Search Results (492)

Oral cancer remains a critical global health challenge, with delayed diagnosis driving high morbidity and mortality. Despite progress in artificial intelligence, computer vision, and medical imaging, early detection tools that are accessible, explainable, and designed for patient engagement remain limited. This study presents a novel system that combines a patient-centred chatbot with a deep learning framework to support early diagnosis, symptom triage, and health education. The system integrates convolutional neural networks, class activation mapping, and natural language processing within a conversational interface. Five deep learning models were evaluated (CNN, DenseNet121, DenseNet169, DenseNet201, and InceptionV3) using two balanced public datasets. Model performance was assessed using accuracy, sensitivity, specificity, diagnostic odds ratio (DOR), and Cohen’s Kappa. InceptionV3 consistently outperformed the other models across these metrics, achieving the highest diagnostic accuracy (77.6%) and DOR (20.67), and was selected as the core engine of the chatbot’s diagnostic module. The deployed chatbot provides real-time image assessments and personalised conversational support via multilingual web and mobile platforms. By combining automated image interpretation with interactive guidance, the system promotes timely consultation and informed decision-making. It offers a prototype for a chatbot, which is scalable and serves as a low-cost solution for underserved populations and demonstrates strong potential for integration into digital health pathways. Importantly, the system is not intended to function as a formal screening tool or replace clinical diagnosis; rather, it provides preliminary guidance to encourage early medical consultation and informed health decisions. Full article

Integrating bioluminescent organisms as passive lighting sources in the built environment is currently a hot topic. However, there are several limitations facing the implementation and up-scaling of these naturally bioluminescent organisms in the built environment on architectural and urban scales, such as the scale, sensitivity, enclosure, and difficulty of maintenance. Moreover, there are complex technicalities and operational aspects of conventional bioreactors that host these bioluminescent agents, especially in terms of managing their recharge and effluent, not to mention their high maintenance cost. The current work offers a sustainable, stand-alone, bioluminescent urban screen system employing Aliivibrio fischeri CECT 524 bioink on 3D-printed customized scaffolds as bioreceptive panel design based on a field-diffusion pattern to host the bioluminescent bacterial bioink. The field-diffusion pattern was employed thanks to its proven efficiency in entrapment of the various microbial cultures. Three different growth media were tested for culturing Aliivibrio fischeri CECT 524, including Luria Bertani Broth (LB), the Tryptone Soy Broth (TSB), and the standard Marine Broth (MB). The results revealed that the Marine Broth (MB) media achieved the highest bioluminescent intensity and duration. The maximum light emission typically in range of ~490 nm of blue–green light captured by a conventional reflex camera (human eye vision) was observed for 10 consecutive days in complete darkness after 3–10 s, at a room temperature of 25 °C. This was visible mainly at the thin curvilinear peaks of the 3D-printed field pattern. P1 achieved the highest performance in terms of visible blue–green light, and a duration of 10 days of active bioluminescence was achieved without the need for refilling, thanks to the high number of peaks and narrow wells at <0.5 cm of its field-diffusion pattern. This study proves the efficiency of this biomimetic pattern in terms of the bioreceptivity of the bioluminescent bacterial bioink. Furthermore, the proposed 3D-printed urban screens proved their economic sustainability in terms of affordability and their minimized production processes, in addition to their easy maintenance and recharge. These results qualify these 3D-printed bioluminescent urban screens for easy and decentralized adoption and application on an architectural and urban scale. Full article

Purpose: To investigate the efficacy of a novel test for diagnosing colour vision deficiencies using reflexive eye movements measured using an unmodified tablet. Methods: This study followed a cross-sectional design, where thirty-three participants aged between 17 and 65 years were recruited. The participant group comprised 23 controls, 8 deuteranopes, and 2 protanopes. An anomaloscope was employed to determine the colour vision status of these participants. The study methodology involved using an Apple iPad Pro’s built-in eye-tracking capabilities to record eye movements in response to coloured patterns drifting on the screen. Through an automated analysis of these movements, the researchers estimated individuals’ red–green equiluminant point and their equivalent luminance contrast. Results: Estimates of the red–green equiluminant point and the equivalent luminance contrast were used to classify participants’ colour vision status with a sensitivity rate of 90.0% and a specificity rate of 91.30%. Conclusions: The novel colour vision test administered using an unmodified tablet was found to be effective in diagnosing colour vision deficiencies and has the potential to be a practical and cost-effective alternative to traditional methods. Translation Relevance: The test’s objectivity, its straightforward implementation on a standard tablet, and its minimal requirement for patient cooperation, all contribute to the wider accessibility of colour vision diagnosis. This is particularly advantageous for demographics like children who might be challenging to engage, but for whom early detection is of paramount importance. Full article

