Optical Coherence Tomography Images Research Articles

Self-supervised based clustering for retinal optical coherence tomography images.

In response to the inadequacy of manual analysis in meeting the rising demand for retinal optical coherence tomography (OCT) images, a self-supervised learning-based clustering model was implemented. A public dataset was utilized, with 83,484 OCT images with categories of choroidal neovascularization (CNV), diabetic macular edema (DME), drusen, and normal fundus. This study employed the Semantic Pseudo Labeling for Image Clustering (SPICE) framework, a self-supervised learning-based method, to cluster unlabeled OCT images into binary and four categories, and the performances were compared with baseline models. We also analysed feature distribution using t-SNE, and explored the cluster centers, attention maps, and misclassified images. In addition, DME and CNV subsets were clustered binarily, and the results were interpreted by two retinal specialists. SPICE demonstrated superior performance in binary and four categories classification tasks, achieving the accuracy of 0.886 and 0.846, respectively. In t-SNE analysis, the four types exhibited significant clustering into distinct groups. The cluster centers corresponded to the human labels, and the heat map revealed that the model focused on important biomarkers. The misclassified images exposed similar features to the inaccurate classes. The model also grouped DME and CNV into two distinct categories respectively. Self-supervised clustering effectively distinguished disease variances and revealed common features, with a notable capability to detect disease heterogeneity through biomarkers.

Plaque phenotyping using deep learning in optical coherence tomography imaging

Abstract Background Artificial intelligence (AI) techniques have emerged as powerful tools in various computer vision tasks, particularly in classification, detection, and segmentation. Purpose This study aimed to develop a deep learning model for categorizing optical coherence tomography (OCT) frames into different plaque phenotypes, including thin-cap fibroatheroma (TCFA) and fibroatheroma. Methods A prospective multicenter observational study enrolled 168 patients who underwent percutaneous coronary intervention (PCI) and received OCT imaging for non-target lesions with angiographically estimated diameter stenosis greater than 30% (177 vessels). Among them, 129 patients (135 vessels) underwent follow-up OCT imaging at 1 year. A total of 45,072 OCT frames were randomly divided into training and validation sets in a 4:1 ratio. An EfficientNet-B3 convolutional neural network model was developed to classify OCT frames into TCFA, thick-cap fibroatheroma, non-fibroatheroma, adaptive intimal thickening, and normal vessel categories. Results A confusion matrix comparing manually assessed plaque phenotypes with AI-predicted phenotypes was presented in Table. In the validation datasets (9,015 frames), the overall accuracy of the model was 85.1%. For predicting the presence of TCFA, sensitivity was 81.1%, specificity was 98.8%, positive predictive value was 63.9%, and negative predictive value was 99.5%. Regarding the presence of fibroatheroma, sensitivity was 90.3%, specificity was 89.4%, positive predictive value was 93.1%, and negative predictive value was 85.4%. Conclusions The deep learning algorithm demonstrated high accuracy in detecting TCFA and fibroatheroma with excellent reproducibility. This automated process enables operators to promptly identify high-risk lesions during PCI, potentially leading to improved clinical outcomes.

Characterization of plaque hemorrhage with optical coherence tomography imaging in statin-treated patients with coronary artery disease: comparison with noncontrast T1-weighted MRI

