Alzheimer's disease pathophysiology in the Retina

Progress in Retinal and Eye Research, Journal Year: 2024, Volume and Issue: 101, P. 101273 - 101273

Published: May 15, 2024

The retina is an emerging CNS target for potential noninvasive diagnosis and tracking of Alzheimer's disease (AD). Studies have identified the pathological hallmarks AD, including amyloid β-protein (Aβ) deposits abnormal tau protein isoforms, in retinas AD patients animal models. Moreover, structural functional vascular abnormalities such as reduced blood flow, Aβ deposition, blood-retinal barrier damage, along with inflammation neurodegeneration, been described mild cognitive impairment dementia. Histological, biochemical, clinical studies demonstrated that nature severity pathologies brain correspond. Proteomics analysis revealed a similar pattern dysregulated proteins biological pathways patients, enhanced inflammatory neurodegenerative processes, impaired oxidative-phosphorylation, mitochondrial dysfunction. Notably, investigational imaging technologies can now detect AD-specific deposits, well vasculopathy neurodegeneration living suggesting alterations at different stages links to pathology. Current exploratory ophthalmic modalities, optical coherence tomography (OCT), OCT-angiography, confocal scanning laser ophthalmoscopy, hyperspectral imaging, may offer promise assessment AD. However, further research needed deepen our understanding AD's impact on its progression. To advance this field, future require replication larger diverse cohorts confirmed biomarkers standardized retinal techniques. This will validate aiding early screening monitoring.

Language: Английский

---

Clearance of interstitial fluid (ISF) and CSF (CLIC) group‐part of Vascular Professional Interest Area (PIA), updates in 2022‐2023. Cerebrovascular disease and the failure of elimination of Amyloid‐β from the brain and retina with age and Alzheimer's disease: Opportunities for therapy

Alzheimer s & Dementia, Journal Year: 2023, Volume and Issue: 20(2), P. 1421 - 1435

Published: Oct. 28, 2023

Abstract This editorial summarizes advances from the Clearance of Interstitial Fluid and Cerebrospinal (CLIC) group, within Vascular Professional Interest Area (PIA) Alzheimer's Association International Society to Advance Research Treatment (ISTAART). The overarching objectives CLIC group are to: (1) understand age‐related physiology changes that underlie impaired clearance interstitial fluid (ISF) cerebrospinal (CSF) (CLIC); (2) cellular molecular mechanisms underlying intramural periarterial drainage (IPAD) in brain; (3) establish novel diagnostic tests for disease (AD), cerebral amyloid angiopathy (CAA), retinal vasculopathy, amyloid‐related imaging abnormalities (ARIA) spontaneous iatrogenic CAA‐related inflammation (CAA‐ri), vasomotion; (4) therapies facilitate IPAD eliminate β (Aβ) aging brain retina, prevent or reduce AD CAA pathology ARIA side events associated with immunotherapy.

Language: Английский

---

Wavelet scattering transform application in classification of retinal abnormalities using OCT images

Scientific Reports, Journal Year: 2023, Volume and Issue: 13(1)

Published: Nov. 3, 2023

To assist ophthalmologists in diagnosing retinal abnormalities, Computer Aided Diagnosis has played a significant role. In this paper, particular Convolutional Neural Network based on Wavelet Scattering Transform (WST) is used to detect one four abnormalities from Optical Coherence Tomography (OCT) images. Predefined wavelet filters network decrease the computation complexity and processing time compared deep learning methods. We use two layers of WST obtain direct efficient model. generates sparse representation images which translation-invariant stable concerning local deformations. Next, Principal Component Analysis classifies extracted features. evaluate model using publicly available datasets have comprehensive comparison with literature. The accuracies classifying OCT OCTID dataset into five classes were [Formula: see text] text], respectively. achieved an accuracy detecting Diabetic Macular Edema Normal ones TOPCON device-based dataset. Heidelberg Duke contain DME, Age-related Degeneration, classes, we A our results state-of-the-art models shows that outperforms these for some assessments or achieves nearly best reported so far while having much smaller computational complexity.

