Deploying efficient net batch normalizations (BNs) for grading diabetic retinopathy severity levels from fundus images.

Diabetic retinopathy (DR) is one of the main causes of blindness in people around the world. Early diagnosis and treatment of DR can be accomplished by organizing large regular screening programs. Still, it is difficult to spot diabetic retinopathy timely because the situation might not indicate signs in the primary stages of the disease. Due to a drastic increase in diabetic patients, there is an urgent need for efficient diabetic retinopathy detecting systems. Auto-encoders, sparse coding, and limited Boltzmann machines were used as a few past deep learning (DL) techniques and features for the classification of DR. Convolutional Neural Networks (CNN) have been identified as a promising solution for detecting and classifying DR. We employ the deep learning capabilities of efficient net batch normalization (BNs) pre-trained models to automatically acquire discriminative features from fundus images. However, we successfully achieved F1 scores above 80% on all efficient net BNs in the EYE-PACS dataset (calculated F1 score for DeepDRiD another dataset) and the results are better than previous studies. In this paper, we improved the accuracy and F1 score of the efficient net BNs pre-trained models on the EYE-PACS dataset by applying a Gaussian Smooth filter and data augmentation transforms. Using our proposed technique, we have achieved F1 scores of 84% and 87% for EYE-PACS and DeepDRiD.

Autonomous assessment of spontaneous retinal venous pulsations in fundus videos using a deep learning framework.

The presence or absence of spontaneous retinal venous pulsations (SVP) provides clinically significant insight into the hemodynamic status of the optic nerve head. Reduced SVP amplitudes have been linked to increased intracranial pressure and glaucoma progression. Currently, monitoring for the presence or absence of SVPs is performed subjectively and is highly dependent on trained clinicians. In this study, we developed a novel end-to-end deep model, called U3D-Net, to objectively classify SVPs as present or absent based on retinal fundus videos. The U3D-Net architecture consists of two distinct modules: an optic disc localizer and a classifier. First, a fast attention recurrent residual U-Net model is applied as the optic disc localizer. Then, the localized optic discs are passed on to a deep convolutional network for SVP classification. We trained and tested various time-series classifiers including 3D Inception, 3D Dense-ResNet, 3D ResNet, Long-term Recurrent Convolutional Network, and ConvLSTM. The optic disc localizer achieved a dice score of 95% for locating the optic disc in 30 milliseconds. Amongst the different tested models, the 3D Inception model achieved an accuracy, sensitivity, and F1-Score of 84 ± 5%, 90 ± 8%, and 81 ± 6% respectively, outperforming the other tested models in classifying SVPs. To the best of our knowledge, this research is the first study that utilizes a deep neural network for an autonomous and objective classification of SVPs using retinal fundus videos.

[A case of von Hippel-Lindau syndrome with optic disc pit and macular hole].

A 69-year-old female patient presented to the ophthalmology department with complaints of blurred vision in the left eye for more than 10 days. Her medical history revealed a history of right renal tumor and left pheochromocytoma, which were treated with surgical resection at an external institution. Ophthalmic examination revealed a temporal superior cup-shaped optic disc pit in the left eye, along with a macular hole approximately 1/5 the size of the optic disc diameter in the macular region. Additionally, peripheral retinal examination at the 6 o'clock and 11 o'clock positions showed vascular tumors, each approximately 1.5 times the size of the optic disc diameter. Based on the patient's medical history, fundus findings, and auxiliary examination results, a diagnosis of macular hole in the left eye, optic disc pit in the left eye, and Von Hippel-Lindau (VHL) syndrome was established. Subsequently, the patient underwent left vitrectomy and macular hole repair surgery, leading to an improvement in visual acuity.

[A case of congenital grouped albinotic spots].

This case report presents a 5-year-old male child with a complaint of poor vision in the left eye for the past 2 years, who sought ophthalmic evaluation. There was no apparent systemic pigment loss, but multiple small, flat, and well-defined white lesions were observed in the retinal pigment epithelium of the left eye. Autofluorescence imaging of the fundus revealed widespread patchy hyperautofluorescence corresponding to the lesions. Fluorescein angiography demonstrated early and stable hyperfluorescence without leakage in these spots. Optical coherence tomography examination revealed thickening and material accumulation in the ellipsoid zone that corresponded to the lesions. Based on clinical findings, the diagnosis of congenital grouped albinotic spots was established.

Optical modeling of the entire visual field of the eye.