Abstract Background Plaque hemorrhage is a phenotypic feature of coronary lesions associated with future coronary events. Noncontrast T1-weighted MRI has a potential to non-invasively identify plaque hemorrhage. Plaque to myocardium ratio (PMR) evaluated by noncontrast T1-weighted MRI has been shown to pathohistologically correspond to plaque hemorrhage. In several case reports comparing OCT images with coronary specimens, low intensity area without its attenuation was observed at lesions presenting plaque hemorrhage. However, whether the association of this OCT-derived feature with PMR remains to be elucidated yet. Purpose To characterize plaque hemorrhage in vivo by comparing with OCT and noncontrast T1-weighted MRI. Methods The current study prospectively enrolled 80 statin-treated patients with coronary artery disease requiring elective PCI. A total of 103 target lesions were imaged by both noncontrast T1-weighted MRI and OCT prior to PCI. Plaque hemorrhage on noncontrast T1-weighted MRI was defined as PMR>1.0. On OCT imaging analysis, low intensity area without attenuation (LIA) was defined according to the following criteria; (1) area presenting homogeneous low intensity signals, (2) clear border and (3) no signal attenuation behind its area. The maximum arc of LIA was manually measured. OCT-derived plaque features were compared in lesions with and without PMR>1.0. Results PMR>1.0 was observed at 49.5% of analyzed lesions. Patients with PMR>1.0 were more likely older (72 vs. 66 years, p=0.02), whereas there was no significant difference in on-treatment LDL-C levels (median: 80 vs. 77 mg/dL, p=0.62) (Table). On lesion-based analysis, while fibrous cap thickness was similar between the two groups (80 vs. 100um, p=0.10), a larger lipid arc (270 vs. 230º, p=0.02) and a greater frequency of macrophage (90.2 vs. 53.9%, p<0.01), layered plaque (62.8 vs. 32.7%, p=0.003) and cholesterol crystal (70.6 vs. 30.8%, p<0.001) were observed at lesions with PMR>1.0. Furthermore, lesions with PMR>1.0 more frequently exhibited LIA (66.7 vs. 19.2%, p<0.001) and its greater maximum arc (33 vs. 0º, p<0.001) (Figure). Of note, PMR was significantly correlated to the maximum arc of LIA (r=0.51, p<0.001) (Figure). Multivariate analysis demonstrated LIA and cholesterol crystal as independent OCT-derived features associated with PMR>1.0 (Table). Conclusion In statin-treated patients, 49.5% of analyzed lesions presented PMR>1.0. Coronary lesions with PMR>1.0 more frequently harbored LIA and cholesterol crystals, in addition to other vulnerable OCT features. Our findings suggest OCT-derived LIA as a potential signature of plaque hemorrhage. Given that cholesterol crystal was another feature of PMR>1.0, plaque inflammation driven by cholesterol crystallization may exist at plaque hemorrhage despite statin use. Whether anti-inflammation agent could modulate plaque hemorrhage requires future investigation.

Relationship between plaque burden and plaque vulnerability in patients with acute coronary syndromes versus chronic coronary syndrome

Abstract Background The relationship between plaque burden and plaque vulnerability based on clinical presentations has not been fully investigated. Objective The aim of this study was to compare the relationship between plaque burden and plaque vulnerability in patients with acute coronary syndromes (ACS) versus chronic coronary syndrome (CCS). Methods This is an observational, single-center cohort study. Patients who underwent both coronary computed tomography angiography (CTA) and optical coherence tomography (OCT) before coronary intervention were enrolled. All OCT and CTA images were analyzed at the core laboratory. All plaques were detected in culprit vessels using CTA, and total plaque volume (TPV) and OCT features were assessed at the corresponding sites. A layered plaque was defined as a plaque with one or more layers of different optical densities and a clear demarcation from underlying components. The layer arc was measured at 1-mm intervals. The layer index, which is equivalent to layer volume, was defined as mean layer arc × layer length. All plaques were divided into three groups according to the tertile range of TPV (low TPV: <96.5 mm3, moderate TPV: 96.5 – 164.7 mm3, high TPV: ≥164.8 mm3). The effect modification of clinical presentations (CCS or ACS) on the association between TPV and OCT features was investigated. Results A total of 990 plaques were imaged by OCT in 419 patients: 445 plaques in 190 (45.3%) patients who presented with ACS and 545 in 229 (54.7%) with CCS. The prevalence of thin-cap fibroatheroma (TCFA) was higher as plaque volume increased in both groups (low vs. moderate vs. high TPV group: 18.4% vs. 37.8% vs. 52.3% in ACS; 16.2% vs. 19.9% vs. 30.6% in CCS; interaction P=0.066). However, the effect modification of clinical presentations on the associations of TPV with macrophage, lipid arc, and layer index was statistically significant. Macrophage was more prevalent in plaques with greater TPV in patients who presented with ACS but not in those who presented with CCS (low vs. moderate vs. high TPV group: macrophage 57.4% vs. 71.8% vs. 82.4% in ACS; 63.4% vs. 67.8% vs. 66.7% in CCS; interaction P=0.004). Lipid arc was more strongly associated with TPV in patients who presented with ACS than in those with CCS (low vs. moderate vs. high TPV group: 116 [76-157] vs. 166 [115-210] vs. 175 [140-221] in ACS; 118 [80-154] vs. 143 [102-187] vs. 150 [102-199] in CCS; interaction P=0.028). Conversely, layer index was more weakly associated with TPV in patients who presented with ACS than in those who presented with CCS (low vs. moderate vs. high TPV group: 0 [0-0] vs. 0 [0-365] vs. 0 [0-903] in ACS; 0 [0-756] vs. 235 [0-1266] vs. 552 [0-1704] in CCS; interaction P=0.046) (Figure 1). Conclusions Greater plaque burden was closely related to the level of plaque vulnerability in ACS, whereas this relationship was less obvious in CCS.Figure 1