Background and Objectives: Children with autism spectrum disorder (ASD) often experience visual problems, yet their ophthalmic health remains underexplored due to testability challenges and limited-service access. This study evaluated ophthalmic screening outcomes in children with ASD and examined whether autism severity influenced ocular findings or cooperation during examinations. Materials and Methods: This cross-sectional study included 210 children with ASD (mean age 8.18 ± 4.99 years; 83.3% male). Examinations were conducted in an autism education center using non-contact methods: stereopsis (LANG I stereotest; LANG-STEREOTEST AG, Küsnacht, Switzerland), cover–uncover, and Hirschberg tests for strabismus, Spot Vision Screener (Welch Allyn Inc., Skaneateles Falls, NY, USA) for refractive errors, and Brückner test for red reflex. Autism severity was assessed with the Turkish version of the Adapted Autism Behavior Checklist (AABC). Results: Refractive errors were identified in 22.3% of participants: astigmatism in 15.2%, myopia in 5.2% (including 3 high myopia), and hyperopia in 1.9%. Strabismus was present in 11.9%, most commonly intermittent exotropia. Nearly half (49.5%) could not complete stereopsis testing, and a weak positive correlation was observed between AABC scores and the higher absolute spherical equivalent (SE) value between the two eyes (r = 0.173, p = 0.044). Children unable to complete stereopsis testing had significantly higher AABC scores (22.66 ± 9.69 vs. 13.39 ± 9.41, p < 0.001). Notably, 50 children (23.8%) had never undergone an eye examination prior to this study. Conclusions: Ophthalmic findings, particularly astigmatism and strabismus, are common in children with ASD. Greater autism severity was associated with reduced testability and modestly worse refractive error status. These findings suggest that tailored, accessible eye-care approaches and systematic vision screening may help to reduce overlooked visual problems and support more equitable care for children with ASD. Full article

Early, non-invasive detection of Parkinson’s Disease (PD) using speech analysis offers promise for scalable screening. In this work, we propose a multi-channel spectro-temporal deep-learning approach for PD detection from sentence-level speech, a clinically relevant yet underexplored modality. We extract and fuse three complementary time–frequency representations—mel spectrogram, constant-Q transform (CQT), and gammatone spectrogram—into a three-channel input analogous to an RGB image. This fused representation is evaluated across CNNs (ResNet, DenseNet, and EfficientNet) and Vision Transformer using the PC-GITA dataset, under 10-fold subject-independent cross-validation for robust assessment. Results showed that fusion consistently improves performance over single representations across architectures. EfficientNet-B2 achieves the highest accuracy (84.39% ± 5.19%) and F1-score (84.35% ± 5.52%), outperforming recent methods using handcrafted features or pretrained models (e.g., Wav2Vec2.0, HuBERT) on the same task and dataset. Performance varies with sentence type, with emotionally salient and prosodically emphasized utterances yielding higher AUC, suggesting that richer prosody enhances discriminability. Our findings indicate that multi-channel fusion enhances sensitivity to subtle speech impairments in PD by integrating complementary spectral information. Our approach implies that multi-channel fusion could enhance the detection of discriminative acoustic biomarkers, potentially offering a more robust and effective framework for speech-based PD screening, though further validation is needed before clinical application. Full article

