ABSTRACT
Purpose: Diabetic macular edema (DME) is one of the leading causes of visual impairment in diabetic retinopathy (DR). Physicians rely on optical coherence tomography (OCT) and baseline visual acuity (VA) to tailor therapeutic regimen. However, best-corrected visual acuity (BCVA) from chart-based examinations may not wholly reflect DME status. Chart-based examinations are subjected findings dependent on the patient's recognition functions and are often confounded by concurrent corneal, lens, retinal, optic nerve, or extraocular disorders. The ability to infer VA from objective optical coherence tomography (OCT) images provides the predicted VA from objective macular structures directly and a better understanding of diabetic macular health. Deviations from chart-based and artificial intelligence (AI) image-based VA will prompt physicians to assess other ocular abnormalities affecting the patients VA and whether pursuing anti-VEGF treatment will likely yield increment in VA. Materials and methods: We enrolled a retrospective cohort of 251 DME patients from Big Data Center (BDC) of Taipei Veteran General Hospital (TVGH) from February 2011 and August 2019. A total of 3,920 OCT images, labeled as "visually impaired" or "adequate" according to baseline VA, were grouped into training (2,826), validation (779), and testing cohort (315). We applied confusion matrix and receiver operating characteristic (ROC) curve to evaluate the performance. Results: We developed an OCT-based convolutional neuronal network (CNN) model that could classify two VA classes by the threshold of 0.50 (decimal notation) with an accuracy of 75.9%, a sensitivity of 78.9%, and an area under the ROC curve of 80.1% on the testing cohort. Conclusion: This study demonstrated the feasibility of inferring VA from routine objective retinal images. Translational relevance: Serves as a pilot study to encourage further use of deep learning in deriving functional outcomes and secondary surrogate endpoints for retinal diseases.
ABSTRACT
Diabetic macular edema (DME) is a highly common cause of vision loss in patients with diabetes. Optical coherence tomography (OCT) is crucial in classifying DME and tracking the results of DME treatment. The presence of intraretinal cystoid fluid (IRC) and subretinal fluid (SRF) and the disruption of the ellipsoid zone (EZ), which is part of the photoreceptor layer, are three crucial factors affecting the best corrected visual acuity (BCVA). However, the manual segmentation of retinal fluid and the EZ from retinal OCT images is laborious and time-consuming. Current methods focus only on the segmentation of retinal features, lacking a correlation with visual acuity. Therefore, we proposed a modified U-net, a deep learning algorithm, to segment these features from OCT images of patients with DME. We also correlated these features with visual acuity. The IRC, SRF, and EZ of the OCT retinal images were manually labeled and checked by doctors. We trained the modified U-net model on these labeled images. Our model achieved Sørensen-Dice coefficients of 0.80 and 0.89 for IRC and SRF, respectively. The area under the receiver operating characteristic curve (ROC) for EZ disruption was 0.88. Linear regression indicated that EZ disruption was the factor most strongly correlated with BCVA. This finding agrees with that of previous studies on OCT images. Thus, we demonstrate that our segmentation network can be feasibly applied to OCT image segmentation and assist physicians in assessing the severity of the disease.
ABSTRACT
X-linked juvenile retinoschisis (XLRS) is one of the common early-onset hereditary retinal degenerative diseases in men. The common symptoms of XLRS range from mild to severe central vision loss and radial stripes created by the fovea, the division of the inner layer of the retina in the peripheral retina and the significant decrease in b-wave amplitude (ERG). Retinoschisin, the 224-amino-acid protein product of the retinoschisis 1 (RS1) gene, contains a discoid domain as the primary structural unit, an N-terminal cleavable signal sequence, and an oligomerization-area component. Retinoschisin is a homo-octamer complex with disulfide links that are released by retinal cells. It helps preserve the retina's integrity by binding to the surface of photoreceptors and bipolar cells. As a recessive genetic disease, XLRS was usually treated by prescribing low vision aids in most clinical cases. A gene replacement therapy based on adeno-associated virus vectors was initiated and showed a breakthrough in treating XLRS in 2014. Understanding the revolution of gene therapy for treating XLRS may accelerate its development and make this gene therapy the template for developing therapeutics against other inherited retinal diseases.
