Foundation models in ophthalmology.

Foundation models represent a paradigm shift in artificial intelligence (AI), evolving from narrow models designed for specific tasks to versatile, generalisable models adaptable to a myriad of diverse applications. Ophthalmology as a specialty has the potential to act as an exemplar for other medical specialties, offering a blueprint for integrating foundation models broadly into clinical practice. This review hopes to serve as a roadmap for eyecare professionals seeking to better understand foundation models, while equipping readers with the tools to explore the use of foundation models in their own research and practice. We begin by outlining the key concepts and technological advances which have enabled the development of these models, providing an overview of novel training approaches and modern AI architectures. Next, we summarise existing literature on the topic of foundation models in ophthalmology, encompassing progress in vision foundation models, large language models and large multimodal models. Finally, we outline major challenges relating to privacy, bias and clinical validation, and propose key steps forward to maximise the benefit of this powerful technology.

Expectation maximisation pseudo labels.

In this paper, we study pseudo-labelling. Pseudo-labelling employs raw inferences on unlabelled data as pseudo-labels for self-training. We elucidate the empirical successes of pseudo-labelling by establishing a link between this technique and the Expectation Maximisation algorithm. Through this, we realise that the original pseudo-labelling serves as an empirical estimation of its more comprehensive underlying formulation. Following this insight, we present a full generalisation of pseudo-labels under Bayes' theorem, termed Bayesian Pseudo Labels. Subsequently, we introduce a variational approach to generate these Bayesian Pseudo Labels, involving the learning of a threshold to automatically select high-quality pseudo labels. In the remainder of the paper, we showcase the applications of pseudo-labelling and its generalised form, Bayesian Pseudo-Labelling, in the semi-supervised segmentation of medical images. Specifically, we focus on: (1) 3D binary segmentation of lung vessels from CT volumes; (2) 2D multi-class segmentation of brain tumours from MRI volumes; (3) 3D binary segmentation of whole brain tumours from MRI volumes; and (4) 3D binary segmentation of prostate from MRI volumes. We further demonstrate that pseudo-labels can enhance the robustness of the learned representations. The code is released in the following GitHub repository: https://github.com/moucheng2017/EMSSL.

Cross-modality Labeling Enables Noninvasive Capillary Quantification as a Sensitive Biomarker for Assessing Cardiovascular Risk.

Purpose: We aim to use fundus fluorescein angiography (FFA) to label the capillaries on color fundus (CF) photographs and train a deep learning model to quantify retinal capillaries noninvasively from CF and apply it to cardiovascular disease (CVD) risk assessment. Design: Cross-sectional and longitudinal study. Participants: A total of 90732 pairs of CF-FFA images from 3893 participants for segmentation model development, and 49229 participants in the UK Biobank for association analysis. Methods: We matched the vessels extracted from FFA and CF, and used vessels from FFA as labels to train a deep learning model (RMHAS-FA) to segment retinal capillaries using CF. We tested the model's accuracy on a manually labeled internal test set (FundusCapi). For external validation, we tested the segmentation model on 7 vessel segmentation datasets, and investigated the clinical value of the segmented vessels in predicting CVD events in the UK Biobank. Main Outcome Measures: Area under the receiver operating characteristic curve (AUC), accuracy, sensitivity, and specificity for segmentation. Hazard ratio (HR; 95% confidence interval [CI]) for Cox regression analysis. Results: On the FundusCapi dataset, the segmentation performance was AUC = 0.95, accuracy = 0.94, sensitivity = 0.90, and specificity = 0.93. Smaller vessel skeleton density had a stronger correlation with CVD risk factors and incidence (P < 0.01). Reduced density of small vessel skeletons was strongly associated with an increased risk of CVD incidence and mortality for women (HR [95% CI] = 0.91 [0.84-0.98] and 0.68 [0.54-0.86], respectively). Conclusions: Using paired CF-FFA images, we automated the laborious manual labeling process and enabled noninvasive capillary quantification from CF, supporting its potential as a sensitive screening method for identifying individuals at high risk of future CVD events. Financial Disclosures: Proprietary or commercial disclosure may be found in the Footnotes and Disclosures at the end of this article.