Vision is rarely evaluated scientifically at very large visual angles, despite being used continuously in everyday life. Furthermore, raytrace calculations indicate that peripheral optical properties are different for a pseudophakic eye, and even though this is rarely noted by patients, it is probably the cause of bothersome "negative dysphotopsia." Simplified paraxial parameters that characterize the basic properties of phakic and pseudophakic eyes are collected together here as a baseline, and then raytracing is used to show that input angles of about 60°, which correspond to obstruction by the nose, eyebrow, and cheek, illuminate a retinal hemisphere. At larger angles in the temporal direction, the image with an intraocular lens (IOL) reaches a limit due to vignetting at about a 90° input angle to the optical axis, in comparison to 105° with the Gullstrand-Emsley eye model, and 109° for the most realistic gradient index crystalline lens model. Scaling the far peripheral vision region more accurately may lead to benefits relating to intraocular lenses, diseases of the peripheral retina, widefield fundus images, and myopia prevention.

Synthetic OCT-A blood vessel maps using fundus images and generative adversarial networks.

Vessel segmentation in fundus images permits understanding retinal diseases and computing image-based biomarkers. However, manual vessel segmentation is a time-consuming process. Optical coherence tomography angiography (OCT-A) allows direct, non-invasive estimation of retinal vessels. Unfortunately, compared to fundus images, OCT-A cameras are more expensive, less portable, and have a reduced field of view. We present an automated strategy relying on generative adversarial networks to create vascular maps from fundus images without training using manual vessel segmentation maps. Further post-processing used for standard en face OCT-A allows obtaining a vessel segmentation map. We compare our approach to state-of-the-art vessel segmentation algorithms trained on manual vessel segmentation maps and vessel segmentations derived from OCT-A. We evaluate them from an automatic vascular segmentation perspective and as vessel density estimators, i.e., the most common imaging biomarker for OCT-A used in studies. Using OCT-A as a training target over manual vessel delineations yields improved vascular maps for the optic disc area and compares to the best-performing vessel segmentation algorithm in the macular region. This technique could reduce the cost and effort incurred when training vessel segmentation algorithms. To incentivize research in this field, we will make the dataset publicly available to the scientific community.

Tear inflammatory cytokines as potential biomarkers for myopic macular degeneration.

Previous studies have reported that inflammatory cytokine levels increase in the intraocular fluids (aqueous humor and vitreous) of highly myopic eyes, However, there has been currently no study revealing the levels of inflammatory cytokines in tear. Therefore, this study aimed to determine tear cytokine levels of highly myopic eyes, and their relationships with myopic macular degeneration (MMD). This case-control study screened inflammatory cytokines of tear samples from 132 highly myopic and 105 emmetropic eyes using a multiplex cytokine antibody array, and cytokines showing significant intergroup differences were further validated using ProQuantum immunoassays in tear samples from another 60 highly myopic and 60 emmetropic eyes. Ultra-widefield fundus photographs of eyes were classified according to the meta-analyses of the Pathologic Myopia Classification. Associations between tear cytokine levels and MMD category were investigated. As a result, tear levels of interleukin (IL)-6, IL-13 and monocyte chemoattractant protein (MCP)-1 were screened significantly higher in highly myopic eyes than in emmetropic controls (IL-6: 11.70 ± 16.81 versus 8.22 ± 10.76 pg/mL; MCP-1: 63.60 ± 54.40 versus 33.87 ± 43.82 pg/mL; both P < 0.05). Validation assays further demonstrated the elevated concentrations of IL-6 and MCP-1 (IL-6: 13.97 ± 8.41 versus 8.06 ± 7.94 pg/mL, P < 0.001; MCP-1: 32.69 ± 8.41 versus 18.07 ± 8.41 pg/mL, P = 0.003). Tear levels of IL-6 and MCP-1 differed significantly among MMD categories (both P < 0.05). The area under receiver operating characteristic curve were 0.783 and 0.682 respectively (both P < 0.05), when using tear IL-6 and MCP-1 levels to predict the presence of MMD (category ≥2). The ordered logistic regression model also indicated that longer axial length, and higher IL-6 and MCP-1 tear levels were independent predictors of higher MMD category. In our study, highly myopic eyes presented significantly higher levels of tear IL-6 and MCP-1, which may also serve as potential biomarkers for MMD.

Genetic Risk Score Analysis Supports a Joint View of Two Classification Systems for Age-Related Macular Degeneration.