Trend in culprit plaque phenotype in ACS patients over the past 10 years

Abstract Background Although the incidence of acute myocardial infarction has been declining over the past decade, the etiology of this trend is unknown. One possible explanation could be changes in coronary plaque phenotype. Purpose The aim of this study was to investigate the changes in plaque characteristics over the last 10 years. Methods Patients who presented with acute coronary syndrome and underwent preintervention optical coherence tomography (OCT) imaging from 2008 to 2018 were included. Patients were divided into 4 periods: (P1) 2008-2010, (P2) 2011-2013, (P3) 2014-2015, and (P4) 2016-2018. Plaque characteristics at culprit lesion in each period were compared. Results Among 1144 patients, 594 (51.9%) presented with ST-segment–elevation myocardial infarction (STEMI). The prevalence of thin-cap fibroatheroma decreased from P1 to P4 in both STEMI group (57.4% to 24.6%, P<0.001) and non-ST-segment elevation acute coronary syndrome (NSTE-ACS) group (40.4% to 17.0%, P<0.001) (Figure). Conclusions The prevalence of OCT features of plaque vulnerability has decreased over the past 10 years. A decrease in plaque vulnerability may explain the favorable trend in cardiovascular events.Prevalence of thin-cap fibroatheroma

Lesion-level changes induced by low-density lipoprotein cholesterol lowering therapy in patients with acute myocardial infarction: the pre-specified PACMAN-AMI lesion-level analysis

Abstract Background Previous intracoronary imaging studies investigated atherosclerotic changes induced by lipid-lowering therapy in extensive coronary segments, based on anatomical landmarks and irrespective of the baseline atheroma burden. The effects of lipid-lowering therapy focusing on coronary lesions with advanced atherosclerotic features and at presumably higher risk for future events remains largely unknown. Purpose To provide a lesion-level analysis reporting changes in high-risk atherosclerotic features and plaque phenotypes induced by alirocumab versus placebo in addition to high-intensity statin therapy in nonculprit coronary lesions of patients admitted with acute myocardial infarction (AMI). Methods In this pre-specified lesion-level analysis of the PACMAN-AMI trial, coronary lesions were identified as segments with plaque burden ≥40% defined by intravascular ultrasound (IVUS). IVUS, near-infrared spectroscopy (NIRS) and optical coherence tomography (OCT) images at baseline and 52-week follow-up were manually matched and lesion-level imaging outcome measures were calculated by analysts blinded to treatment allocation and patient clinical data. The evolution of high-risk plaque phenotypes as defined by each imaging modality (IVUS-defined plaque burden ≥70%, NIRS-defined maximum LCBI 4 mm ≥400, OCT-defined thin-cap fibroatheroma <65µm) was investigated. Results Overall, 591 lesions with available baseline and follow-up imaging were included. Of these, 287 were identified in 118 patients (214 vessels) in the alirocumab arm, and 304 lesions were identified in 127 patients (239 vessels) in the placebo arm. At lesion-level, mean change in percent atheroma volume (PAV) by IVUS was -4.86% with alirocumab vs -2.78% with placebo (difference, -2.02 [-3.00 to -1.05], p<.001), substantially greater than that at vessel-level (Figure 1). At the minimum lumen area (MLA) site, mean change in PAV was -10.14% with alirocumab vs -6.70% with placebo (difference, -3.36 (-4.98 to -1.75, p<0.001). MLA increased by 0.15 mm2 with alirocumab and decreased by 0.07 mm2 in the placebo arm (difference, +0.21, 0.01 to 0.41, p=0.036), whereas arterial remodeling was similar between arms (-0.73±1.61 in alirocumab vs -0.63±1.31 in placebo group, p=0.482). Among lesions with high-risk plaque phenotypes at baseline, those in the alirocumab arm more frequently showed a more stable and less lipid-rich plaque phenotype at follow-up than those in the placebo arm (Figure 2). Conclusions At lesion-level, intense lipid-lowering therapy with alirocumab in addition to high-intensity statin therapy induced extensive PAV regression, substantially greater than described in previous vessel-level analyses. Compared to statin therapy alone, alirocumab treatment was associated with a greater enlargement of the lesion MLA, similar negative arterial remodeling, and a more frequent transition of high-risk plaque phenotypes into more stable, less lipid-rich plaque phenotypes.Figure 1.Figure 2

