retinal disorders that significantly impact vision, requiring accurate classification for effective management. This study aims to compare the effectiveness of an expert-designed custom deep learning (DL) model and a code free Auto Machine Learning (ML) model in classifying optical coherence tomography (OCT) images of VMI disorders. Materials and Methods: A balanced dataset of OCT images across five classes—normal, epiretinal membrane (ERM), idiopathic full-thickness macular hole (FTMH), lamellar macular hole (LMH), and vitreomacular traction (VMT)—was used. The expert-designed model combined ResNet-50 and EfficientNet-B0 architectures with Monte Carlo cross-validation. The AutoML model was created on Google Vertex AI, which handled data processing, model selection, and hyperparameter tuning automatically. Performance was evaluated using average precision, precision, and recall metrics. Results: The expert-designed model achieved an overall balanced accuracy of 95.97% and a Matthews Correlation Coefficient (MCC) of 94.65%. Both models attained 100% precision and recall for normal cases. For FTMH, the expert model reached perfect precision and recall, while the AutoML model scored 97.8% average precision, and 97.4% recall. In VMT detection, the AutoML model showed 99.5% average precision with a slightly lower recall of 94.7% compared to the expert model’s 95%. For ERM, the expert model achieved 95% recall, while the AutoML model had higher precision at 93.9% but a lower recall of 79.5%. In LMH classification, the expert model exhibited 95% precision, compared to 72.3% for the AutoML model, with similar recall for both (88% and 87.2%, respectively). Conclusions: While the AutoML model demonstrated strong performance, the expert-designed model achieved superior accuracy across certain classes. AutoML platforms, although accessible to healthcare professionals, may require further advancements to match the performance of expert-designed models in clinical applications. Full article

There is a wide spectrum of macular conditions that are characterized by an irregular foveal contour caused by a break in the inner fovea. These include full-thickness macular hole (FTMH), foveal pseudocyst, lamellar macular hole (LMH) and macular pseudohole (MPH). Clinical examination of vitreomacular interface disorders is notoriously poor in differentiating these conditions. These conditions were initially described with slit-lamp biomicroscopy, and the main goal was to distinguish an FTMH from the others. The introduction of optical coherence tomography (OCT) has revolutionized our understanding of the foveal microstructural anatomy and has facilitated differentiating these conditions from an FTMH. However, the definitions of the other conditions, particularly LMH, has evolved over the past two decades. Initially the term LMH encompassed a wide spectrum of clinical conditions. As OCT became more widely used and observations became more refined, two different phenotypes of LMH became apparent, raising the question of different pathogenic mechanisms for each phenotype. Tractional and degenerative pathological mechanisms were proposed. Epiretinal membranes (ERMs) associated with each phenotype were identified. Typical ERMs were associated with a tractional mechanism, whereas an epiretinal proliferation was associated with a degenerative mechanism. Epiretinal proliferation represents Müller cell proliferation as a reactive process to retinal injury. These two types of ERM were differentiated by their characteristics on SD-OCT. The latest consensus definitions take into account this phenotypic differentiation and classifies these entities into LMH, MPH and ERM foveoschisis. The initial event in both ERM foveoschisis and LMH is a tractional event that disrupts the Müller cell cone in the foveola or the foveal walls. Depending on the extent of Müller cell disruption, either a LMH or an ERM foveoschisis may develop. Although surgical intervention for LMH remains controversial and no clear guidelines exist for pars plana vitrectomy (PPV), eyes with symptomatic, progressive ERM foveoschisis and LMH may benefit from surgical intervention. Full article

Vitreomacular interface plays an important role in the pathogenesis and progression of proliferative diabetic retinopathy (PDR). This study investigated the prevalence and risk factors of vitreomacular interface disorders (VMID) in PDR. The macular optical coherence tomography (OCT) scans of 493 eyes from 378 PDR patients were retrospectively reviewed to detect VMID, including vitreomacular adhesion (VMA), vitreomacular traction (VMT), epiretinal membrane (ERM), lamellar hole–associated epiretinal proliferation (LHEP), and macular hole (MH). The associations between VMID and baseline factors, intraretinal structure, and visual acuity were analyzed. The prevalence was 78.9% for ERM, 13.4% for VMT, 4.8% for MH, 2.2% for LHEP, and 2.0% for VMA, respectively. On multivariable analyses (odds ratio, 95% confidence interval), fibrovascular proliferation (FVP) was positively associated with MH (8.029, 1.873–34.420), VMT (3.774, 1.827–7.798), and ERM (2.305, 1.460–3.640). High-risk PDR was another risk factor of ERM (1.846, 1.101–3.090). Female gender was positively associated with MH (3.836, 1.132–13.006), while vitreous hemorrhage was negatively associated with MH (0.344, 0.133–0.890). Eyes with all VMID subtypes showed more frequent macular cysts and tractional retinal detachment with poorer visual acuity (p ≤ 0.001). Therefore, the prevalence of VMID was considerably high, indicating that this distinct entity should be considered in interventions for PDR. Full article