CF-Loss: Clinically-relevant feature optimised loss function for retinal multi-class vessel segmentation and vascular feature measurement.

Characterising clinically-relevant vascular features, such as vessel density and fractal dimension, can benefit biomarker discovery and disease diagnosis for both ophthalmic and systemic diseases. In this work, we explicitly encode vascular features into an end-to-end loss function for multi-class vessel segmentation, categorising pixels into artery, vein, uncertain pixels, and background. This clinically-relevant feature optimised loss function (CF-Loss) regulates networks to segment accurate multi-class vessel maps that produce precise vascular features. Our experiments first verify that CF-Loss significantly improves both multi-class vessel segmentation and vascular feature estimation, with two standard segmentation networks, on three publicly available datasets. We reveal that pixel-based segmentation performance is not always positively correlated with accuracy of vascular features, thus highlighting the importance of optimising vascular features directly via CF-Loss. Finally, we show that improved vascular features from CF-Loss, as biomarkers, can yield quantitative improvements in the prediction of ischaemic stroke, a real-world clinical downstream task. The code is available at https://github.com/rmaphoh/feature-loss.

Development and validation of a diagnostic nomogram model for predicting monoclonal gammopathy of renal significance.

In patients with kidney disease, the presence of monoclonal gammopathy necessitates the exploration of potential causal relationships. Therefore, in this study, we aimed to address this concern by developing a nomogram model for the early diagnosis of monoclonal gammopathy of renal significance (MGRS). Univariate and multivariate logistic regression analyses were employed to identify risk factors for MGRS. Verification and evaluation of the nomogram model's differentiation, calibration, and clinical value were conducted using the receiver operating characteristic (ROC) curve, calibration curve, and decision curve analysis. The study encompassed 347 patients who underwent kidney biopsy, among whom 116 patients (33.4%) were diagnosed with MGRS and 231 (66.6%) with monoclonal gammopathy of undetermined significance. Monoclonal Ig-related amyloidosis (n = 86) and membranous nephropathy (n = 86) was the most common renal pathological type in each group. Notably, older age, abnormal serum-free light chain ratio, and the absence of microscopic hematuria were identified as independent prognostic factors for MGRS. The areas under the ROC curves for the training and verification sets were 0.848 and 0.880, respectively. In conclusion, the nomogram model demonstrated high accuracy and clinical applicability for diagnosing MGRS, potentially serving as a valuable tool for noninvasive early MGRS diagnosis.

Aurora kinase B disruption suppresses pathological retinal angiogenesis by affecting cell cycle progression.

PURPOSE: The detrimental effects of pathological angiogenesis on the visual function are indisputable. Within a prominent role in chromosome segregation and tumor progression, aurora kinase B (AURKB) assumes a prominent role. However, its role in pathological retinal angiogenesis remains unclear. This study explores this latent mechanism. METHODS: To inhibit AURKB expression, we designed specific small interfering RNAs targeting AURKB and transfected them into vascular endothelial cells. Barasertib was selected as the AURKB inhibitor. The anti-angiogenic effects of both AURKB siRNA and barasertib were assessed in vitro by cell proliferation, transwell migration, and tube formation. To evaluate the angiogentic effects of AURKB in vivo, neonatal mice were exposed to 75% oxygen followed by normoxic repositioning to establish an oxygen-induced retinopathy (OIR) model. Subsequently, phosphate-buffered saline and barasertib were administered into OIR mice via intravitreal injection. The effects of AURKB on cell cycle proteins were determined by western blot analysis. RESULTS: We found that AURKB was overexpressed during pathological angiogenesis. AURKB siRNA and barasertib significantly inhibited endothelial cell proliferation, migration, and tube formation in vitro. Furthermore, AURKB inhibition attenuated retinal angiogenesis in the OIR model. A possible mechanism is the disruption of cell cycle by AURKB inhibition. CONCLUSION: In conclusion, AURKB significantly influenced pathological retinal angiogenesis, thereby presenting a promising therapeutic target in ocular neovascular diseases.

Precipitation Kinetics of Water-Cooled Copper Mold Al-Mg-Si(-Mn, Zr) Alloy during Aging.