Background/Objectives: Childhood eye disorders, including refractive errors, strabismus, and amblyopia, are prevalent yet often underdiagnosed in Saudi Arabia. Limited data on barriers to pediatric eye care hinder efforts to optimize service delivery. This study aimed to identify barriers to accessing pediatric eye care and to develop consensus-based strategies for improvement. Methods: A Delphi technique involving three iterative rounds of questionnaires was conducted with a panel of 22 eye care experts across Saudi Arabia. Consensus was defined as ≥80% agreement among participants. In total, 30 statements were developed from thematic analysis of open-ended responses and a supporting literature review. Panelists rated each statement on a five-point Likert scale, and descriptive statistics were applied. Internal consistency across rounds was assessed using Cronbach’s alpha. Results: Of the 30 proposed statements, 25 (83.3%) reached consensus, with a mean agreement score of 4.45 ± 0.59. Internal consistency was high (Cronbach’s alpha = 0.92). High-priority recommendations included implementing mandatory vision screening, integrating optometrists into primary healthcare, and establishing specialized pediatric eye care centers. Other recommendations emphasized expanding mobile clinics and increasing public awareness. Areas that did not reach consensus included referral inefficiencies, adequacy of the current workforce, and school accommodations for children with visual impairment. Conclusions: This study presents the first national consensus on pediatric eye care in Saudi Arabia and provides actionable recommendations to strengthen services. The findings offer a strategic framework to guide policy, enhance workforce development, and reduce childhood visual impairment through early detection and intervention. Full article

Background: Vision loss due to chronic eye diseases remains a significant public health challenge. Early detection through screening programs may reduce the burden of vision loss. This study aimed to assess the detection rate of eye diseases (glaucoma, AMD, and diabetic retinopathy), including those newly detected during opportunistic screening and ophthalmological consultations within the local government health policy program “Good Vision for Mazovians” in Mazovia, Poland. Material and methods: This study is a retrospective analysis of medical data from the registry of the Ophthalmology Department of the Międzylesie Specialist Hospital in Warsaw, which implemented the local government preventive program “Good Vision for Mazovians. Data from 1812 individuals (aged 18–92 years) participating in the “Good Vision for Mazovians” preventive program were analyzed. Results: Most participants were female (59.7%), aged over 60, and took medications regularly (62.7%). Excluding subjects with prior diagnosis of eye conditions, the detection rate was 38 suspected cases (3.8%) of glaucoma cases, 84 suspected cases of AMD (4.6%), and 21 suspected cases of diabetic retinopathy (1.2%). Most participants had not visited an ophthalmologist in the past two years (58.6%), reported low or average knowledge of eye health, had difficulty accessing ophthalmology services in their region (57%), and identified long waiting times for appointments as the main barrier to care (83.5%). Conclusions: Opportunistic screening for eye diseases in populations with limited access to eye care should be considered as a method for detecting common causes of irreversible visual impairment, particularly AMD. Older adults and individuals without higher education appear to face the greatest barriers to accessing ophthalmology services and may benefit the most from targeted opportunistic screening initiatives. Full article

Background/Objectives: Diabetes Mellitus (DM) remains a critical public health issue in Saudi Arabia, shaped by complex interactions among genetic, lifestyle, and sociodemographic factors. This study explores interplay of modifiable and non-modifiable determinants of DM among Saudi adults. Methods: An analytical cross-sectional study was conducted among 3411 adults aged 18 years and above in the Jazan region, southwest of Saudi Arabia, in May–June 2024. Data was collected via a structured, pretested questionnaire assessing sociodemographic, dietary patterns, physical activity, smoking habits, and family history of DM. Bivariate analysis and logistic regression were used to identify associations with self-reported diabetes. Results: Out of 3411 participants (1735 males and 1676 females), 424 (12.4%) reported DM. Diabetics were older (48 vs. 32 years), more often male, married, had lower education, had larger families, had higher BMIs, and exhibited more tobacco use (p < 0.05), and a family history of diabetes was strongly associated with diagnosis of DM (p < 0.001). Diabetics were more likely to choose low-fat meats, avoid sugary foods, and select low-fat products (p < 0.05). In multivariate analysis, predictors were age (OR = 1.07, 95% CI: 1.06–1.09), male sex (OR = 1.65, 95% CI: 1.26–2.16), family history (OR = 7.68, 95% CI: 5.67–10.57), traditional housing (OR = 1.82, 95% CI: 1.11–3.05), and whole grain intake (OR = 0.67, 95% CI: 0.52–0.85). Conclusions: DM in Saudi Arabia is driven by both inherited and behavioral risks. These findings support the urgent need for integrated, culturally tailored prevention strategies that combine early screening for individuals with higher risk. Targeted actions such as relevant lifestyle interventions can help reduce disease burden and align with Saudi Vision 2030 health priorities. Full article

