Development of a transformer-based deep learning algorithm for diabetic peripheral neuropathy classification using corneal confocal microscopy images DOI Creative Commons
Woei‐Luen Chen,

Danling Liao,

Yuyang Deng

и другие.

Frontiers in Cell and Developmental Biology, Год журнала: 2024, Номер 12

Опубликована: Окт. 14, 2024

Background Diabetic peripheral neuropathy (DPN) is common and can go unnoticed until it firmly developed. This study aims to establish a transformer-based deep learning algorithm (DLA) classify corneal confocal microscopy (CCM) images, identifying DPN in diabetic patients. Methods Our classification model differs from traditional convolutional neural networks (CNNs) using Swin transformer network with hierarchical architecture backbone. Participants included those (DPN+, n = 57) or without (DPN−, 37) as determined by the updated Toronto consensus criteria. The CCM image dataset (consisting of 570 DPN+ 370 DPN− five images selected each participant’s left right eyes) was randomly divided into training, validation, test subsets at 7:1:2 ratio, considering individual participants. effectiveness assessed diagnostic accuracy measures, such sensitivity, specificity, accuracy, conjunction Grad-CAM visualization techniques interpret model’s decisions. Results In + group (n 12), successfully predicted all participants, while 7), one participant misclassified DPN+, an area under curve (AUC) 0.9405 (95% CI 0.8166, 1.0000). Among 120), 117 were correctly classified, among 70), 49 AUC 0.8996 0.8502, 0.9491). For single-image predictions, achieved superior relative ResNet50 (0.8761, 95% 0.8155, 0.9366), Inception_v3 (0.8802, 0.8231, 0.9374), DenseNet121 (0.8965, 0.8438, Conclusion Transformer-based outperform CNN-based rapid binary classification. DLAs have clinical screening potential.

Язык: Английский

Advances in Understanding Biomarkers and Treating Neurological Diseases -Role of the Cerebellar Dysfunction and Emerging Therapies DOI
Azhagu Madhavan Sivalingam

Ageing Research Reviews, Год журнала: 2024, Номер unknown, С. 102519 - 102519

Опубликована: Сен. 1, 2024

Язык: Английский

Процитировано

4

Association between lipoprotein(a) and diabetic peripheral neuropathy in patients with type 2 diabetes: a meta-analysis DOI Creative Commons
Sheng Li, Yiwen Yang,

Yun-Qing Zhou

и другие.

Diabetology & Metabolic Syndrome, Год журнала: 2025, Номер 17(1)

Опубликована: Март 4, 2025

Diabetic peripheral neuropathy (DPN) is a common complication of type 2 diabetes (T2D). Lipoprotein(a) [Lp(a)], known cardiovascular risk factor, has been hypothesized to influence the development DPN. This meta-analysis aimed investigate relationship between Lp(a) levels and DPN in patients with T2D. Following PRISMA 2020 guidelines, systematic search PubMed, Embase, Web Science, Wanfang, CNKI databases was performed up October 12, 2024. Observational studies assessing blood T2D without or evaluating association were included. Data synthesis utilized random-effects model calculate standardized mean differences (SMDs) odds ratios (ORs) corresponding 95% confidence intervals (CIs). Eleven 18,022 Patients had significantly higher than those (SMD: 0.10, CI: 0.02–0.19, p = 0.01; I² 43%). High associated (OR: 1.31, 1.07–1.60, 0.009; 62%). Subgroup analyses according study design, age patients, methods for measuring concentration, cutoff values high Lp(a), quality scores showed consistent results (p subgroup difference all > 0.05). A from Asian countries, but not European countries 0.001). Elevated are particularly countries.

Язык: Английский

Процитировано

0

Causal Association Between Cholesterol‐Lowering Drugs and Diabetic Microvascular Complications: A Drug‐Target Mendelian Randomization Study DOI Creative Commons
Bo Yang, Bo Yao,

Qu Zou

и другие.

Journal of Diabetes Research, Год журнала: 2025, Номер 2025(1)