The aging precipitation behavior of 6061 aluminum alloy that underwent iron casting and water-cooled copper casting and 6061 aluminum with Mn and Zr elements added was studied. Firstly, the hardness curves, tensile properties, and fracture morphology of four aging alloys-6061 (iron mold casting), 6061 (water-cooled copper mold casting), 6061-0.15Mn-0.05Zr (iron mold casting), and 6061-0.15Mn-0.05Zr (water-cooled copper mold casting)-were studied. The results of the aging hardness curve show that the aging precipitated phase of the 6061 alloy cast with a water-cooled copper mold is dispersed. The addition of Mn increases the amount of coarse inclusion α-(AlMnFeSi) in the alloy, resulting in a decrease in the age hardening property. The addition of Zr is related to the nucleation and growth of the G.P. region in the early aging period, mainly changing the formation rate and quantity of the G.P. region, leading to the advancement of peak aging and an increase in hardness. After the G.P. region gradually transforms into the ß phase, the hardness of the alloy increases with the increase in the volume fraction of the ß phase. When the ß″ phase is coarsened to the point where the fault line can be bypassed, the transitional metastable ß' phase begins to precipitate, and the coherent distortion around it weakens, indicating over-aging. Finally, the equilibrium phase Mg2Si is formed. The results of the tensile tests indicate that the tensile strength and yield strength of the 6061-0.15Mn-0.05Zr alloy produced by water-cooled copper casting after aging are 356 Mpa and 230 Mpa, respectively. These values are 80 MPa and 75 MPa higher, respectively, than those of the 6061 aluminum alloy produced via iron casting. However, the elongation is by 5%. The fracture morphology of the tensile sample of the aging alloy shows that dislocation slip in the alloy results in dislocation plugging, stress concentration, and the initiation of crack cleavage on the surface. The fracture of the water-cooled copper mold-casting alloy is a ductile fracture of the microporous aggregation type, and the macroscopic fracture exhibits an obvious "neck shrinkage" phenomenon. The fracture analysis is consistent with the mechanical properties. The DSC curve shows that there is no enrichment process of solute atoms during the heating process, and the aging precipitation process after homogenization is as follows: G.P. zone → ß″ phase → ß' phase. The aging precipitation process of the water-cooled copper casting alloy after homogenization treatment is as follows: ß″ phase → ß' phase (no precipitation in the G.P. zone was observed). The results of the differential scanning calorimetry (DSC) analysis show that the main strengthening phase in the experimental alloy system is the ß″ phase. The activation energies for the ß″ phase precipitation were calculated and found to be 147 KJ/mol, 217 KJ/mol, 185 KJ/mol, and 235 KJ/mol, respectively. Additionally, a kinetic equation for the ß″ phase precipitation during alloy aging was fitted.

Diagnostic value of α1-MG and URBP in early diabetic renal impairment.

Aims/Introduction: Diabetic kidney disease (DKD) is defined as diabetes with impaired renal function, elevated urinary albumin excretion, or both. DKD is one of the most common microvascular complications of diabetes and plays an important role in the cause of end-stage renal disease (ESRD). About 5% of people with type 2 diabetes (T2DM) already have kidney damage at the time they are diagnosed, but other triggers of renal insufficiency, such as obesity, hyperlipidemia, glomerular atherosclerosis are often present, making it difficult to define "diabetic kidney disease" or "diabetic nephropathy" precisely in epidemiology or clinical practice. Therefore, the aim of this study is to identify diabetic patients with CKD at an early stage, and evaluate the value of tubular injury markers including α1-microglobulin (α1-MG), ß2-microglobulin (ß2-MG), N-acetyl-beta-D-glucosaminidase (NAG) and Urinary retinol binding protein (URBP) in the development of diabetes to DKD. Materials and methods: We recruited a total of 182 hospitalized patients with T2DM in the First Affiliated Hospital of Zhengzhou University from February 2018 to April 2023. We collected basic clinical characteristics and laboratory biochemical parameters of the patients. Based on their levels of urinary albumin creatinine ratio (UACR) and glomerular filtration rate (GFR), patients were divided into DM group (UACR≤30 mg/g and eGFR≥90 mL/min/1.73 m2, n = 63) and DKD group (UACR>30 mg/g or eGFR<90 mL/min/1.73 m2, n = 119) excluding other causes of chronic kidney disease. We further developed diagnostic models to improve the ability to predict the risk of developing DKD by screening potential risk factors using univariate and multivariate logistic regression analysis. Calibration plots and curve analysis were used to validate the model and clinical usefulness. Next, we screened patients with relatively normal estimated glomerular filtration rate (eGFR) (≥90 mL/min/1.73 m2) to investigate whether tubular injury markers could accurately predict the risk of DKD in patients with normal renal function. We defined the rate of GFR decline as a prognostic indicator of renal function in patients and collected the information of the re-hospitalized DKD patients to determine whether the relevant indicators had an impact on the renal prognosis. Results: The patients with DKD had higher levels of tubular injury markers than patients with DM. URBP, α1-MG, eGFR were statistically different in both univariate and multivariate logistic regression analyses and displayed great predictive power after modeling with an area under curve of 0.987. The calibration curve showed medium agreement. Decision curve showed it would add more net benefits for clinical decision. After adjusting eGFR and serum creatinine (Scr), URBP was demonstrated to be associated with early renal function impairment. Conclusion: Tubular injury markers play an important role in early diabetic renal function impairment.