Subject(s)
Retinoschisis , Electroretinography , Eye Proteins/genetics , Genetic Therapy , Humans , Male , Retina , Retinoschisis/genetics , Retinoschisis/metabolism , Retinoschisis/therapyABSTRACT
Background: Adequate bowel cleansing is important for colonoscopy performance evaluation. Current bowel cleansing evaluation scales are subjective, with a wide variation in consistency among physicians and low reported rates of accuracy. We aim to use machine learning to develop a fully automatic segmentation method for the objective evaluation of the adequacy of colon preparation. Methods: Colonoscopy videos were retrieved from a video data cohort and transferred to qualified images, which were randomly divided into training, validation, and verification datasets. The fecal residue was manually segmented. A deep learning model based on the U-Net convolutional network architecture was developed to perform automatic segmentation. The performance of the automatic segmentation was evaluated on the overlap area with the manual segmentation. Results: A total of 10,118 qualified images from 119 videos were obtained. The model averaged 0.3634 s to segmentate one image automatically. The models produced a strong high-overlap area with manual segmentation, with 94.7% ± 0.67% of that area predicted by our AI model, which correlated well with the area measured manually (r = 0.915, p < 0.001). The AI system can be applied in real-time qualitatively and quantitatively. Conclusions: We established a fully automatic segmentation method to rapidly and accurately mark the fecal residue-coated mucosa for the objective evaluation of colon preparation.
ABSTRACT
Electrotherapy or electrical stimulation (ES) is a part of clinical intervention in the rehabilitation field. With rehabilitation intervention, electrotherapy may be provided as a treatment for pain relief, strengthening, muscle education, wound recovery, or functional training. Although these interventions may not be considered as the primary therapy for patients, the advantages of the ease of operation, lower costs, and lower risks render ES to be applied frequently in clinics. There have also been emerging ES tools for brain modulation in the past decade. ES interventions are not only considered analgesics but also as an important assistive therapy for motor improvement in orthopedic and neurological rehabilitation. In addition, during the coronavirus disease pandemic, lockdowns and self-quarantine policies have led to the discontinuation of orthopedic and neurological rehabilitation interventions. Therefore, the feasibility and effectiveness of home-based electrotherapy may provide opportunities for the prevention of deterioration or extension of the original therapy. The most common at-home applications in previous studies showed positive effects on pain relief, functional ES, muscle establishment, and motor training. Currently, there is a lack of certain products for at-home brain modulation; however, transcranial direct current stimulation has shown the potential of future home-based rehabilitation due to its relatively small and simple design. We have organized the features and applications of ES tools and expect the future potential of remote therapy during the viral pandemic.
Subject(s)
COVID-19/epidemiology , Electric Stimulation Therapy/methods , Neurological Rehabilitation , Orthopedic Procedures , SARS-CoV-2 , Electric Stimulation Therapy/adverse effects , Humans , Transcranial Direct Current Stimulation , Transcutaneous Electric Nerve StimulationABSTRACT
OBJECTIVES: Computer-aided diagnosis (CAD)-based artificial intelligence (AI) has been shown to be highly accurate for detecting and characterizing colon polyps. However, the application of AI to identify normal colon landmarks and differentiate multiple colon diseases has not yet been established. We aimed to develop a convolutional neural network (CNN)-based algorithm (GUTAID) to recognize different colon lesions and anatomical landmarks. METHODS: Colonoscopic images were obtained to train and validate the AI classifiers. An independent dataset was collected for verification. The architecture of GUTAID contains two major sub-models: the Normal, Polyp, Diverticulum, Cecum and CAncer (NPDCCA) and Narrow-Band Imaging for Adenomatous/Hyperplastic polyps (NBI-AH) models. The development of GUTAID was based on the 16-layer Visual Geometry Group (VGG16) architecture and implemented on Google Cloud Platform. RESULTS: In total, 7838 colonoscopy images were used for developing and validating the AI model. An additional 1273 images were independently applied to verify the GUTAID. The accuracy for GUTAID in detecting various colon lesions/landmarks is 93.3% for polyps, 93.9% for diverticula, 91.7% for cecum, 97.5% for cancer, and 83.5% for adenomatous/hyperplastic polyps. CONCLUSIONS: A CNN-based algorithm (GUTAID) to identify colonic abnormalities and landmarks was successfully established with high accuracy. This GUTAID system can further characterize polyps for optical diagnosis. We demonstrated that AI classification methodology is feasible to identify multiple and different colon diseases.