Опубликована: Янв. 1, 2025

Background: It remains unclear whether cholesterol-lowering therapy can reduce the incidence of microvascular complications in patients with diabetes. We aim to explore potential causal relationship between three common types drugs and diabetic through drug-target Mendelian randomization (MR) study, laying groundwork for development new medications. Methods: In this we collected single nucleotide polymorphisms (SNPs) associated HMGCR (3-hydroxy-3-methylglutaryl-CoA reductase) inhibitors, PCSK9 (proprotein convertase subtilisin/kexin type 9) NPC1L1 (Niemann-Pick C1-Like 1) inhibitors from published genome-wide association study statistics. Subsequently, MR analyses were performed investigate effects these on low-density lipoprotein cholesterol (LDL-C) level-mediated diabetes mellitus. Coronary atherosclerosis as a positive control. Primary outcomes included nephropathy, retinopathy, neuropathy FinnGen Consortium. Results: The analysis revealed significant associations inhibition increased risks nephropathy (OR [95%confidence interval (CI)] = 1.88 [1.50, 2.36], p 5.55 × 10-8), retinopathy [95%CI] 1.86 [1.54, 2.24], 6.28 10-11), 2.63 [1.84, 3.75], 1.14 10-7) using inverse variance weighted method. have been an risk 1.30 [1.07, 1.58], 0.009) 1.40 [1.15, 1.72], 0.001); significantly 0.48 [0.28, 0.85], 0.01). coronary heart disease Conclusions: findings show that may increase complications. However, provide protection against retinopathy.

Язык: Английский

Процитировано

0

What does nerve ultrasound contribute to the evaluation of diabetic polyneuropathy over time? A prospective follow-up observational study of people with type 2 diabetes DOI
Bianka Heiling, Thomas Lehmann, N Müller

и другие.

Diabetes Research and Clinical Practice, Год журнала: 2025, Номер unknown, С. 112115 - 112115

Опубликована: Март 1, 2025

Язык: Английский

Процитировано

0

Early diagnosis and risk factors of diabetic peripheral neuropathy in type 1 diabetes: insights from current perception threshold testing DOI Creative Commons
Qian Zhao, Jialin Wang, Fangfang Liu

и другие.

Frontiers in Endocrinology, Год журнала: 2025, Номер 16

Опубликована: Март 31, 2025

This study investigates nerve fiber dysfunction in type 1 diabetes (T1D) patients and identifies risk factors for diabetic peripheral neuropathy (DPN). It evaluates the relationship between current perception threshold (CPT) tests conduction velocity (NCV), assesses CPT's diagnostic accuracy early DPN detection. enrolled 110 with T1D 26 healthy controls January 2020 December 2021, accordance predefined inclusion/exclusion criteria. CPT testing at 2000 Hz, 250 5 Hz assessed Aβ, Aδ, C function, while NCV was measured 47 patients. Subgroups were stratified by disease duration (>5 years vs ≤5 years). Multivariate logistic regression identified factors, CPT-NCV correlation analyzed using Chi-square Kappa tests. Receiver operating characteristic (ROC) curves evaluated efficacy. The incidence of 78%, no significant difference subgroups (78.3% vs. 78.0%). Neurological abnormalities significantly more common lower extremities compared to upper (67.27% 49.09%, P < 0.05). analysis revealed that a waist-to-hip ratio (WHR) greater than 0.85 an independent factor (OR = 3.01, 95% CI: 1.03-8.80, Patients >5 demonstrated higher 2000Hz abnormality rates (68.09% 46.15%, 0.05) severe neurological lesions (57.45% 35.90%, In contrast, those exhibited elevated 5Hz (30.77% 10.64%, predominantly milder (56.41% 31.91%, Statistical analyses association (P<0.001), moderate consistency further supported Cohen's Test (κ=0.515, P<0.001). ROC curve detecting 5Hz, area under (AUC) 0.723. showed unmyelinated (C) fibers detection, routine screening high-risk groups (central obesity/short duration) enables timely intervention prevent irreversible damage.

Язык: Английский

Процитировано

0

Diabetes Mellitus and Associated Vascular Disease: Pathogenesis, Complications, and Evolving Treatments DOI Creative Commons
Kazi Nazrul Islam,

Rahib K. Islam,

Ivan Nguyen

и другие.

Advances in Therapy, Год журнала: 2025, Номер unknown

Опубликована: Апрель 19, 2025

Diabetes mellitus is a metabolic disorder, characterized by elevated blood sugar levels (hyperglycemia) and insulin dysregulation. This disease associated with morbidity mortality, including significant potential vascular complications. High of hyperglycemia lead to not only reactive oxygen species but also advanced glycation end products, which are detrimental the endothelium reduce protective compounds such as nitric oxide prostacyclin. damage contributes development both macrovascular microvascular The present investigation explores pathophysiological mechanisms diabetic complications evaluates current management strategies, lifestyle modifications, pharmacological treatments, emerging therapies. review underscores importance ongoing progress in diabetes patient education optimal patient-health outcomes quality life for individuals mellitus.

Язык: Английский

Процитировано

0

Diabetic neuropathy: cutting-edge research and future directions DOI Creative Commons
Yang Yang, Bing Zhao,

Yuanzhe Wang

и другие.