Epiretinal membrane: an overview and update.

Epiretinal membrane (ERM) is a frequently diagnosed macular disease associated with aging, characterized by a fibrous membrane forming on the internal limiting membrane (ILM) and leading to visual dysfunctions such as metamorphopsia. Various hypotheses regarding the pathology of metamorphopsia have been proposed; however, the complete pathophysiologic mechanism underlying ERM remains unclear. Optical coherence tomography (OCT) provides detailed images enabling precise diagnosis and characterization of ERM, with several recent studies using the latest OCT imaging techniques. Surgical removal of ERM is the only treatment option; however, criteria for surgical intervention are not established, complicating the decision-making processes. Furthermore, the debate on whether simultaneous peeling of the ILM during ERM surgery enhances outcomes or poses unnecessary risks is ongoing, with no definite conclusion having yet been reached. This review also focuses on epiretinal proliferation, which is different from ERM and is characteristic of lamellar macular hole (LMH). Recently, diagnostic criteria for LMH and related diseases were proposed. Reports on effective surgical procedures for LMH exist, although more research is needed to confirm the long-term outcomes. Thus, this review article aims to provide an overview and updated knowledge of ERM, LMH, and related diseases.

Deep Learning With Optical Coherence Tomography for Melanoma Identification and Risk Prediction.

Malignant melanoma is the most severe skin cancer with a rising incidence rate. Several noninvasive image techniques and computer-aided diagnosis systems have been developed to help find melanoma in its early stages. However, most previous research utilized dermoscopic images to build a diagnosis model, and only a few used prospective datasets. This study develops and evaluates a convolutional neural network (CNN) for melanoma identification and risk prediction using optical coherence tomography (OCT) imaging of mice skin. Longitudinal tests are performed on four animal models: melanoma mice, dysplastic nevus mice, and their respective controls. The CNN classifies melanoma and healthy tissues with high sensitivity (0.99) and specificity (0.98) and also assigns a risk score to each image based on the probability of melanoma presence, which may facilitate early diagnosis and management of melanoma in clinical settings.

Bioresorbable stent unloading during percutaneous coronary intervention: Early detection and management

In this letter, we comment on a recent case report by Sun et al in the World Journal of Cardiology . The report describes the successful management of a rare complication: The unloading or detachment of a bioresorbable stent (BRS) during percutaneous coronary intervention (PCI) in a male patient. The unloading of BRS was detected via angiography and intravascular ultrasound (IVUS) imaging of the left coronary artery and left anterior descending artery. Although this case is interesting, the authors’ report lacked crucial details. Specifically, insufficient information about the type of BRS used, potential causes of BRS unloading, or whether optical coherence tomography (OCT) imaging for coronary arteries was performed before, during, or after PCI. The OCT imaging of coronary arteries before PCI can potentially prevent BRS unloading due to its higher resolution compared to IVUS. In addition, despite detecting myocardial bridging during the PCI, the authors did not provide any details regarding this variation. Here we discuss the various types of BRS, the importance of OCT in PCI, and the clinical relevance of myocardial bridging.