Purpose: The purpose of this study was to evaluate the utility of combining the Clinical Classification (CC) and the Three Continent age-related macular degeneration (AMD) Consortium Severity Scale (3CACSS) for classification of AMD. Methods: In two independent cross-sectional datasets of our population-based AugUR study (Altersbezogene Untersuchungen zur Gesundheit der Universität Regensburg), we graded AMD via color fundus images applying two established classification systems (CC and 3CACSS). We calculated the genetic risk score (GRS) across 50 previously identified variants for late AMD, its association via logistic regression, and area under the curve (AUC) for each AMD stage. Results: We analyzed 2188 persons aged 70 to 95 years. When comparing the two classification systems, we found a distinct pattern: CC "age-related changes" and CC "early AMD" distinguished individuals with 3CACSS "no AMD"; 3CACSS "mild/moderate/severe early AMD" stages, and distinguished CC "intermediate AMD". This suggested a 7-step scale combining the 2 systems: (i) "no AMD", (ii) "age-related changes", (iii) "very early AMD", (i.e. CC "early"), (iv) "mild early AMD", (v) "moderate early AMD", (vi) "severe early AMD", and (vii) "late AMD". GRS association and diagnostic accuracy increased stepwise by increased AMD severity in the 7-step scale and by increased restriction of controls (e.g. for CC "no AMD without age-related changes": AUC = 55.1%, 95% confidence interval [CI] = 51.6, 58.6, AUC = 62.3%, 95% CI = 59.1, 65.6, AUC = 63.8%, 95% CI = 59.3, 68.3, AUC = 78.1%, 95% CI = 73.6, 82.5, AUC = 82.2%, 95% CI = 78.4, 86.0, and AUC = 79.2%, 95% CI = 75.4, 83.0). A stepwise increase was also observed by increased drusen size and area. Conclusions: The utility of a 7-step scale is supported by our clinical and GRS data. This harmonization and full data integration provides an immediate simplification over using either CC or 3CACSS and helps to sharpen the control group.

Dual-path multi-scale context dense aggregation network for retinal vessel segmentation.

There has been steady progress in the field of deep learning-based blood vessel segmentation. However, several challenging issues still continue to limit its progress, including inadequate sample sizes, the neglect of contextual information, and the loss of microvascular details. To address these limitations, we propose a dual-path deep learning framework for blood vessel segmentation. In our framework, the fundus images are divided into concentric patches with different scales to alleviate the overfitting problem. Then, a Multi-scale Context Dense Aggregation Network (MCDAU-Net) is proposed to accurately extract the blood vessel boundaries from these patches. In MCDAU-Net, a Cascaded Dilated Spatial Pyramid Pooling (CDSPP) module is designed and incorporated into intermediate layers of the model, enhancing the receptive field and producing feature maps enriched with contextual information. To improve segmentation performance for low-contrast vessels, we propose an InceptionConv (IConv) module, which can explore deeper semantic features and suppress the propagation of non-vessel information. Furthermore, we design a Multi-scale Adaptive Feature Aggregation (MAFA) module to fuse the multi-scale feature by assigning adaptive weight coefficients to different feature maps through skip connections. Finally, to explore the complementary contextual information and enhance the continuity of microvascular structures, a fusion module is designed to combine the segmentation results obtained from patches of different sizes, achieving fine microvascular segmentation performance. In order to assess the effectiveness of our approach, we conducted evaluations on three widely-used public datasets: DRIVE, CHASE-DB1, and STARE. Our findings reveal a remarkable advancement over the current state-of-the-art (SOTA) techniques, with the mean values of Se and F1 scores being an increase of 7.9% and 4.7%, respectively. The code is available at https://github.com/bai101315/MCDAU-Net.

RP2 X-LINKED RETINITIS PIGMENTOSA CARRIER STATE PRESENTING WITH VASCULAR LEAKAGE AND UNILATERAL MACULAR ATROPHY.

PURPOSE: We describe the unusual clinical presentation of a 33-year-old woman subsequently identified as a carrier of RP2-associated X-linked retinitis pigmentosa. METHODS: Case report. RESULTS: A 33-year-old woman without a known family history of retinal disease presented with unilateral reduced visual acuity and central scotoma in the left eye. Examination showed underlying macular atrophy in the left eye and a bilateral tapetal-like reflex. Full-field electroretinogram was abnormal in the left eye but normal in the right eye. Notable findings on wide-field imaging included bilateral peripheral vascular leakage on fluorescein angiography and a bilaterally symmetric radial pattern of hyperfluorescence on fundus autofluorescence. Genetic testing demonstrated a pathogenic variant in the gene RP2 confirming that she was a carrier of X-linked retinitis pigmentosa. CONCLUSION: We describe clinical features of the carrier state of RP2-XLRP and expand potential findings to include peripheral vascular leakage. This case highlights the importance of awareness of the carrier state, particularly if a family history cannot be provided.