Signal Transduction and Targeted Therapy, Год журнала: 2025, Номер 10(1)

Опубликована: Апрель 24, 2025

Abstract Diabetic neuropathy (DN) is a prevalent and debilitating complication of diabetes mellitus, significantly impacting patient quality life contributing to morbidity mortality. Affecting approximately 50% patients with diabetes, DN predominantly characterized by distal symmetric polyneuropathy, leading sensory loss, pain, motor dysfunction, often resulting in diabetic foot ulcers lower-limb amputations. The pathogenesis multifaceted, involving hyperglycemia, dyslipidemia, oxidative stress, mitochondrial inflammation, which collectively damage peripheral nerves. Despite extensive research, disease-modifying treatments remain elusive, current management primarily focusing on symptom control. This review explores the complex mechanisms underlying highlights recent advances diagnostic therapeutic strategies. Emerging insights into molecular cellular pathways have unveiled potential targets for intervention, including neuroprotective agents, gene stem cell therapies, innovative pharmacological approaches. Additionally, novel tools, such as corneal confocal microscopy biomarker-based tests, improved early detection intervention. Lifestyle modifications multidisciplinary care strategies can enhance outcomes. While significant progress has been made, further research required develop therapies that effectively halt or reverse disease progression, ultimately improving lives individuals DN. provides comprehensive overview understanding future directions management.

Язык: Английский

Процитировано

0

Thermography, Temperature, Pressure Force Distribution and Physical Activity in Diabetic Foot: A Systematic Review DOI Creative Commons

Maria P. Rosell-Diago,

Marta Izquierdo-Renau, Iván Julían-Rochina

и другие.

Applied Sciences, Год журнала: 2024, Номер 14(19), С. 8726 - 8726

Опубликована: Сен. 27, 2024

One of the most significant complications diabetes mellitus is diabetic sensory neuropathy. A reduction in pain perception increases risk patients developing foot ulcers and requiring amputation, which turn results impaired mobility a diminished quality life. Despite recent advances our understanding aetiology (DFUs), there remains gap clinical prevention management these ulcers. Given that inflammation represents initial stage formation skin wounds, utilisation infrared thermography for early detection inflammatory tissue, analysis plantar pressures monitoring controlled physical activity (PA) may facilitate close observation alterations are susceptible to development DFUs, thereby enabling implementation timely interventions such as personalised PA patients. It recommended an integrated approach be adopted encompasses all emerging technologies order enhance outcomes DFUs. The aim this study examine existing literature on relationship between temperature, pressure foot. To end, systematic review was conducted accordance with Preferred Reporting Items Systematic Reviews Meta-Analyses (PRISMA) format across PubMed, Scopus Web Science databases, no temporal restriction. total 14 studies were included review. use can changes pose UFDs provide intervention, personalising patient.

Язык: Английский

Процитировано

0

Development of a transformer-based deep learning algorithm for diabetic peripheral neuropathy classification using corneal confocal microscopy images DOI Creative Commons
Woei‐Luen Chen,

Danling Liao,

Yuyang Deng

и другие.

Frontiers in Cell and Developmental Biology, Год журнала: 2024, Номер 12

Опубликована: Окт. 14, 2024

Background Diabetic peripheral neuropathy (DPN) is common and can go unnoticed until it firmly developed. This study aims to establish a transformer-based deep learning algorithm (DLA) classify corneal confocal microscopy (CCM) images, identifying DPN in diabetic patients. Methods Our classification model differs from traditional convolutional neural networks (CNNs) using Swin transformer network with hierarchical architecture backbone. Participants included those (DPN+, n = 57) or without (DPN−, 37) as determined by the updated Toronto consensus criteria. The CCM image dataset (consisting of 570 DPN+ 370 DPN− five images selected each participant’s left right eyes) was randomly divided into training, validation, test subsets at 7:1:2 ratio, considering individual participants. effectiveness assessed diagnostic accuracy measures, such sensitivity, specificity, accuracy, conjunction Grad-CAM visualization techniques interpret model’s decisions. Results In + group (n 12), successfully predicted all participants, while 7), one participant misclassified DPN+, an area under curve (AUC) 0.9405 (95% CI 0.8166, 1.0000). Among 120), 117 were correctly classified, among 70), 49 AUC 0.8996 0.8502, 0.9491). For single-image predictions, achieved superior relative ResNet50 (0.8761, 95% 0.8155, 0.9366), Inception_v3 (0.8802, 0.8231, 0.9374), DenseNet121 (0.8965, 0.8438, Conclusion Transformer-based outperform CNN-based rapid binary classification. DLAs have clinical screening potential.

Язык: Английский

Процитировано

0