Subject(s)
Artificial Intelligence , Colonic Polyps , Algorithms , Colonic Polyps/diagnostic imaging , Colonoscopy/methods , Humans , Machine LearningABSTRACT
BACKGROUND: Acute myocardial infarction (AMI) and atrial fibrillation (AF) are risk factors for stroke. The risk of stroke after AMI may differ between patients with and without AF. The aim of this study was to evaluate the impact of AF on stroke in patients after the first AMI. METHODS: This was a retrospective, nationwide cohort study. Patients with a primary diagnosis of a first AMI between 2000 and 2012 were included. All patients were followed up until ischemic stroke or transient ischemic attack (TIA), or December 31, 2012, whichever occurred first. Kaplan-Meier cumulative survival curves were constructed to compare ischemic stroke or TIA between AMI patients with and without AF. RESULTS: A total of 170 472 patients were enrolled in this study. Among them, 8530 patients with AF were identified. The propensity score matching technique was used to match 8530 patients without AF of similar ages and sexes. Overall, the 12-year stroke rate was significantly higher in patients with AF than in those without AF (log-rank p < 0.001), including different sexes, ages, and interventional therapy subgroups. Patients with pre-existing AF had higher stroke rates than those with newly diagnosed AF in male sex, age below 65 years, and those receiving interventional therapy subgroups. In Cox proportional-hazard regression analysis, AF was an independent risk factor for stroke after the first AMI (hazard ratio, 1.67; 95% CI: 1.5-1.87). CONCLUSION: AF significantly increases stroke risk after the first AMI. In patients with AF, those with pre-existing AF have higher stroke risks in male sex, age below 65 years, and those with interventional therapy than those with newly diagnosed AF.
Subject(s)
Atrial Fibrillation , Myocardial Infarction/complications , Stroke/etiology , Aged , Cohort Studies , Female , Humans , Male , Middle Aged , Risk AssessmentABSTRACT
BACKGROUND: The coronavirus disease 2019 (COVID-19) pandemic continues to affect countries worldwide. To inhibit the transmission of severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2), testing of patients, contact tracing, and quarantine of their close contacts have been used as major nonpharmaceutical interventions. The advantages of antigen tests, such as low cost and rapid turnaround, may allow for the rapid identification of larger numbers of infectious persons. This meta-analysis aimed to evaluate the diagnostic accuracy of antigen tests for SARS-CoV-2. METHODS: We searched PubMed, Embase, Cochrane Library, and Biomed Central databases from inception to January 2, 2021. Studies evaluating the diagnostic accuracy of antigen testing for SARS-CoV-2 with reference standards were included. We included studies that provided sufficient data to construct a 2 × 2 table on a per-patient basis. Only articles in English were reviewed. Summary sensitivity and specificity for antigen tests were generated using a random-effects model. RESULTS: Fourteen studies with 8624 participants were included. The meta-analysis for antigen testing generated a pooled sensitivity of 79% (95% CI, 66%-88%; 14 studies, 8624 patients) and a pooled specificity of 100% (95% CI, 99%-100%; 14 studies, 8624 patients). The subgroup analysis of studies that reported specimen collection within 7 days after symptom onset showed a pooled sensitivity of 95% (95% CI, 78%-99%; four studies, 1342 patients) and pooled specificity of 100% (95% CI, 97%-100%; four studies, 1342 patients). Regarding the applicability, the patient selection, index tests, and reference standards of studies in our meta-analysis matched the review title. CONCLUSION: Antigen tests have moderate sensitivity and high specificity for the detection of SARS-CoV-2. Antigen tests might have a higher sensitivity in detecting SARS-CoV-2 within 7 days after symptom onset. Based on our findings, antigen testing might be an effective method for identifying contagious individuals to block SARS-CoV-2 transmission.