A fundus image classification framework for learning with noisy labels.

Fundus images are widely used in the screening and diagnosis of eye diseases. Current classification algorithms for computer-aided diagnosis in fundus images rely on large amounts of data with reliable labels. However, the appearance of noisy labels degrades the performance of data-dependent algorithms, such as supervised deep learning. A noisy label learning framework suitable for the multiclass classification of fundus diseases is presented in this paper, which combines data cleansing (DC), adaptive negative learning (ANL), and sharpness-aware minimization (SAM) modules. Firstly, the DC module filters the noisy labels in the training dataset based on the prediction confidence. Then, the ANL module modifies the loss function by choosing complementary labels, which are neither the given labels nor the labels with the highest confidence. Moreover, for better generalization, the SAM module is applied by simultaneously optimizing the loss and its sharpness. Extensive experiments on both private and public datasets show that our method greatly promotes the performance for classification of multiple fundus diseases with noisy labels.

Primary versus Secondary Elevations in Fundus Autofluorescence.

The method of quantitative fundus autofluorescence (qAF) can be used to assess the levels of bisretinoids in retinal pigment epithelium (RPE) cells so as to aid the interpretation and management of a variety of retinal conditions. In this review, we focused on seven retinal diseases to highlight the possible pathways to increased fundus autofluorescence. ABCA4- and RDH12-associated diseases benefit from known mechanisms whereby gene malfunctioning leads to elevated bisretinoid levels in RPE cells. On the other hand, peripherin2/RDS-associated disease (PRPH2/RDS), retinitis pigmentosa (RP), central serous chorioretinopathy (CSC), acute zonal occult outer retinopathy (AZOOR), and ceramide kinase like (CERKL)-associated retinal degeneration all express abnormally high fundus autofluorescence levels without a demonstrated pathophysiological pathway for bisretinoid elevation. We suggest that, while a known link from gene mutation to increased production of bisretinoids (as in ABCA4- and RDH12-associated diseases) causes primary elevation in fundus autofluorescence, a secondary autofluorescence elevation also exists, where an impairment and degeneration of photoreceptor cells by various causes leads to an increase in bisretinoid levels in RPE cells.

Cat-scratch disease manifesting as uveitis and binocular fundus nodular lesions: a case report.

BACKGROUND: Cat-scratch disease typically presents with various ocular manifestations such as uveitis, vitritis, retinitis, retinochoroiditis, and optic neuritis. However, fundus nodular lesions was rarely reported. In our study, we reported a case of Cat-Scratch disease with binocular fundus nodular lesions. CASE PRESENTATION: An 11-year old male presented with uveitis in the right eye and bilateral fundus nodular lesions after indirect contact with unvaccinated cats. Comprehensive ancillary examinations including wide-angle fundus photography, ultrasonography, fluorescein fundus angiography, optical coherence tomography, and orbital magnetic resonance imaging were performed to elucidate the multidimensional features of the binocular lesions. Metagenomics next-generation sequencing was utilized to confirm the diagnosis of Cat-scratch disease. The patient's condition showed improvement after a 6-month combination treatment regimen involving systemic administration of doxycycline hyclate and methylprednisolone tablets, as well as local application of mydriatic and corticosteroid eye drops. CONCLUSIONS: We firstly reported a case of Cat-scratch disease presenting simultaneously with uveitis and fundus nodular lesions caused by Bartonella henselae infection in a child. Timely diagnosis and treatment with antibiotics and corticosteroids showed promising outcomes for the prognosis of these ocular disorders.

Semi-supervised segmentation of retinoblastoma tumors in fundus images.

Retinoblastoma is a rare form of cancer that predominantly affects young children as the primary intraocular malignancy. Studies conducted in developed and some developing countries have revealed that early detection can successfully cure over 90% of children with retinoblastoma. An unusual white reflection in the pupil is the most common presenting symptom. Depending on the tumor size, shape, and location, medical experts may opt for different approaches and treatments, with the results varying significantly due to the high reliance on prior knowledge and experience. This study aims to present a model based on semi-supervised machine learning that will yield segmentation results comparable to those achieved by medical experts. First, the Gaussian mixture model is utilized to detect abnormalities in approximately 4200 fundus images. Due to the high computational cost of this process, the results of this approach are then used to train a cost-effective model for the same purpose. The proposed model demonstrated promising results in extracting highly detailed boundaries in fundus images. Using the Sørensen-Dice coefficient as the comparison metric for segmentation tasks, an average accuracy of 93% on evaluation data was achieved.

