Multimodal retinal imaging to detect and understand Alzheimer's and Parkinson's disease

Classification of diabetic retinopathy algorithm based on a novel dual-path multi-module model

Diabetic retinopathy is a chronic disease of the eye that is precipitated via diabetes. As the disease progresses, the blood vessels in the retina are issue to modifications such as dilation, leakage, and new blood vessel formation. Early detection and treatment of the lesions are vital for the prevention and reduction of imaginative and prescient loss. A new dual-path multi-module network algorithm for diabetic retinopathy classification is proposed in this paper, aiming to accurately classify the diabetic retinopathy stage to facilitate early diagnosis and intervention. To obtain the purpose of fact augmentation, the algorithm first enhances retinal lesion features using color correcting and multi-scale fusion algorithms. It then optimizes the local records via a multi-path multiplexing structure with convolutional kernels of exclusive sizes. Finally, a multi-feature fusion module is used to improve the accuracy of the diabetic retinopathy classification model. Two public datasets and a real hospital dataset are used to validate the algorithm. The accuracy is 98.9%, 99.3%, and 98.3%, respectively. The experimental results not only confirm the advancement and practicability of the algorithm in the field of automatic DR diagnosis, but also foretell its broad application prospects in clinical settings, which is expected to provide strong technical support for the early screening and treatment of diabetic retinopathy.

Analysis of the Relationship Between Parkinson's Disease and Diabetic Retinopathy Based on Bioinformatics Methods

The objective of the study was to explore the relationship and potential mechanism between Parkinson's disease (PD) and diabetic retinopathy (DR) using bioinformatics methods. We first examined the causal relationship between PD and DR by Mendelian randomization (MR) analysis. The datasets of PD and DR patients from the Gene Expression Omnibus database were used to identify differentially expressed genes (DEGs). Then, we performed the Gene Ontology, Kyoto Encyclopedia of Genes and Genomes, and immune infiltration analysis. We also constructed a protein-protein interaction network and receiver operating characteristic (ROC) curve. Finally, an online website was used for drug prediction. The MR analysis demonstrated a causal relationship between DR and PD (odds ratio [OR] = 0.86; 95% confidence interval [CI] 0.79-0.93; p = 3.24E - 04), in which DR acted as a protective factor against PD. There were 81 DEGs identified from the PD and DR datasets, of which 29 genes had protein interaction relationships, and enrichment analysis showed that these genes were mainly related to immune pathways. As indicated by immune cell infiltration analysis, the expression of immune cells between PD and the control group was significantly different. ROC curve results showed five genes had diagnostic value, and several potential chemical compounds were predicted to target the genes. Our findings demonstrate a reduced risk of PD in patients with DR. We also found that PD and DR are closely related in terms of inflammation, which provides clues for further exploring the common mechanisms and interaction of these two diseases.

An update on new-age potential biomarkers for Parkinson's disease

Parkinson's disease (PD) is a progressive neurodegenerative disorder that deals with dopaminergic deficiency in Substantia nigra pars compact (SNpc) region of the brain. Dopaminergic deficiency manifests into motor dysfunction. Alpha-synuclein protein aggregation is the source for inception of the pathology. Motor symptoms include rigidity, akinesia, tremor and gait dysfunction. Pre-motor symptoms are also seen in early stage of the disease; however, they are not distinguishable. Lack of early diagnosis in PD pathology poses a major challenge for development of disease modifying therapeutics. Substantial neuronal loss has already been occurred before the clinical manifestations appear and hence, it becomes impossible to halt the disease progression. Current diagnostics are majorly based on the clinical symptoms and thus fail to detect early progression of the disease. Thus, there is need for early diagnosis of PD, for detection of the disease at its inception. This will facilitate the effective use of therapies that halt the progression and will make remission possible. Many novel biomarkers are being developed that include blood-based biomarker, CSF biomarker. Other than that, there are non-invasive techniques that can detect biomarkers. We aim to discuss potential role of these new age biomarkers and their association with PD pathogenesis in this review.