Subject(s)
Antigens, Viral/analysis , COVID-19 Serological Testing/methods , COVID-19/diagnosis , SARS-CoV-2/immunology , Humans , Sensitivity and SpecificityABSTRACT
Non-viral gene delivery holds promises for treating inherited diseases. However, the limited cloning capacity of plasmids may hinder the co-delivery of distinct genes to the transfected cells. Previously, the conjugation of maleimide-functionalized polyurethane grafted with small molecular weight polyethylenimine (PU-PEI600-Mal) using 1,6-hexanedithiol (HDT) could promote the co-delivery and extensive co-expression of two different plasmids in target cells. Herein, we designed HDT-conjugated PU-PEI600-Mal for the simultaneous delivery of CRISPR/Cas9 components to achieve efficient gene correction in the induced pluripotent stem cell (iPSC)-derived model of Fabry cardiomyopathy (FC) harboring GLA IVS4 + 919 G > A mutation. This FC in vitro model recapitulated several clinical FC features, including cardiomyocyte hypertrophy and lysosomal globotriaosylceramide (Gb3) deposition. As evidenced by the expression of two reporter genes, GFP and mCherry, the addition of HDT conjugated two distinct PU-PEI600-Mal/DNA complexes and promoted the co-delivery of sgRNA/Cas9 and homology-directed repair DNA template into target cells to achieve an effective gene correction of IVS4 + 919 G > A mutation. PU-PEI600-Mal/DNA with or without HDT-mediated conjugation consistently showed neither the cytotoxicity nor an adverse effect on cardiac induction of transfected FC-iPSCs. After the gene correction and cardiac induction, disease features, including cardiomyocyte hypertrophy, the mis-regulated gene expressions, and Gb3 deposition, were remarkably rescued in the FC-iPSC-differentiated cardiomyocytes. Collectively, HDT-conjugated PU-PEI600-Mal-mediated dual DNA transfection system can be an ideal approach to improve the concurrent transfection of non-viral-based gene editing system in inherited diseases with specific mutations.
ABSTRACT
The landscape of gastrointestinal endoscopy continues to evolve as new technologies and techniques become available. The advent of image-enhanced and magnifying endoscopies has highlighted the step toward perfecting endoscopic screening and diagnosis of gastric lesions. Simultaneously, with the development of convolutional neural network, artificial intelligence (AI) has made unprecedented breakthroughs in medical imaging, including the ongoing trials of computer-aided detection of colorectal polyps and gastrointestinal bleeding. In the past demi-decade, applications of AI systems in gastric cancer have also emerged. With AI's efficient computational power and learning capacities, endoscopists can improve their diagnostic accuracies and avoid the missing or mischaracterization of gastric neoplastic changes. So far, several AI systems that incorporated both traditional and novel endoscopy technologies have been developed for various purposes, with most systems achieving an accuracy of more than 80%. However, their feasibility, effectiveness, and safety in clinical practice remain to be seen as there have been no clinical trials yet. Nonetheless, AI-assisted endoscopies shed light on more accurate and sensitive ways for early detection, treatment guidance and prognosis prediction of gastric lesions. This review summarizes the current status of various AI applications in gastric cancer and pinpoints directions for future research and clinical practice implementation from a clinical perspective.
Subject(s)
Artificial Intelligence , Stomach Neoplasms , Early Detection of Cancer , Endoscopy, Gastrointestinal , Humans , Neural Networks, Computer , Stomach Neoplasms/diagnostic imagingABSTRACT
Osteoarthritis (OA) is a common degenerative disease; however, its exact pathophysiology and early diagnosis are still a challenge. Growing attention to the exosomes may inspire innovations that would make the current management of OA more effective. The exosomes in synovial fluid are relatively stable, and they can be easily isolated by the relatively noninvasive procedure of liquid biopsy to provide diagnostic and monitoring value. Some miRNAs (miR-504, miR-146a, miR-26a, miR-200c, and miR-210) have been known to be secreted in exosomes of OA patients. On the other hand, intraarticular injection of platelet-rich plasma (PRP) is becoming a popular therapy for OA patients. PRP is also a source of exosomes and their numerous contents. It is evident from the literature that PRP-derived exosomes can induce chondrogenic gene expression in OA chondrocytes. Here, we review the latest findings on the roles of exosomes in OA with the emphasis on PRP-derived exosomes and their potential applications for treating OA.
Subject(s)
Exosomes/metabolism , Osteoarthritis/physiopathology , Biomarkers , Female , Humans , Liquid Biopsy , Male , Synovial FluidABSTRACT
Inherited retinal dystrophies (IRDs) are rare but highly heterogeneous genetic disorders that affect individuals and families worldwide. However, given its wide variability, its analysis of the driver genes for over 50% of the cases remains unexplored. The present study aims to identify novel driver genes, disease-causing variants, and retinitis pigmentosa (RP)-associated pathways. Using family-based whole-exome sequencing (WES) to identify putative RP-causing rare variants, we identified a total of five potentially pathogenic variants located in genes OR56A5, OR52L1, CTSD, PRF1, KBTBD13, and ATP2B4. Of the variants present in all affected individuals, genes OR56A5, OR52L1, CTSD, KBTBD13, and ATP2B4 present as missense mutations, while PRF1 and CTSD present as frameshift variants. Sanger sequencing confirmed the presence of the novel pathogenic variant PRF1 (c.124_128del) that has not been reported previously. More causal-effect or evidence-based studies will be required to elucidate the precise roles of these SNPs in the RP pathogenesis. Taken together, our findings may allow us to explore the risk variants based on the sequencing data and upgrade the existing variant annotation database in Taiwan. It may help detect specific eye diseases such as retinitis pigmentosa in East Asia.