Learning from small data: Classifying sex from retinal images via deep learning.

Deep learning (DL) techniques have seen tremendous interest in medical imaging, particularly in the use of convolutional neural networks (CNNs) for the development of automated diagnostic tools. The facility of its non-invasive acquisition makes retinal fundus imaging particularly amenable to such automated approaches. Recent work in the analysis of fundus images using CNNs relies on access to massive datasets for training and validation, composed of hundreds of thousands of images. However, data residency and data privacy restrictions stymie the applicability of this approach in medical settings where patient confidentiality is a mandate. Here, we showcase results for the performance of DL on small datasets to classify patient sex from fundus images-a trait thought not to be present or quantifiable in fundus images until recently. Specifically, we fine-tune a Resnet-152 model whose last layer has been modified to a fully-connected layer for binary classification. We carried out several experiments to assess performance in the small dataset context using one private (DOVS) and one public (ODIR) data source. Our models, developed using approximately 2500 fundus images, achieved test AUC scores of up to 0.72 (95% CI: [0.67, 0.77]). This corresponds to a mere 25% decrease in performance despite a nearly 1000-fold decrease in the dataset size compared to prior results in the literature. Our results show that binary classification, even with a hard task such as sex categorization from retinal fundus images, is possible with very small datasets. Our domain adaptation results show that models trained with one distribution of images may generalize well to an independent external source, as in the case of models trained on DOVS and tested on ODIR. Our results also show that eliminating poor quality images may hamper training of the CNN due to reducing the already small dataset size even further. Nevertheless, using high quality images may be an important factor as evidenced by superior generalizability of results in the domain adaptation experiments. Finally, our work shows that ensembling is an important tool in maximizing performance of deep CNNs in the context of small development datasets.

Can Glaucoma Suspect Data Help to Improve the Performance of Glaucoma Diagnosis?

Purpose: The presence of imbalanced datasets in medical applications can negatively affect deep learning methods. This study aims to investigate how the performance of convolutional neural networks (CNNs) for glaucoma diagnosis can be improved by addressing imbalanced learning issues through utilizing glaucoma suspect samples, which are often excluded from studies because they are a mixture of healthy and preperimetric glaucomatous eyes, in a semi-supervised learning approach. Methods: A baseline 3D CNN was developed and trained on a real-world glaucoma dataset, which is naturally imbalanced (like many other real-world medical datasets). Then, three methods, including reweighting samples, data resampling to form balanced batches, and semi-supervised learning on glaucoma suspect data were applied to practically assess their impacts on the performances of the trained methods. Results: The proposed method achieved a mean accuracy of 95.24%, an F1 score of 97.42%, and an area under the curve of receiver operating characteristic (AUC ROC) of 95.64%, whereas the corresponding results for the traditional supervised training using weighted cross-entropy loss were 92.88%, 96.12%, and 92.72%, respectively. The obtained results show statistically significant improvements in all metrics. Conclusions: Exploiting glaucoma suspect eyes in a semi-supervised learning method coupled with resampling can improve glaucoma diagnosis performance by mitigating imbalanced learning issues. Translational Relevance: Clinical imbalanced datasets may negatively affect medical applications of deep learning. Utilizing data with uncertain diagnosis, such as glaucoma suspects, through a combination of semi-supervised learning and class-imbalanced learning strategies can partially address the problems of having limited data and learning on imbalanced datasets.

Diabetic microaneurysms detected by fluorescein angiography spatially correlate with regions of macular ischemia delineated by optical coherence tomography angiography.