Automatically transferring supervised targets method for segmenting lung lesion regions with CT imaging.

BACKGROUND: To present an approach that autonomously identifies and selects a self-selective optimal target for the purpose of enhancing learning efficiency to segment infected regions of the lung from chest computed tomography images. We designed a semi-supervised dual-branch framework for training, where the training set consisted of limited expert-annotated data and a large amount of coarsely annotated data that was automatically segmented based on Hu values, which were used to train both strong and weak branches. In addition, we employed the Lovasz scoring method to automatically switch the supervision target in the weak branch and select the optimal target as the supervision object for training. This method can use noisy labels for rapid localization during the early stages of training, and gradually use more accurate targets for supervised training as the training progresses. This approach can utilize a large number of samples that do not require manual annotation, and with the iterations of training, the supervised targets containing noise become closer and closer to the fine-annotated data, which significantly improves the accuracy of the final model. RESULTS: The proposed dual-branch deep learning network based on semi-supervision together with cost-effective samples achieved 83.56 ± 12.10 and 82.67 ± 8.04 on our internal and external test benchmarks measured by the mean Dice similarity coefficient (DSC). Through experimental comparison, the DSC value of the proposed algorithm was improved by 13.54% and 2.02% on the internal benchmark and 13.37% and 2.13% on the external benchmark compared with U-Net without extra sample assistance and the mean-teacher frontier algorithm, respectively. CONCLUSION: The cost-effective pseudolabeled samples assisted the training of DL models and achieved much better results compared with traditional DL models with manually labeled samples only. Furthermore, our method also achieved the best performance compared with other up-to-date dual branch structures.

A foundation model for generalizable disease detection from retinal images.

Medical artificial intelligence (AI) offers great potential for recognizing signs of health conditions in retinal images and expediting the diagnosis of eye diseases and systemic disorders1. However, the development of AI models requires substantial annotation and models are usually task-specific with limited generalizability to different clinical applications2. Here, we present RETFound, a foundation model for retinal images that learns generalizable representations from unlabelled retinal images and provides a basis for label-efficient model adaptation in several applications. Specifically, RETFound is trained on 1.6 million unlabelled retinal images by means of self-supervised learning and then adapted to disease detection tasks with explicit labels. We show that adapted RETFound consistently outperforms several comparison models in the diagnosis and prognosis of sight-threatening eye diseases, as well as incident prediction of complex systemic disorders such as heart failure and myocardial infarction with fewer labelled data. RETFound provides a generalizable solution to improve model performance and alleviate the annotation workload of experts to enable broad clinical AI applications from retinal imaging.

Retinal Optical Coherence Tomography Features Associated With Incident and Prevalent Parkinson Disease.