Multimodal Optical Imaging of Ex Vivo Fallopian Tubes to Distinguish Early and Occult Tubo-Ovarian Cancers

Background: There are currently no effective screening measures to detect early or occult tubo-ovarian cancers, resulting in late-stage detection and high mortality. This work explores whether an optical imaging catheter can detect early-stage tubo-ovarian cancers or precursor lesions where they originate in the fallopian tubes. Methods: This device collects co-registered optical coherence tomography (OCT) and autofluorescence imaging (AFI). OCT provides three-dimensional assessment of underlying tissue structures; autofluorescence imaging provides functional contrast of endogenous fluorophores. Ex vivo fallopian tubes (n = 28; n = 7 cancer patients) are imaged; we present methods for the calculation of and analyze eleven imaging biomarkers related to fluorescence, optical attenuation, and OCT texture for their potential to detect tubo-ovarian cancers and other lesions of interest. Results: We visualize folded plicae, vessel-like structures, tissue layering, hemosiderin deposits, and regions of fibrotic change. High-grade serous ovarian carcinoma appears as reduced autofluorescence paired with homogenous OCT and reduced mean optical attenuation. Specimens containing cancerous lesions demonstrate a significant increase in median autofluorescence intensity and decrease in Shannon entropy compared to specimens with no lesion. Non-cancerous specimens demonstrate an increase in optical attenuation in the fimbriae when compared to the isthmus or the ampulla. Conclusions: We conclude that this approach shows promise and merits further investigation of its diagnostic potential.

Persistence and recurrence after removal of idiopathic epiretinal membrane.

To analyse the incidence of persistence and recurrence after the peeling of idiopathic epiretinal membrane (ERM) and to describe its clinical features. This retrospective study included 666 eyes (645 patients) that underwent macular surgery for ERM removal. Optical coherence tomographic (OCT) images taken within three months after surgery and at the following visits, clinical parameters and surgery related factors were analysed to investigate the incidence and associated factors of ERM persistence and recurrence. Postoperative ERM types were categorised depending on the size ( < 100 µm, ≥100 µm) and the location (foveal, parafoveal, outside the parafovea) RESULTS: The mean follow-up time was 29.4 months. ERM persistence (examination within 3 months) was found in 29.6% of all eyes. Only 1.9% of the eyes presented foveal ERM persistence. Foveal recurrence, defined as reappearance or growth of persistent ERM covering the fovea, was found in 8.2%. In 84.4% of eyes with foveal ERM recurrence, postoperative persistence of ERM of varying severity were identified. None of the pre-operative or surgery related factors were found significantly associated with ERM recurrence. Persistent ERM within the parafovea was the most significant risk factor for foveal ERM recurrence. Recurrence of ERM is generally preceded by the persistence of ERM fragments found in the early postoperative period. Growth of ERM persistence from the parafoveal region was often the origin of foveal ERM recurrence. Insufficient peeling seems to be the most significant predisposing factor for foveal ERM recurrence.

Multilayered Inverted Internal Limiting Membrane Flap Technique in Optic Disc Pit Maculopathy.