Bridging Retinal and Cerebral Neurodegeneration: A Focus on Crosslinks between Alzheimer-Perusini's Disease and Retinal Dystrophies

In the early stages of Alzheimer-Perusini's disease (AD), individuals often experience vision-related issues such as color vision impairment, reduced contrast sensitivity, and visual acuity problems. As the disease progresses, there is a connection with glaucoma and age-related macular degeneration (AMD) leading to retinal cell death. The retina's involvement suggests a link with the hippocampus, where most AD forms start. A thinning of the retinal nerve fiber layer (RNFL) due to the loss of retinal ganglion cells (RGCs) is seen as a potential AD diagnostic marker using electroretinography (ERG) and optical coherence tomography (OCT). Amyloid beta fragments (Aβ), found in the eye's vitreous and aqueous humor, are also present in the cerebrospinal fluid (CSF) and accumulate in the retina. Aβ is known to cause tau hyperphosphorylation, leading to its buildup in various retinal layers. However, diseases like AD are now seen as mixed proteinopathies, with deposits of the prion protein (PrP) and α-synuclein found in affected brains and retinas. Glial cells, especially microglial cells, play a crucial role in these diseases, maintaining immunoproteostasis. Studies have shown similarities between retinal and brain microglia in terms of transcription factor expression and morphotypes. All these findings constitute a good start to achieving better comprehension of neurodegeneration in both the eye and the brain. New insights will be able to bring the scientific community closer to specific disease-modifying therapies.

APOE ε4 Gene Carriers Demonstrate Reduced Retinal Capillary Densities in Asymptomatic Older Adults

Early identification of Apolipoprotein E (APOE)-related microvascular pathology will help to study the microangiopathic contribution to Alzheimer's disease and provide a therapeutic target for early intervention. To evaluate the differences in retinal microvasculature parameters between APOE ε4 carriers and non-carriers, asymptomatic older adults aged ≥ 55 years underwent APOE ε4 genotype analysis, neuropsychological examination, and optical coherence tomography angiography (OCTA) imaging. One hundred sixty-three older adults were included in the data analysis. Participants were also defined as cognitively impaired (CI) and non-cognitively impaired (NCI) according to their MoCA scores and educational years. APOE ε4 carriers demonstrated reduced SVC (p = 0.023) compared to APOE ε4 non-carriers. Compared to NCI, CI participants showed reduced SVC density (p = 0.006). In the NCI group, no significant differences (p > 0.05) were observed in the microvascular densities between APOE ε4 carriers and non-carriers. In the CI group, APOE ε4 carriers displayed reduced microvascular densities compared to non-carriers (SVC, p = 0.006; DVC, p = 0.048). We showed that CI and APOE ε4 affect retinal microvasculature in older adults. Quantitative measures of the retinal microvasculature could serve as surrogates for brain microcirculation, providing an opportunity to study microvascular contributions to AD.

Age-associated changes in electroretinography measures in companion dogs

PURPOSE

To determine the association between age and retinal full-field electroretinographic (ERG) measures in companion (pet) dogs, an important translational model species for human neurologic aging.

METHODS

Healthy adult dogs with no significant ophthalmic abnormalities were included. Unilateral full-field light- and dark-adapted electroretinography was performed using a handheld device, with mydriasis and topical anaesthesia. Partial least squares effect screening analysis was performed to determine the effect of age, sex, body weight and use of anxiolytic medication on log-transformed ERG peak times and amplitudes; age and anxiolytic usage had significant effects on multiple ERG outcomes. Mixed model analysis was performed on data from dogs not receiving anxiolytic medications.