BACKGROUND AND OBJECTIVES: Cadaveric studies have shown disease-related neurodegeneration and other morphological abnormalities in the retina of individuals with Parkinson disease (PD); however, it remains unclear whether this can be reliably detected with in vivo imaging. We investigated inner retinal anatomy, measured using optical coherence tomography (OCT), in prevalent PD and subsequently assessed the association of these markers with the development of PD using a prospective research cohort. METHODS: This cross-sectional analysis used data from 2 studies. For the detection of retinal markers in prevalent PD, we used data from AlzEye, a retrospective cohort of 154,830 patients aged 40 years and older attending secondary care ophthalmic hospitals in London, United Kingdom, between 2008 and 2018. For the evaluation of retinal markers in incident PD, we used data from UK Biobank, a prospective population-based cohort where 67,311 volunteers aged 40-69 years were recruited between 2006 and 2010 and underwent retinal imaging. Macular retinal nerve fiber layer (mRNFL), ganglion cell-inner plexiform layer (GCIPL), and inner nuclear layer (INL) thicknesses were extracted from fovea-centered OCT. Linear mixed-effects models were fitted to examine the association between prevalent PD and retinal thicknesses. Hazard ratios for the association between time to PD diagnosis and retinal thicknesses were estimated using frailty models. RESULTS: Within the AlzEye cohort, there were 700 individuals with prevalent PD and 105,770 controls (mean age 65.5 ± 13.5 years, 51.7% female). Individuals with prevalent PD had thinner GCIPL (-2.12 µm, 95% CI -3.17 to -1.07, p = 8.2 × 10-5) and INL (-0.99 µm, 95% CI -1.52 to -0.47, p = 2.1 × 10-4). The UK Biobank included 50,405 participants (mean age 56.1 ± 8.2 years, 54.7% female), of whom 53 developed PD at a mean of 2,653 ± 851 days. Thinner GCIPL (hazard ratio [HR] 0.62 per SD increase, 95% CI 0.46-0.84, p = 0.002) and thinner INL (HR 0.70, 95% CI 0.51-0.96, p = 0.026) were also associated with incident PD. DISCUSSION: Individuals with PD have reduced thickness of the INL and GCIPL of the retina. Involvement of these layers several years before clinical presentation highlight a potential role for retinal imaging for at-risk stratification of PD.

MisMatch: Calibrated Segmentation via Consistency on Differential Morphological Feature Perturbations With Limited Labels.

Semi-supervised learning (SSL) is a promising machine learning paradigm to address the ubiquitous issue of label scarcity in medical imaging. The state-of-the-art SSL methods in image classification utilise consistency regularisation to learn unlabelled predictions which are invariant to input level perturbations. However, image level perturbations violate the cluster assumption in the setting of segmentation. Moreover, existing image level perturbations are hand-crafted which could be sub-optimal. In this paper, we propose MisMatch, a semi-supervised segmentation framework based on the consistency between paired predictions which are derived from two differently learnt morphological feature perturbations. MisMatch consists of an encoder and two decoders. One decoder learns positive attention for foreground on unlabelled data thereby generating dilated features of foreground. The other decoder learns negative attention for foreground on the same unlabelled data thereby generating eroded features of foreground. We normalise the paired predictions of the decoders, along the batch dimension. A consistency regularisation is then applied between the normalised paired predictions of the decoders. We evaluate MisMatch on four different tasks. Firstly, we develop a 2D U-net based MisMatch framework and perform extensive cross-validation on a CT-based pulmonary vessel segmentation task and show that MisMatch statistically outperforms state-of-the-art semi-supervised methods. Secondly, we show that 2D MisMatch outperforms state-of-the-art methods on an MRI-based brain tumour segmentation task. We then further confirm that 3D V-net based MisMatch outperforms its 3D counterpart based on consistency regularisation with input level perturbations, on two different tasks including, left atrium segmentation from 3D CT images and whole brain tumour segmentation from 3D MRI images. Lastly, we find that the performance improvement of MisMatch over the baseline might originate from its better calibration. This also implies that our proposed AI system makes safer decisions than the previous methods.

Association Between Retinal Features From Multimodal Imaging and Schizophrenia.