To evaluate the anatomical and visual outcomes of the multilayered inverted internal limiting membrane (ML-ILM) flap technique in the treatment of optic disc pit maculopathy (ODPM). In this retrospective interventional case series, medical records and macular spectral-domain optical coherence tomography images of patients who underwent combined pars plana vitrectomy with ML-ILM flap surgery for ODPM were analyzed. Best-corrected visual acuity (BCVA) and central macular thickness (CMT) at postoperative 6 months were compared with baseline findings. Intraoperative and postoperative complications, fluid resolution time, and recurrence during follow-up were recorded. Five eyes of 5 patients with ODPM were included in the study. According to the preoperative macular fluid characteristics, 2 patients had only intraretinal fluid, while 3 patients had intraretinal and subretinal fluid. The preoperative median BCVA was 1.0 logarithm of the minimum angle of resolution (logMAR) (range, 1.0-1.3 logMAR), and the CMT was 560 μm (range, 452-667 μm). At the 6-month postoperative follow-up, the median BCVA was 0.40 logMAR (range, 0.1-0.7 logMAR), and CMT was 315 μm (range, 265-326 μm) (p=0.042 and p=0.043, respectively). During the 6-month follow-up period, no recurrence or full-thickness macular hole formation was observed. The ML-ILM flap technique is a preferable surgical option to achieve both high anatomical and functional success and flap stabilization.

Distribution and associated factors of choroidal thickness in highly myopic eyes-a real-world study based on a Chinese population.

To measure the subfoveal choroidal thickness (SFCT) in highly myopic eyes at different locations using enhanced depth imaging spectral-domain optical coherence tomography (EDI SD-OCT). To identify the ocular and systemic risk factors associated with choroidal thinning in high myopia. Based on the Beijing Eye Study, a detailed ophthalmic examination was performed including EDI SD-OCT for the measurement of SFCT. OCT images were obtained from 103 highly myopic eyes (≥ -6.00 dioptres) and 227 normal eyes randomly selected from the baseline population, matched for age and sex. The mean SFCT was 110.6 ± 85.2 μm in highly myopic eyes (range, 3-395 μm). Mean regional choroidal thickness was lowest on the nasal and inferior sides of the macula, and slightly higher on the temporal and superior sides than at the fovea. On multivariate analysis, SFCT was associated with age (b = -0.48; P < 0.001), axial length (b = -0.44; P < 0.001), gender (b = -0.31; P < 0.05) and staphyloma (b = -0.26; P = 0.05). In highly myopic eyes, SFCT decreased by 5.1 μm/year of age, by 9.2 µm/D of myopia, and by 22.6 µm/mm of axial length. The SFCT decreases with age and increased axial length in highly myopic eyes. The formation of a posterior staphyloma has been identified as a major contributor to choroidal thinning and is therefore a reliable indicator for risk management. The involvement of choroidal abnormalities may be a significant factor in the development of myopic degeneration.

Impact of retinal traction induced by epiretinal membrane on aniseikonia

We investigated the effect of retinal traction caused by epiretinal membranes (ERMs) on aniseikonia and retinal microstructures in 81 unilateral ERMs. Retinal traction was quantified by measuring the maximum depth of the retinal fold (MDRF) using en face optical coherence tomography (OCT) images. Measurements included the mean inner nuclear layer (INL), outer plexiform layer (OPL), outer nuclear layer (ONL), central retinal thickness (CRT), and interocular ratios of the foveal avascular zone (FAZ) area (FAZ ratio). Significant correlations were found between the preoperative MDRF and preoperative aniseikonia (P < 0.001), INL thickness (P < 0.001), CRT (P < 0.001), and FAZ ratio (P = 0.003). Preoperative aniseikonia was significantly correlated with preoperative INL and OPL-ONL thicknesses (P < 0.001 and P = 0.020, respectively) and CRT (P = 0.003). Multiple regression analysis revealed that preoperative aniseikonia was significantly associated with preoperative MDRF, INL, and OPL-ONL thicknesses (P = 0.029, 0.006, and 0.006, respectively). Twenty-nine eyes underwent membrane peeling, resolving all retinal folds 6 months postoperatively. A significant correlation was observed between preoperative MDRF and postoperative aniseikonia (P = 0.011). Our findings suggest that retinal traction by ERM is significantly associated with aniseikonia both pre- and postoperatively, alongside other OCT parameters.