RESULTS

In dogs not receiving anxiolytics, median age was 118 months (interquartile range 72-140 months, n = 77, 44 purebred, 33 mixed breed dogs). Age was significantly associated with prolonged peak times of a-waves (dark-adapted 3 and 10 cds/m2 flash p < 0.0001) and b-waves (cone flicker p = 0.03, dark-adapted 0.01 cds/m2 flash p = 0.001). Age was also significantly associated with reduced amplitudes of a-waves (dark-adapted 3 cds/m2 flash p < 0.0001, 10 cds/m2 flash p = 0.005) and b-waves (light-adapted 3 cds/m2 flash p < 0.0001, dark-adapted 0.01 cds/m2 flash p = 0.0004, 3 cds/m2 flash p < 0.0001, 10 cds/m2 flash p = 0.007) and flicker (light-adapted 30 Hz 3 cds/m2 p = 0.0004). Within the Golden Retriever breed, these trends were matched in a cross-sectional analysis of 6 individuals that received no anxiolytic medication.

CONCLUSIONS

Aged companion dogs have slower and reduced amplitude responses in both rod- and cone-mediated ERG. Consideration of anxiolytic medication use should be made when conducting ERG studies in dogs.

Retinal organoids from human-induced pluripotent stem cells: From studying retinal dystrophies to early diagnosis of Alzheimer's and Parkinson's disease

Human-induced pluripotent stem cells (hiPSCs) have provided new methods to study neurodegenerative diseases. In addition to their wide application in neuronal disorders, hiPSCs technology can also encompass specific conditions, such as inherited retinal dystrophies. The possibility of evaluating alterations related to retinal disorders in 3D organoids increases the truthfulness of in vitro models. Moreover, both Alzheimer's (AD) and Parkinson's disease (PD) have been described as causing early retinal alterations, generating beta-amyloid protein accumulation, or affecting dopaminergic amacrine cells. This review addresses recent advances and future perspectives obtained from in vitro modeling of retinal diseases, focusing on retinitis pigmentosa (RP). Additionally, we depicted the possibility of evaluating changes related to AD and PD in retinal organoids obtained from potential patients long before the onset of the disease, constituting a valuable tool in early diagnosis. With this, we pointed out prospects in the study of retinal dystrophies and early diagnosis of AD and PD.

OCT2Former: A retinal OCT-angiography vessel segmentation transformer

BACKGROUND AND OBJECTIVE

Retinal vessel segmentation plays an important role in the automatic retinal disease screening and diagnosis. How to segment thin vessels and maintain the connectivity of vessels are the key challenges of the retinal vessel segmentation task. Optical coherence tomography angiography (OCTA) is a noninvasive imaging technique that can reveal high-resolution retinal vessels. Aiming at make full use of its characteristic of high resolution, a new end-to-end transformer based network named as OCT²Former (OCT-a Transformer) is proposed to segment retinal vessel accurately in OCTA images.

METHODS

The proposed OCT²Former is based on encoder-decoder structure, which mainly includes dynamic transformer encoder and lightweight decoder. Dynamic transformer encoder consists of dynamic token aggregation transformer and auxiliary convolution branch, in which the multi-head dynamic token aggregation attention based dynamic token aggregation transformer is designed to capture the global retinal vessel context information from the first layer throughout the network and the auxiliary convolution branch is proposed to compensate for the lack of inductive bias of the transformer and assist in the efficient feature extraction. A convolution based lightweight decoder is proposed to decode features efficiently and reduce the complexity of the proposed OCT²Former.

RESULTS

The proposed OCT²Former is validated on three publicly available datasets i.e. OCTA-SS, ROSE-1, OCTA-500 (subset OCTA-6M and OCTA-3M). The Jaccard indexes of the proposed OCT²Former on these datasets are 0.8344, 0.7855, 0.8099 and 0.8513, respectively, outperforming the best convolution based network 1.43, 1.32, 0.75 and 1.46%, respectively.

CONCLUSION

The experimental results have demonstrated that the proposed OCT²Former can achieve competitive performance on retinal OCTA vessel segmentation tasks.