Language: Английский

---

Early detection of dementia through retinal imaging and trustworthy AI

npj Digital Medicine, Journal Year: 2024, Volume and Issue: 7(1)

Published: Oct. 20, 2024

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

Language: Английский

---

Beyond the eye: A relational model for early dementia detection using retinal OCTA images

Medical Image Analysis, Journal Year: 2025, Volume and Issue: 102, P. 103513 - 103513

Published: Feb. 26, 2025

Language: Английский

---

Retinal microvascular density and inner thickness in Alzheimer’s disease and mild cognitive impairment

Frontiers in Aging Neuroscience, Journal Year: 2025, Volume and Issue: 17

Published: Feb. 28, 2025

Background Alzheimer’s disease (AD) is a major healthcare challenge, with existing diagnostics being costly/infeasible. This study explores retinal biomarkers from optical coherence tomography (OCT) and OCT angiography (OCTA) as cost-effective non-invasive solution to differentiate AD, mild cognitive impairment (MCI), healthy controls (HCs). Methods Participants the CALLIOPE Research Program were classified “Dem” (AD early AD), “MCI,” “HCs” using neuropsychological tests clinical diagnosis by neurologist. OCT/OCTA examinations conducted RTVue XR 100 Avanti SD-OCT system (VISIONIX), parameters extracted. Statistical analysis included normality homogeneity of variance (HOV) select ANOVA methods. Post-hoc analyses utilized Mann–Whitney U , Dunnett, or Tukey-HSD based on parameters’ HOV. Correlations age assessed via Pearson Spearman tests. A generalized linear model (GLM) Tweedie regression modeled relationship between MMSE scores, correcting for age. Another ordinal logistic GLM (OL-GLM) against classes, adjusting multiple confounders. Results We analyzed 357 participants: 44 Dem, 139 MCI, 174 HCs. Significant microvascular density (VD) reductions around fovea linked MCI Dem compared Age-related associated thickness HCs’ old Our OL-GLM demonstrated significant thickness/volume in Inner_Retina Full_Retina layers. Foveal avascular zone (FAZ) area perimeter initially not correlated decline; however, significantly FAZ enlargement groups. average inferior peripapillary RNFL thinning Conclusion first examine VD changes G grid sections among found association various decline. Most macular did correlate decline initially; our succeeded, highlighting importance confounders’ corrections. excluded individual layer due limitations; literature suggests their value. confirmed biomarkers’ efficacy uncovered novel decline, requiring further validation.

Language: Английский

---

Relationships between quantitative retinal microvascular characteristics and cognitive function based on automated artificial intelligence measurements

Frontiers in Cell and Developmental Biology, Journal Year: 2023, Volume and Issue: 11

Published: June 21, 2023

Introduction: The purpose of this study is to assess the relationship between retinal vascular characteristics and cognitive function using artificial intelligence techniques obtain fully automated quantitative measurements morphological parameters. Methods: A deep learning-based semantic segmentation network ResNet101-UNet was used construct a model for measurement parameters on fundus photographs. Retinal photographs centered optic disc 3107 participants (aged 50-93 years) from Beijing Eye Study 2011, population-based cross-sectional study, were analyzed. main included branching angle, fractal dimension, diameter, tortuosity, density. Cognitive assessed Mini-Mental State Examination (MMSE). Results: results showed that mean MMSE score 26.34 ± 3.64 (median: 27; range: 2-30). Among participants, 414 (13.3%) classified as having impairment (MMSE < 24), 296 (9.5%) mild (MMSE: 19-23), 98 (3.2%) moderate 10-18), 20 (0.6%) severe 10). Compared with normal group, venular average diameter significantly larger (p = 0.013), dimension density smaller (both p 0.001) in group. arteriole-to-venular ratio 0.003) 0.033) decreased group compared In multivariate analysis, better cognition (i.e., higher score) associated (b 0.134, 0.043) 0.152, 0.023) after adjustment age, best corrected visual acuity (BCVA) (logMAR) education level. Discussion: conclusion, our findings derived an intelligence-based parameter method several correlated impairment. decrease may serve candidate biomarkers early identification observed reduction occurs late stages