Subject(s)
Cathepsin D/genetics , Genetic Predisposition to Disease , Retinal Dystrophies/genetics , Adult , Aged , Cathepsin D/blood , Female , Frameshift Mutation , Gene Ontology , Humans , Male , Middle Aged , Muscle Proteins/genetics , Mutation, Missense , Pedigree , Perforin/genetics , Plasma Membrane Calcium-Transporting ATPases/genetics , Polymorphism, Single Nucleotide , Protein Interaction Maps , Retinal Dystrophies/congenital , Retinal Dystrophies/pathology , Retinitis Pigmentosa/congenital , Retinitis Pigmentosa/diagnostic imaging , Retinitis Pigmentosa/genetics , Retinitis Pigmentosa/pathology , Risk Factors , Tomography, Optical Coherence , Exome SequencingABSTRACT
Angiotensin-converting enzyme 2 (ACE2) was identified as the main host cell receptor for the entry of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) and its subsequent infection. In some coronavirus disease 2019 (COVID-19) patients, it has been reported that the nervous tissues and the eyes were also affected. However, evidence supporting that the retina is a target tissue for SARS-CoV-2 infection is still lacking. This present study aimed to investigate whether ACE2 expression plays a role in human retinal neurons during SARS-CoV-2 infection. Human induced pluripotent stem cell (hiPSC)-derived retinal organoids and monolayer cultures derived from dissociated retinal organoids were generated. To validate the potential entry of SARS-CoV-2 infection in the retina, we showed that hiPSC-derived retinal organoids and monolayer cultures endogenously express ACE2 and transmembrane serine protease 2 (TMPRSS2) on the mRNA level. Immunofluorescence staining confirmed the protein expression of ACE2 and TMPRSS2 in retinal organoids and monolayer cultures. Furthermore, using the SARS-CoV-2 pseudovirus spike protein with GFP expression system, we found that retinal organoids and monolayer cultures can potentially be infected by the SARS-CoV-2 pseudovirus. Collectively, our findings highlighted the potential of iPSC-derived retinal organoids as the models for ACE2 receptor-based SARS-CoV-2 infection.
Subject(s)
Angiotensin-Converting Enzyme 2/genetics , COVID-19/genetics , Gene Expression , Induced Pluripotent Stem Cells/cytology , Retina/cytology , SARS-CoV-2/physiology , Angiotensin-Converting Enzyme 2/metabolism , COVID-19/metabolism , Cell Culture Techniques , Cell Line , Humans , Induced Pluripotent Stem Cells/metabolism , Organoids/cytology , Organoids/metabolism , Retina/metabolism , Serine Endopeptidases/genetics , Serine Endopeptidases/metabolism , Virus InternalizationABSTRACT
A single-blind study to investigate the effects of noisy galvanic vestibular stimulation (nGVS) in straight walking and 2 Hz head yaw walking for healthy and bilateral vestibular hypofunction (BVH) participants in light and dark conditions. The optimal stimulation intensity for each participant was determined by calculating standing stability on a force plate while randomly applying six graded nGVS intensities (0-1000 µA). The chest-pelvic (C/P) ratio and lateral deviation of the center of mass (COM) were measured by motion capture during straight and 2 Hz head yaw walking in light and dark conditions. Participants were blinded to nGVS served randomly and imperceivably. Ten BVH patients and 16 healthy participants completed all trials. In the light condition, the COM lateral deviation significantly decreased only in straight walking (p = 0.037) with nGVS for the BVH. In the dark condition, both healthy (p = 0.026) and BVH (p = 0.017) exhibited decreased lateral deviation during nGVS. The C/P ratio decreased significantly in BVH for 2 Hz head yaw walking with nGVS (p = 0.005) in light conditions. This study demonstrated that nGVS effectively reduced walking deviations, especially in visual deprived condition for the BVH. Applying nGVS with different head rotation frequencies and light exposure levels may accelerate the rehabilitation process for patients with BVH.Clinical Trial Registration This clinical trial was prospectively registered at www.clinicaltrials.gov with the Unique identifier: NCT03554941. Date of registration: (13/06/2018).