Application of optical coherence tomography in neurodegenerative diseases: a focus on Parkinson’s, Alzheimer’s, and Schizophrenia in Bulgarian patients

This study investigates retinal alterations in patients with Alzheimer’s disease (AD), Parkinson’s disease (PD), and schizophrenia (SZ) using Optical Coherence Tomography (OCT). The primary objective was to determine whether these neurodegenerative diseases manifest in measurable retinal changes and to assess the impact of disease duration, medication use, and cognitive decline on these alterations. A cross-sectional observational design was employed, including 132 patients and age- and gender-matched controls. OCT imaging was performed using the Topcon 3D OCT-1 Maestro 2, focusing on the retinal nerve fibre layer (RNFL), ganglion cell complex (GCC), and macular thickness. Significant retinal thinning was observed in the AD and PD groups, correlating with disease severity and cognitive decline, and was more pronounced with longer disease duration. In contrast, no significant retinal changes were identified in the SZ group. The study also explored the effects of different drug classes, revealing correlations between specific medications and retinal parameters, particularly in the AD and PD cohorts. To our knowledge, this is the first study of its kind conducted with Bulgarian patients. These findings suggest that OCT may serve as a non-invasive biomarker for neurodegeneration in AD and PD, though its utility in SZ remains limited. Future research should aim to standardize OCT protocols, further investigate the potential of retinal imaging in tracking neurodegenerative disease progression, and explore its integration with other neuroimaging techniques for a more comprehensive diagnostic and monitoring approach.

Microvascular and Structural Characterization of Birdshot Chorioretinitis in Active and Inactive Phases.

Objective: This study aimed to examine microvascular changes and identify predictors of short-term quiescence in active birdshot chorioretinitis (BSCR). Methods: An observational, prospective, 12-month follow-up cohort study was conducted. BSCR eyes were clinically assessed at baseline, categorized as active or inactive, and reevaluated at 12 months. Based on their clinical activity at both timepoints, eyes were divided into three subgroups: active-to-inactive (A-I), consistently active (A-A), and consistently inactive (I-I). Structural OCT, OCT-angiography (OCT-A), and ultra-widefield imaging were utilized. Exam data from fundus and nasal subfields were analyzed for microvascular changes and quiescence predictors. Results: Sixty eyes from 30 BSCR patients (47% women, 53% men, mean age 59.7 ± 12.3 years) were included. In the A-I group (16 eyes), vascular density and perfusion indices increased in all subfields post-quiescence, contrasting with the other groups. Perifoveal looping in the superficial capillary plexus predicted quiescence at 12 months compared with the A-A group. Conclusions: Vascular density rises after complete inflammation control in BSCR, and perifoveal capillary loops serve as potential predictors of short-term quiescence in active BSCR.

Early detection of dementia through retinal imaging and trustworthy AI

Alzheimer’s disease (AD) is a global healthcare challenge lacking a simple and affordable detection method. We propose a novel deep learning framework, Eye-AD, to detect Early-onset Alzheimer’s Disease (EOAD) and Mild Cognitive Impairment (MCI) using OCTA images of retinal microvasculature and choriocapillaris. Eye-AD employs a multilevel graph representation to analyze intra- and inter-instance relationships in retinal layers. Using 5751 OCTA images from 1671 participants in a multi-center study, our model demonstrated superior performance in EOAD (internal data: AUC = 0.9355, external data: AUC = 0.9007) and MCI detection (internal data: AUC = 0.8630, external data: AUC = 0.8037). Furthermore, we explored the associations between retinal structural biomarkers in OCTA images and EOAD/MCI, and the results align well with the conclusions drawn from our deep learning interpretability analysis. Our findings provide further evidence that retinal OCTA imaging, coupled with artificial intelligence, will serve as a rapid, noninvasive, and affordable dementia detection.

Influence of scleral thickness on photodynamic therapy outcomes in central serous chorioretinopathy.