Importance: The potential association of schizophrenia with distinct retinal changes is of clinical interest but has been challenging to investigate because of a lack of sufficiently large and detailed cohorts. Objective: To investigate the association between retinal biomarkers from multimodal imaging (oculomics) and schizophrenia in a large real-world population. Design, Setting, and Participants: This cross-sectional analysis used data from a retrospective cohort of 154 830 patients 40 years and older from the AlzEye study, which linked ophthalmic data with hospital admission data across England. Patients attended Moorfields Eye Hospital, a secondary care ophthalmic hospital with a principal central site, 4 district hubs, and 5 satellite clinics in and around London, United Kingdom, and had retinal imaging during the study period (January 2008 and April 2018). Data were analyzed from January 2022 to July 2022. Main Outcomes and Measures: Retinovascular and optic nerve indices were computed from color fundus photography. Macular retinal nerve fiber layer (RNFL) and ganglion cell-inner plexiform layer (mGC-IPL) thicknesses were extracted from optical coherence tomography. Linear mixed-effects models were used to examine the association between schizophrenia and retinal biomarkers. Results: A total of 485 individuals (747 eyes) with schizophrenia (mean [SD] age, 64.9 years [12.2]; 258 [53.2%] female) and 100 931 individuals (165 400 eyes) without schizophrenia (mean age, 65.9 years [13.7]; 53 253 [52.8%] female) were included after images underwent quality control and potentially confounding conditions were excluded. Individuals with schizophrenia were more likely to have hypertension (407 [83.9%] vs 49 971 [48.0%]) and diabetes (364 [75.1%] vs 28 762 [27.6%]). The schizophrenia group had thinner mGC-IPL (-4.05 µm, 95% CI, -5.40 to -2.69; P = 5.4 × 10-9), which persisted when investigating only patients without diabetes (-3.99 µm; 95% CI, -6.67 to -1.30; P = .004) or just those 55 years and younger (-2.90 µm; 95% CI, -5.55 to -0.24; P = .03). On adjusted analysis, retinal fractal dimension among vascular variables was reduced in individuals with schizophrenia (-0.14 units; 95% CI, -0.22 to -0.05; P = .001), although this was not present when excluding patients with diabetes. Conclusions and Relevance: In this study, patients with schizophrenia had measurable differences in neural and vascular integrity of the retina. Differences in retinal vasculature were mostly secondary to the higher prevalence of diabetes and hypertension in patients with schizophrenia. The role of retinal features as adjunct outcomes in patients with schizophrenia warrants further investigation.

Characterization and diagnostic value of the gut microbial composition in patients with minimal change disease.

Background: Minimal change disease (MCD) is one of the most common causes of primary nephrotic syndrome with high morbidity. This study aimed to explore the typical alterations of gut microbiota in MCD and establish a non-invasive classifier using key gut microbiome. We also aimed to evaluate the therapeutic efficiency of gut microbiota intervention in MCD through animal experiments. Methods: A total of 222 stool samples were collected from MCD patients and healthy controls at the First Affiliated Hospital of Zhengzhou University and Shandong Provincial Hospital for 16S rRNA sequencing. Optimum operational taxonomic units (OTUs) were obtained for constructing a diagnostic model. MCD rat models were established using doxorubicin hydrochloride for exploring the therapeutic efficiency of gut microbial intervention through fecal microbiota transplantation (FMT). Results: The α-diversity of gut microbiota decreased in MCD patients when compared with healthy controls. The relative abundance of bacterial species also changed significantly. We constructed a diagnostic model based on eight optimal OTUs and it achieved efficiency of 97.81% in discovery cohort. The high efficiency of diagnostic model was also validated in the patients with different disease states and cross-regional cohorts. The treatment partially recovered the gut microbial dysbiosis in patients with MCD. In animal experiments, likewise, the gut microbiota changed sharply in MCD rats. However, gut microbial interventions did not reduce urinary protein or pathological kidney damage. Conclusion: Gut Microbiota shifts sharply in both patients and rats with MCD. Typical microbial changes can be used as biomarkers for MCD diagnosis. The gut microbiota compositions in patients with MCD tended to normalize after treatment. However, the intervention of gut microbiota seems to have no therapeutic effect on MCD.

AutoMorph: Automated Retinal Vascular Morphology Quantification Via a Deep Learning Pipeline.