Subject(s)
Electric Stimulation , Noise , Postural Balance , Sensory Deprivation , Vestibule, Labyrinth/physiopathology , Vision, Ocular , Adult , Aged , Female , Gait , Humans , Male , Middle Aged , Motor Activity , Psychomotor Performance , WalkingABSTRACT
BACKGROUND: Cardiotocography is a common method of electronic fetal monitoring (EFM) for fetal well-being. Data-driven analyses have shown potential for automated EFM assessment. For this preliminary study, we used a novel artificial intelligence method based on fully convolutional networks (FCNs), with deep learning for EFM evaluation and correct recognition, and its possible role in evaluation of nonreassuring fetal status. METHODS: We retrospectively collected 3239 EFM labor records from 292 deliveries and neonatal Apgar scores between December 2018 and July 2019 at a single medical center. We analyzed these data using an FCN model and compared the results with clinical practice. RESULTS: The FCN model recognized EFM traces like physicians, with an average Cohen's kappa coefficient of agreement of 0.525 and average area under the receiver operating characteristic curve of 0.892 for six fetal heart rate (FHR) categories. The FCN model showed higher sensitivity for predicting fetal compromise (0.528 vs 0.132) but a higher false-positive rate (0.632 vs 0.012) compared with clinical practice. CONCLUSION: FCN is a modern technique that may be useful for EFM trace recognition based on its multiconvolutional layered analysis. Our model showed a competitive ability to identify FHR patterns and the potential for evaluation of nonreassuring fetal status.
Subject(s)
Artificial Intelligence , Cardiotocography/methods , Fetal Monitoring/instrumentation , Fetal Monitoring/methods , Heart Rate, Fetal/physiology , Adult , Female , Humans , Medical Audit , Pregnancy , Retrospective StudiesABSTRACT
BACKGROUND: Embedded within most rapid head rotations are gaze shifts, which is an initial eye rotation to a target of interest, followed by a head rotation towards the same target. Gaze shifts are used to acquire an image that initially is outside of the participant's current field of vision. Currently, there are no tools available that evaluate the functional relevance of a gaze shift. OBJECTIVE: The purpose of our study was to measure dynamic visual acuity (DVA) while performing a gaze shift. METHODS: Seventy-one healthy participants (42.79±16.89 years) and 34 participants with unilateral vestibular hypofunction (UVH) (54.59±20.14 years) were tested while wearing an inertial measurement unit (IMU) sensor on the head and walking on a treadmill surrounded by three monitors. We measured visual acuity during three subcomponent tests: standing (static visual acuity), while performing an active head rotation gaze shift, and an active head rotation gaze shift while walking (gsDVAw). RESULTS: While doing gsDVAw, patients with Left UVH (nâ=â21) had scores worse (pâ=â0.023) for leftward (0.0446±0.0943 LogMAR) head rotation compared with the healthy controls (-0.0075±0.0410 LogMAR). Similarly, patients with right UVH (Nâ=â13) had worse (pâ=â0.025) gsDVAw for rightward head motion (0.0307±0.0481 LogMAR) compared with healthy controls (-0.0047±0.0433 LogMAR). As a whole, gsDVAw scores were worse in UVH compared to the healthy controls when we included the ipsilesional head rotation on both sides gsDVAw (0.0061±0.0421 LogMAR healthy vs. 0.03926±0.0822 LogMAR UVH, pâ=â0.003). Controlling for age had no effect, the gsDVAw scores of the patients were always worse (pâ<â0.01). CONCLUSION: The gaze shift DVA test can distinguish gaze stability in patients with UVH from healthy controls. This test may be a useful measure of compensation for patients undergoing various therapies for their vestibular hypofunction.