Label-free hyperspectral imaging and deep-learning prediction of retinal amyloid β-protein and phosphorylated tau

Alzheimer's disease (AD) is a major risk for the aging population. The pathological hallmarks of AD-an abnormal deposition of amyloid β-protein (Aβ) and phosphorylated tau (pTau)-have been demonstrated in the retinas of AD patients, including in prodromal patients with mild cognitive impairment (MCI). Aβ pathology, especially the accumulation of the amyloidogenic 42-residue long alloform (Aβ42), is considered an early and specific sign of AD, and together with tauopathy, confirms AD diagnosis. To visualize retinal Aβ and pTau, state-of-the-art methods use fluorescence. However, administering contrast agents complicates the imaging procedure. To address this problem from fundamentals, ex-vivo studies were performed to develop a label-free hyperspectral imaging method to detect the spectral signatures of Aβ42 and pS396-Tau, and predicted their abundance in retinal cross-sections. For the first time, we reported the spectral signature of pTau and demonstrated an accurate prediction of Aβ and pTau distribution powered by deep learning. We expect our finding will lay the groundwork for label-free detection of AD.

Retinal Ganglion Cells: Global Number, Density and Vulnerability to Glaucomatous Injury in Common Laboratory Mice

How many RBPMS+ retinal ganglion cells (RGCs) does a standard C57BL/6 laboratory mouse have on average and is this number substrain- or sex-dependent? Do RGCs of (European) C57BL/6J and -N mice show a different intrinsic vulnerability upon glaucomatous injury? Global RGC numbers and densities of common laboratory mice were previously determined via axon counts, retrograde tracing or BRN3A immunohistochemistry. Here, we report the global RGC number and density by exploiting the freely available tool RGCode to automatically count RGC numbers and densities on entire retinal wholemounts immunostained for the pan-RGC marker RBPMS. The intrinsic vulnerability of RGCs from different substrains to glaucomatous injury was evaluated upon introduction of the microbead occlusion model, followed by RBPMS counts, retrograde tracing and electroretinography five weeks post-injury. We demonstrate that the global RGC number and density varies between substrains, yet is not sex-dependent. C57BL/6J mice have on average 46K ± 2K RBPMS+ RGCs per retina, representing a global RGC density of 3268 ± 177 RGCs/mm2. C57BL/6N mice, on the other hand, have on average less RBPMS+ RGCs (41K ± 3K RGCs) and a lower density (3018 ± 189 RGCs/mm2). The vulnerability of the RGC population of the two C57BL/6 substrains to glaucomatous injury did, however, not differ in any of the interrogated parameters.

Predicting Systemic Health Features from Retinal Fundus Images Using Transfer-Learning-Based Artificial Intelligence Models

While color fundus photos are used in routine clinical practice to diagnose ophthalmic conditions, evidence suggests that ocular imaging contains valuable information regarding the systemic health features of patients. These features can be identified through computer vision techniques including deep learning (DL) artificial intelligence (AI) models. We aim to construct a DL model that can predict systemic features from fundus images and to determine the optimal method of model construction for this task. Data were collected from a cohort of patients undergoing diabetic retinopathy screening between March 2020 and March 2021. Two models were created for each of 12 systemic health features based on the DenseNet201 architecture: one utilizing transfer learning with images from ImageNet and another from 35,126 fundus images. Here, 1277 fundus images were used to train the AI models. Area under the receiver operating characteristics curve (AUROC) scores were used to compare the model performance. Models utilizing the ImageNet transfer learning data were superior to those using retinal images for transfer learning (mean AUROC 0.78 vs. 0.65, p-value < 0.001). Models using ImageNet pretraining were able to predict systemic features including ethnicity (AUROC 0.93), age > 70 (AUROC 0.90), gender (AUROC 0.85), ACE inhibitor (AUROC 0.82), and ARB medication use (AUROC 0.78). We conclude that fundus images contain valuable information about the systemic characteristics of a patient. To optimize DL model performance, we recommend that even domain specific models consider using transfer learning from more generalized image sets to improve accuracy.