Purpose: To externally validate a deep learning pipeline (AutoMorph) for automated analysis of retinal vascular morphology on fundus photographs. AutoMorph has been made publicly available, facilitating widespread research in ophthalmic and systemic diseases. Methods: AutoMorph consists of four functional modules: image preprocessing, image quality grading, anatomical segmentation (including binary vessel, artery/vein, and optic disc/cup segmentation), and vascular morphology feature measurement. Image quality grading and anatomical segmentation use the most recent deep learning techniques. We employ a model ensemble strategy to achieve robust results and analyze the prediction confidence to rectify false gradable cases in image quality grading. We externally validate the performance of each module on several independent publicly available datasets. Results: The EfficientNet-b4 architecture used in the image grading module achieves performance comparable to that of the state of the art for EyePACS-Q, with an F1-score of 0.86. The confidence analysis reduces the number of images incorrectly assessed as gradable by 76%. Binary vessel segmentation achieves an F1-score of 0.73 on AV-WIDE and 0.78 on DR HAGIS. Artery/vein scores are 0.66 on IOSTAR-AV, and disc segmentation achieves 0.94 in IDRID. Vascular morphology features measured from the AutoMorph segmentation map and expert annotation show good to excellent agreement. Conclusions: AutoMorph modules perform well even when external validation data show domain differences from training data (e.g., with different imaging devices). This fully automated pipeline can thus allow detailed, efficient, and comprehensive analysis of retinal vascular morphology on color fundus photographs. Translational Relevance: By making AutoMorph publicly available and open source, we hope to facilitate ophthalmic and systemic disease research, particularly in the emerging field of oculomics.

Graphene quantum dots rescue angiogenic retinopathy via blocking STAT3/Periostin/ERK signaling.

BACKGROUND: Pathological retinal angiogenesis resulting from a variety of ocular diseases including oxygen induced retinopathy, diabetic retinopathy and ocular vein occlusion, is one of the major reasons for vision loss, yet the therapeutic option is limited. Multiple nanoparticles have been reported to alleviate angiogenic retinopathy. However, the adverse effect cannot be ignored due to the relatively large scale. Graphene quantum dots (GQDs) have shown potential in drug delivery and have been proved biocompatible. In this study, Graphene quantum dots are extensively investigated for their application in angiogenic retinopathy therapy. RESULTS: We showed that GQDs were biocompatible nanomaterials in vitro and in vivo. The nanoparticles have a dose-dependent inhibitory effect on proliferation, migration, tube formation and sprouting of human umbilical vein endothelial cells (HUVECs). Further data show that GQDs could inhibit pathological retinal neovascularization in an oxygen-induced retinopathy (OIR) model. The data of RNA sequencing suggested that periostin is involved in this process. GQDs inhibit the expression of periostin via STAT3, and further regulated cell cycle-related protein levels through ERK pathway. The signaling pathway was conformed in vivo using OIR mouse model. CONCLUSIONS: The present study indicated that GQDs could be a biocompatible anti-angiogenic nanomedicine in the treatment of pathological retinal neovascularization via disrupting periostin/ERK pathway and subsequent cell cycle.

Radiomics Based on Contrast-Enhanced MRI in Differentiation Between Fat-Poor Angiomyolipoma and Hepatocellular Carcinoma in Noncirrhotic Liver: A Multicenter Analysis.

OBJECTIVE: This study aims to develop and externally validate a contrast-enhanced magnetic resonance imaging (CE-MRI) radiomics-based model for preoperative differentiation between fat-poor angiomyolipoma (fp-AML) and hepatocellular carcinoma (HCC) in patients with noncirrhotic livers and to compare the diagnostic performance with that of two radiologists. METHODS: This retrospective study was performed with 165 patients with noncirrhotic livers from three medical centers. The dataset was divided into a training cohort (n = 99), a time-independent internal validation cohort (n = 24) from one center, and an external validation cohort (n = 42) from the remaining two centers. The volumes of interest were contoured on the arterial phase (AP) images and then registered to the venous phase (VP) and delayed phase (DP), and a total of 3,396 radiomics features were extracted from the three phases. After the joint mutual information maximization feature selection procedure, four radiomics logistic regression classifiers, including the AP model, VP model, DP model, and combined model, were built. The area under the receiver operating characteristic curve (AUC), diagnostic accuracy, sensitivity, and specificity of each radiomics model and those of two radiologists were evaluated and compared. RESULTS: The AUCs of the combined model reached 0.789 (95%CI, 0.579-0.999) in the internal validation cohort and 0.730 (95%CI, 0.563-0.896) in the external validation cohort, higher than the AP model (AUCs, 0.711 and 0.638) and significantly higher than the VP model (AUCs, 0.594 and 0.610) and the DP model (AUCs, 0.547 and 0.538). The diagnostic accuracy, sensitivity, and specificity of the combined model were 0.708, 0.625, and 0.750 in the internal validation cohort and 0.619, 0.786, and 0.536 in the external validation cohort, respectively. The AUCs for the two radiologists were 0.656 and 0.594 in the internal validation cohort and 0.643 and 0.500 in the external validation cohort. The AUCs of the combined model surpassed those of the two radiologists and were significantly higher than that of the junior one in both validation cohorts. CONCLUSIONS: The proposed radiomics model based on triple-phase CE-MRI images was proven to be useful for differentiating between fp-AML and HCC and yielded comparable or better performance than two radiologists in different centers, with different scanners and different scanning parameters.