For more than 60 years, newborn (or neonatal) screening has flourished through global collaboration, demonstrating that collective action is key to success. This unity proved to be especially vital during the COVID-19 pandemic, when, despite severe disruptions, NBS services were largely preserved, reflecting the high value placed on early detection and care for vulnerable newborns. Today, the International Society for Neonatal Screening (ISNS) recognises that NBS programmes face increasing challenges due to global instability. While direct assistance is not always possible, ISNS emphasises the strength of the international NBS community—scientists, clinicians, patient groups, and industry partners—who are committed to mutual support and knowledge-sharing. Building on the proud legacy inspired by pioneers like Bob Guthrie, this community is enriched by diverse voices and is unified by a shared vision: to ensure that all children with rare disorders have access to life-saving screening and care. Safeguarding and advancing this foundation is a responsibility owed to future generations. Full article

Background: Diabetic retinopathy (DR) is a leading cause of preventable vision impairment in individuals with diabetes. Early detection is essential, yet often hindered by subtle disease progression and reliance on manual expert screening. This study introduces an AI-based framework designed to achieve robust multiclass DR classification from retinal fundus images, addressing the challenges of early diagnosis and fine-grained lesion discrimination. Methods: The framework incorporates preprocessing steps such as pixel intensity normalization and geometric correction. A Hybrid Local-Global Retina Super-Resolution (HLG-RetinaSR) module is developed, combining deformable convolutional networks for local lesion enhancement with vision transformers for global contextual representation. Classification is performed using a hierarchical approach that integrates three models: a Convolutional Neural Network (CNN), DenseNet-121, and a custom multi-branch RefineNet-U architecture. Results: Experimental evaluation demonstrates that the combined HLG-RetinaSR and RefineNet-U approach consistently achieves precision, recall, F1-score, and accuracy values exceeding 99% across all DR severity levels. The system effectively emphasizes vascular abnormalities while suppressing background noise, surpassing existing state-of-the-art methods in accuracy and robustness. Conclusions: The proposed hybrid pipeline delivers a scalable, interpretable, and clinically relevant solution for DR screening. By improving diagnostic reliability and supporting early intervention, the system holds strong potential to assist ophthalmologists in reducing preventable vision loss. Full article

Background/Objectives: Keratoconus (KC) can rapidly erode vision in children with Leber congenital amaurosis (LCA), yet screening usually depends on costly corneal imaging that is often unavailable. We evaluated whether a lightweight, image-free neural network fed only routine clinical and genetic variables can detect KC in patients with AIPL1-related LCA. Methods: This retrospective, proof-of-concept pilot study analyzed chart data for 19 children with biallelic AIPL1 mutations (6 with KC) seen at five tertiary eye centers between January and December 2004. Ten baseline predictors were entered into a feed-forward neural network. Records were randomly split 60/20/20 into training, validation and test sets; 20 replicate networks were trained. The mean test accuracy, sensitivity and specificity across runs were the primary outcomes. Results: The ensemble achieved a mean test accuracy of 91.6% (SD 12.8%), sensitivity of 87.5% (SD 13.1%) and specificity of 93.5% (SD 17.0%). A total of 6 of the 20 runs made no test-set errors, and 16 achieved 100% specificity. The median training time per network was less than 1 s on a laptop CPU. Conclusions: This exploratory pilot shows that a point-of-care, image-free neural network using readily available clinical and genetic data accurately identified KC in AIPL1-LCA. External validation in larger, contemporary cohorts is warranted, but the approach could help triage scarce imaging resources and enable timely corneal–collagen cross-linking in settings where tomography is inaccessible. Full article

Background/Objectives: Early and accurate detection of glaucoma is vital for preventing irreversible vision loss, yet traditional diagnostic approaches relying solely on unimodal retinal imaging are limited by data sparsity and constrained context. Furthermore, real-world clinical data are often fragmented across institutions under strict privacy regulations, posing significant challenges for centralized machine learning methods. Methods: To address these barriers, this study proposes a novel Quality Aware Vertical Federated Learning (QAVFL) framework for decentralized multimodal glaucoma detection. The proposed system dynamically integrates clinical text, retinal fundus images, and biomedical signal data through modality-specific encoders, followed by a Fusion Attention Module (FAM) that adaptively weighs the reliability and contribution of each modality. Unlike conventional early fusion or horizontal federated learning methods, QAVFL operates in vertically partitioned environments and employs secure aggregation mechanisms incorporating homomorphic encryption and differential privacy to preserve patient confidentiality. Results: Extensive experiments conducted under heterogeneous non-IID settings demonstrate that QAVFL achieves an accuracy of 98.6%, a recall of 98.6%, an F1-score of 97.0%, and an AUC of 0.992, outperforming unimodal and early fusion baselines with statistically significant improvements (p < 0.01). Conclusions: The findings validate the effectiveness of dynamic multimodal fusion under privacy-preserving decentralized learning and highlight the scalability and clinical applicability of QAVFL for robust glaucoma screening across fragmented healthcare environments. Full article