Language: Английский

---

Deep Learning Models for the Screening of Cognitive Impairment Using Multimodal Fundus Images

Ophthalmology Retina, Journal Year: 2024, Volume and Issue: 8(7), P. 666 - 677

Published: Jan. 26, 2024

We aimed to develop a deep learning system capable of identifying subjects with cognitive impairment quickly and easily based on multimodal ocular images. Cross-sectional study Participants Beijing Eye Study 2011 patients attending Tongren Center Hospital Physical Examination Center. trained validated algorithm assess using retrospectively collected data from the 2011. Cognitive was defined as Mini–Mental State (MMSE) score <24. Based fundus photographs optical coherence tomography (OCT) images, we developed five models following sets images: macula-centered photographs, optic disc-centered both fields, fields OCT (multi-modal). The performance evaluated compared in an external validation dataset, which Area under curve (AUC). A total 9,424 retinal 4,712 images were used model. each center included 1,180 590 Model comparison revealed that multi-modal performed best, achieving AUC 0.820 internal set, 0.786 set 1 0.784 2. multi-model different sexes age groups; there no significant differences. heatmap analysis showed signals around disc retina choroid macular regions by identify participants impairment. Fundus can provide valuable information function. Multi-modal richer single-mode models. Deep algorithms may be screening This technique has potential value for broader implementation community-based or clinic settings.

Language: Английский

---

A Novel Single-Sample Retinal Vessel Segmentation Method Based on Grey Relational Analysis

Sensors, Journal Year: 2024, Volume and Issue: 24(13), P. 4326 - 4326

Published: July 3, 2024

Accurate segmentation of retinal vessels is great significance for computer-aided diagnosis and treatment many diseases. Due to the limited number vessel samples scarcity labeled samples, since grey theory excels in handling problems "few data, poor information", this paper proposes a novel relational-based method segmentation. Firstly, noise-adaptive discrimination filtering algorithm based on relational analysis (NADF-GRA) designed enhance image. Secondly, threshold model (TS-GRA) segment enhanced Finally, post-processing stage involving hole filling removal isolated pixels applied obtain final output. The performance proposed evaluated using multiple different measurement metrics publicly available digital DRIVE, STARE HRF datasets. Experimental showed that average accuracy specificity DRIVE dataset were 96.03% 98.51%. mean 95.46% 97.85%. Precision, F1-score, Jaccard index all demonstrated high-performance levels. superior current mainstream methods.

Language: Английский

---

Dual-stream disentangled model for microvascular extraction in five datasets from multiple OCTA instruments

Frontiers in Medicine, Journal Year: 2025, Volume and Issue: 12

Published: Jan. 29, 2025

Introduction Optical Coherence Tomography Angiography (OCTA) is a cutting-edge imaging technique that captures retinal capillaries at micrometer resolution using optical instrument. Accurate segmentation of vasculature essential for eye related diseases measurement and diagnosis. However, noise artifacts from different instruments can interfere with segmentation, most existing deep learning models struggle segmenting small vessels capturing low-dimensional structural information. These challenges typically results in less precise performance. Methods Therefore, we propose novel robust Dual-stream Disentangled Network (D2Net) OCTA microvascular segmentation. Specifically, the D2Net includes dual-stream encoder separately learns image latent vascular features. By introducing structure as prior constraint constructing auxiliary information, network achieves disentangled representation learning, effectively minimizing interference artifacts. The introduced neighborhood energy Distance Correlation Energy (DCE) module, which helps to better perceive information continuous vessels. Results discussion To precisely evaluate our method on vessels, delicately establish labels by performing comprehensive detailed annotations FOCA dataset, data collected instruments, evaluated proposed mitigates microvasculature region recognition caused more refined In addition, validated performance four datasets (OCTA-500, ROSE-O, ROSE-Z, ROSE-H) acquired demonstrating its robustness generalization capabilities vessel compared other state-of-the-art methods.

Language: Английский

---