To test the prognostic role of anterior scleral substantia propria (ASSP) thickness in predicting the 3-month response after half-dose photodynamic therapy (PDT) in central serous chorioretinopathy (CSCR) and to assess its clinical relevance of ASSP in different CSCR phenotypes. A prospective, exploratory, multi-centre cohort study conducted at IRCCS San Martino Hospital (Genoa, Italy) and Jules-Gonin Eye Hospital (Lausanne, Switzerland). Demographic and clinical data, and optical coherence tomography (OCT) were collected at baseline and 3 months after PDT. Based on OCT images, we categorized CSCR phenotypes and collected clinically relevant imaging metrics. ASSP thickness was obtained from four different measurements using anterior segment (AS) OCT. Multivariable regression models were performed to evaluate the distribution of ASSP thicknesses among different CSCR phenotypes and to test the prognostic role of ASSP thickness in discriminating between PDT responders (complete subretinal fluid reabsorption) and partial responders. The study cohort comprised 109 Caucasian patients (82 males, 75.2%) with a total of 142 eyes: 84 eyes simple (59.1%) versus 58 eyes complex (40.9%) CSCR. A linear normal model confirmed a positive association between complex CSCR and higher ASSP thickness (β = 26.1, 95% CL = 12.1/40.1, p < 0.001), with a low prevalence of ciliochoroidal effusion loculations in AS-OCT (1/142 eyes, 0.7%). ASSP thickening was positively linked to the presence of posterior cystoid retinal degeneration (PCRD; p = 0.002), indicating a potential role in the pathogenesis of severe CSCR phenotypes. In the subgroup of treated patients (61 eyes), 63.9% had a complete response after PDT. In these patients a logistic binary model highlighted a significantly higher risk of PDT non-responsiveness (OR = 9.62, 95% CL = 2.44/37.9, p = 0.001) associated with a 60-unit increase in ASSP thickness levels. By contrast, other anatomical parameters (i.e., body surface area, age, gender, axial length) showed no remarkable prognostic roles. This research highlighted the association of ASSP thickening with complex CSCR phenotype in Caucasian patients and its role in predicting PDT efficacy. These findings enhance our comprehension of the anatomical risk factors in patients affected with CSCR and potentially guide a better understanding of non-responsive cases to PDT treatment.

Deep attention for enhanced OCT image analysis in clinical retinal diagnosis

AbstractRetinal illnesses such as age-related macular degeneration (AMD) and diabetic maculopathy pose serious risks to vision in the developed world. The diagnosis and assessment of these disorders have undergone revolutionary change with the development of optical coherence tomography (OCT). This study proposes a novel method for improving clinical precision in retinal disease diagnosis by utilizing the strength of Attention-Based DenseNet, a deep learning architecture with attention processes. For model building and evaluation, a dataset of 84495 high-resolution OCT images divided into NORMAL, CNV, DME, and DRUSEN classes was used. Data augmentation techniques were employed to enhance the model's robustness. The Attention-Based DenseNet model achieved a validation accuracy of 0.9167 with a batch size of 32 and 50 training epochs. This discovery presents a promising route for more precise and speedy identification of retinal illnesses, ultimately enhancing patient care and outcomes in clinical settings by integrating cutting-edge technology with powerful neural network architectures.

Clinical safety and efficacy of elliptical thin-flap LASIK using a low-pulse energy femtosecond laser.

This study assessed the clinical safety and efficacy of elliptical thin-flap LASIK with a low-pulse energy femtosecond laser in 3 thickness subgroups (85, 90, and 100μm). A total of 80 patients who underwent bilateral LASIK surgery at Taiwan between April and September 2019 were retrospectively enrolled. Elliptical corneal flaps with wide temporal hinges and inverted-angled side cuts were created. Target flap thickness was calculated on the basis of residual stromal bed thickness and percent tissue altered. Before flap creation, an optical coherence tomography image for visualization of the precut flap position was obtained with the built-in camera. At postoperative month 1, the overall mean logMAR uncorrected distance visual acuity (UDVA) was - 0.04 ± 0.07 (20/18 Snellen), with 96% of eyes achieving UDVA of 20/20 or better. Postoperative mean manifest spherical equivalent was - 0.37 ± 0.42 D at 1 month. The visual and refractive outcomes in each of the 3 subgroups were similar. The achieved flap thicknesses were found to be highly predictable and consistent in the respective thickness subgroups. Optical coherence tomography-guided thin-flap LASIK performed with the low-pulse energy femtosecond laser was found to be safe with no intraoperative or postoperative complications.