Background/Objectives: The inner blood–retinal barrier (iBRB) is a specialized neurovascular interface essential for retinal homeostasis and visual function and is compromised in several vision-threating conditions. Therefore, the ability to model iBRB function and dysfunction in a controlled, reproducible and scalable manner is crucial for pharmaceutical research. However, the complex anatomy and physiology of the iBRB raise challenges for cell-based in vitro modeling. Methods/Results: This review follows the evolution of iBRB models—from simple monolayers of retinal endothelial cells (ECs) to sophisticated multicellular microphysiological systems (MPs). Advanced diverse microfluidic platforms aim to replicate key structural, biochemical and functional aspects of the iBRB, each incorporating distinct strategies regarding cell sourcing, device design, flow dynamics and functional readouts. Conclusions: Despite their limitations, these models are highly valuable for drug screening and mechanistic studies aimed at preserving or restoring barrier integrity while also helping to bridge the translational gap in ophthalmic drug discovery. Full article

Background/Objectives: Immune checkpoint inhibitors (ICIs) have been extensively used for the treatment of non-small cell lung cancer patients in recent years, providing a significant survival benefit. However, a major drawback of ICI-related immunotherapy is the risk of developing post-surgical pneumonitis. Methods: In this study, we propose a deep learning-embedded, multi-modality prediction approach to assess whether patients will develop ICI-pneumonitis after receiving ICI-based immunotherapy. This approach utilizes multi-modal data, including clinical data and pre-treatment lung screening computed tomography (CT) images. We extracted three types of features: (1) deep learning features from CT scans using a pre-trained vision transformer; (2) radiomic features from CT scans using pre-defined radiomic algorithms; (3) clinical features from patients’ electronic health records. We then compared ten machine learning algorithms for prediction based on these extracted features. Results: Our experiments demonstrated that using all three types of features leads to the best prediction result, with a prediction accuracy rate of 0.823 and an area under the receiver operating characteristic curve of 0.895. Conclusion: Multimodal approaches can result in superior prediction results compared to single modality approaches. This study demonstrates the feasibility of developing machine learning algorithms to accurately predict ICI-pneumonitis and contributes to the early identification of patients who are at a higher risk of developing pneumonitis. Full article

Glaucoma is an eye disease caused by increased intraocular pressure (IOP) that affects the optic nerve head (ONH), leading to vision problems and irreversible blindness. Background/Objectives: Glaucoma is the second leading cause of blindness worldwide, and the number of people affected is increasing each year, with the number expected to reach 111.8 million by 2040. This escalating trend is alarming due to the lack of ophthalmology specialists relative to the population. This study proposes an explainable end-to-end pipeline for automated glaucoma diagnosis from fundus images. It also evaluates the performance of Vision Transformers (ViTs) relative to traditional CNN-based models. Methods: The proposed system uses three datasets: REFUGE, ORIGA, and G1020. It begins with YOLOv11 for object detection of the optic disc. Then, the optic disc (OD) and optic cup (OC) are segmented using U-Net with ResNet50, VGG16, and MobileNetV2 backbones, as well as MaskFormer with a Swin-Base backbone. Glaucoma is classified based on the vertical cup-to-disc ratio (vCDR). Results: MaskFormer outperforms all models in segmentation in all aspects, including IoU OD, IoU OC, DSC OD, and DSC OC, with scores of 88.29%, 91.09%, 93.83%, and 93.71%. For classification, it achieved accuracy and F1-scores of 84.03% and 84.56%. Conclusions: By relying on the interpretable features of the vCDR, the proposed framework enhances transparency and aligns well with the principles of explainable AI, thus offering a trustworthy solution for glaucoma screening. Our findings show that Vision Transformers offer a promising approach for achieving high segmentation performance with explainable, biomarker-driven diagnosis. Full article