Purpose: To characterize the relationship between diabetic macular ischemia (DMI) delineated by optical coherence tomography angiography (OCTA) and microaneurysms (MAs) identified by fundus fluorescein angiography (FFA). Methods: Patients with diabetic retinopathy (DR) who underwent OCTA and FFA were retrospectively identified. FFA images were cropped and aligned with their respective OCTA images using i2k Align Retina software (Dual-Align, Clifton Park, NY, USA). Foveal avascular zone (FAZ) and ischemic areas were manually delineated on OCTA images, and MAs were marked on the corresponding FFA images before overlaying paired scans for analysis (ImageJ; National Institutes of Health, Bethesda, MD, USA). Results: Twenty-eight eyes of 20 patients were included. The average number of MAs identified in cropped FFA images was 127 ± 42. More DMI was noted in the superficial capillary plexus (SCP; 36 ± 13%) compared to the deep capillary plexus (DCP; 28 ± 14%, P < 0.001). Similarly, more MAs were associated with ischemic areas in SCP compared to DCP (92.0 ± 35.0 vs. 76.8 ± 36.5, P < 0.001). Most MAs bordered ischemic areas; fewer than 10% localized inside these regions. As DMI area increased, so did associated MAs (SCP: r = 0.695, P < 0.001; DCP: r = 0.726, P < 0.001). Density of MAs surrounding FAZ (7.7 ± 6.0 MAs/mm2) was similar to other DMI areas (SCP: 7.0 ± 4.0 MAs/mm2, P = 0.478; DCP: 9.2 ± 10.9 MAs/mm2, P = 0.394). Conclusion: MAs identified in FFA strongly associate with, and border areas of, DMI delineated by OCTA. Although more MAs are localized to SCP ischemia, the concentration of MAs associated with DCP ischemia is greater. By contrast, few MAs are present inside low-flow regions, likely because capillary loss is associated with their regression.

Stimulus-guided adaptive transformer network for retinal blood vessel segmentation in fundus images.

Automated retinal blood vessel segmentation in fundus images provides important evidence to ophthalmologists in coping with prevalent ocular diseases in an efficient and non-invasive way. However, segmenting blood vessels in fundus images is a challenging task, due to the high variety in scale and appearance of blood vessels and the high similarity in visual features between the lesions and retinal vascular. Inspired by the way that the visual cortex adaptively responds to the type of stimulus, we propose a Stimulus-Guided Adaptive Transformer Network (SGAT-Net) for accurate retinal blood vessel segmentation. It entails a Stimulus-Guided Adaptive Module (SGA-Module) that can extract local-global compound features based on inductive bias and self-attention mechanism. Alongside a light-weight residual encoder (ResEncoder) structure capturing the relevant details of appearance, a Stimulus-Guided Adaptive Pooling Transformer (SGAP-Former) is introduced to reweight the maximum and average pooling to enrich the contextual embedding representation while suppressing the redundant information. Moreover, a Stimulus-Guided Adaptive Feature Fusion (SGAFF) module is designed to adaptively emphasize the local details and global context and fuse them in the latent space to adjust the receptive field (RF) based on the task. The evaluation is implemented on the largest fundus image dataset (FIVES) and three popular retinal image datasets (DRIVE, STARE, CHASEDB1). Experimental results show that the proposed method achieves a competitive performance over the other existing method, with a clear advantage in avoiding errors that commonly happen in areas with highly similar visual features. The sourcecode is publicly available at: https://github.com/Gins-07/SGAT.

SPONTANEOUS CLOSURE AND RECURRENT OPENING TIMES TWO OF A MACULAR HOLE IN A SURGICALLY NAIVE EYE.

PURPOSE: To report a case of an idiopathic macular hole with recurrent opening and spontaneous closure in a surgically naive eye. METHODS: A retrospective review of medical records was performed in addition to a review of the current literature. RESULTS: An 82-year-old man was referred for the management of a full-thickness macular hole in the right eye. Visual acuity was 20/60, and dilated fundus examination was notable for a posterior vitreous detachment, macular hole, and mild epiretinal membrane. Optical coherence tomography confirmed the presence of a full-thickness macular hole. The patient declined surgical intervention and elected to observe. Five weeks later, optical coherence tomography confirmed spontaneous closure. One year later, a recurrent partial thickness outer retinal hole was noted on dilated fundus examination and optical coherence tomography that subsequently spontaneously closed for the second time. The following year, the patient represented with a new scotoma and metamorphopsia and was found to have a full-thickness macular hole. This time the patient was elected for surgical intervention (25-gauge pars plana vitrectomy, epiretinal membrane peel, and 14% C3F8), resulting in closure of the macular hole and improvement in visual acuity to 20/25+1. CONCLUSION: This case highlights a rare presentation of a see-saw pattern of opening and closing of a macular hole in a treatment-naive eye. The presence of a posterior vitreous detachment and epiretinal membrane suggests that other factors than anterior-posterior and tangential traction may be a contributing in the formation and closure of idiopathic macular holes.