Comparative analysis of the clinical outcomes between wavefront-guided and conventional femtosecond LASIK in myopia and myopia astigmatism.

AIM: To compare the clinical outcomes of wavefront guided femtosecond LASIK (WFG LASIK) and conventional femtosecond LASIK (NWFG LASIK) in eyes with myopia and myopia astigmatism. METHODS: This was a retrospective, nonrandomized, comparative investigation enrolling 236 eyes of 122 patients (18-50y) with low & moderate and high myopia. The WFG group including 97 eyes (50 patients) undergone WFG LASIK and the NWFG group including 139 eyes (72 patients) undergone conventional LASIK. Mean efficacy index, high order aberrations (HOAs), pupil size and the quality of visual questionnaire were evaluated 6mo postoperatively. RESULTS: There is no difference between WFG group (-0.054±0.049 in logMAR) and NWFG group (-0.040±0.056) in uncorrected distance visual acuity (UDVA) postoperatively. The myopia astigmatism is higher in WFG group than that in NWFG group (P<0.05). However, the mean efficacy index (MEI) in the WFG group (1.09±0.106) is better than that in the NWFG group (1.036±0.124; P<0.001). Increased HOAs were observed in NWFG group (0.30±0.196) than that in WFG group (0.146±0.188; P<0.001). The pupil size is larger in WFG group (5.15±0.76 mm) than that in NWFG group (4.32±0.52 mm). The patients are satisfied with the clinical surgery, yet WFG group showed better visual quality using the questionnaire survey. Meanwhile, high myopia would result in worse MEI, HOAs and visual quality than low & moderate myopia. CONCLUSION: WFG and NWFG FS-LASIK are both effective and safe procedures to correct low & moderate and high myopia, but WFG FS-LASIK gives a better postoperative MEI, aberrometric control and predictable outcome. Meanwhile, WFG FS-LASIK is better than NWFG FS-LASIK in correction of myopia astigmatism. Low & moderate myopia allow better clinical outcomes than high myopia using any surgical method.

A Stable NanoPAA-ZnO/ZnCl2 Composite with Variable 3D Structured Morphology and Sustained Superhydrophilicity.

A ZnO/ZnCl2 composite with stable 3D structural morphologies and long lasting superhydrophilicity was synthesized on the top surface of a nano porous anodic alumina (nanoPAA) substrate. The wettability of a nanoPAA-ZnO/ZnCl2 was systematically characterized and the experimental data indicated that the water contact angle (WCA) of 0° could be achieved as well as maintained over 7 days and still remained at 4.36° after 50 days, and its 3D structural morphology had no clearly observable change during this period. The mechanism for the superhydrophilicity of the composites was interpreted in terms of the inherent hydrophilicity of ZnO/ZnCl2 nanofilm, the three-dimensional structures of wrinkled nanoflakes, the nanogaps between neighbor nanoflakes, the difference of structual morphologies (i.e., size, shape, and upright posture of nanoflakes), and the measured True Volume of voids in the nanocomposite. The structural morphologies were mainly determined by the parameters such as the original concentration of precursor ZnCl2 and the pore diameter of nanoPAA substrate. The study proposes a promising superhydrophilic nanomaterial and a cost-effective synthesis method, which will play a practical role in the fields of biomedical molecular sensors and micro/nanofluidic chips.