Subject(s)
Head Movements , Vestibular Diseases , Humans , Reflex, Vestibulo-Ocular , Vestibular Diseases/diagnosis , Vision Tests , Visual Acuity , WalkingABSTRACT
BACKGROUND: Diabetic macular edema (DME) is a sight-threatening condition that needs regular examinations and remedies. Optical coherence tomography (OCT) is the most common used examination to evaluate the structure and thickness of the macula, but the software in the OCT machine does not tell the clinicians whether DME exists directly. Recently, artificial intelligence (AI) is expected to aid in diagnosis generation and therapy selection. We thus develop a smartphone-based offline AI system that provides diagnostic suggestions and medical strategies through analyzing OCT images from diabetic patients at the risk of developing DME. METHODS: DME patients receiving treatments in 2017 at Taipei Veterans General Hospital were included in this study. We retrospectively collected the OCT images of these patients from January 2008 to July 2018. We established the AI model based on MobileNet architecture to classify the OCT images conditions. The confusion matrix has been applied to present the performance of the trained AI model. RESULTS: Based on the convolutional neural network with the MobileNet model, our AI system achieved a high DME diagnostic accuracy of 90.02%, which is comparable to other AI systems such as InceptionV3 and VGG16. We further developed a mobile-application based on this AI model available at https://aicl.ddns.net/DME.apk. CONCLUSION: We successful integrated an AI model into the mobile device to provide an offline method to provide the diagnosis for quickly screening the risk of developing DME. With the offline property, our model could help those nonophthalmological healthcare providers in offshore islands or underdeveloped countries.
Subject(s)
Artificial Intelligence , Diabetic Retinopathy/diagnostic imaging , Macular Edema/diagnostic imaging , Smartphone , Humans , Image Processing, Computer-Assisted , Neural Networks, Computer , Taiwan , Tomography, Optical CoherenceABSTRACT
Patients with bilateral vestibular hypofunction (BVH) often suffer from imbalance, gait problems, and oscillopsia. Noisy galvanic vestibular stimulation (GVS), a technique that non-invasively stimulates the vestibular afferents, has been shown to enhance postural and walking stability. However, no study has investigated how it affects stability and neural activities while standing and walking with a 2 Hz head yaw turning. Herein, we investigated this issue by comparing differences in neural activities during standing and walking with a 2 Hz head turning, before and after noisy GVS. We applied zero-mean gaussian white noise signal stimulations in the mastoid processes of 10 healthy individuals and seven patients with BVH, and simultaneously recorded electroencephalography (EEG) signals with 32 channels. We analyzed the root mean square (RMS) of the center of pressure (COP) sway during 30 s of standing, utilizing AMTI force plates (Advanced Mechanical Technology Inc., Watertown, MA, USA). Head rotation quality when walking with a 2 Hz head yaw, with and without GVS, was analyzed using a VICON system (Vicon Motion Systems Ltd., Oxford, UK) to evaluate GVS effects on static and dynamic postural control. The RMS of COP sway was significantly reduced during GVS while standing, for both patients and healthy subjects. During walking, 2 Hz head yaw movements was significantly improved by noisy GVS in both groups. Accordingly, the EEG power of theta, alpha, beta, and gamma bands significantly increased in the left parietal lobe after noisy GVS during walking and standing in both groups. GVS post-stimulation effect changed EEG activities in the left and right precentral gyrus, and the right parietal lobe. After stimulation, EEG activity changes were greater in healthy subjects than in patients. Our findings reveal noisy GVS as a non-invasive therapeutic alternative to improve postural stability in patients with BVH. This novel approach provides insight to clinicians and researchers on brain activities during noisy GVS in standing and walking conditions in both healthy and BVH patients.
ABSTRACT
To evaluate vestibular function in the clinic, current assessments are applied under static conditions, such as with the subject in a sitting or supine position. Considering the complexities of daily activities, the combination of dynamic activities, dynamic visual acuity (DVA) and postural control could produce an evaluation that better reflects vestibular function in daily activities. Objective: To develop a novel sensor-based system to investigate DVA, walking trajectory, head and trunk movements and the chest-pelvis rotation ratio during forward and backward overground walking in both healthy individuals and patients with vestibular hypofunction. Methods: Fifteen healthy subjects and 7 patients with bilateral vestibular hypofunction (BVH) were recruited for this study. Inertial measurement units were placed on each subject's head and torso. Each subject walked forward and backward for 5 m twice with 2 Hz head yaw. Our experiment comprised 2 stages. In stage 1, we measured forward (FW), backward (BW), and medial-lateral (MLW) walking trajectories; head and trunk movements; and the chest-pelvis rotation ratio. In stage 2, we measured standing and locomotion DVA (loDVA). Using Mann-Whitney U-test, we compared the abovementioned parameters between the 2 groups. Results: Patients exhibited an in-phase chest/pelvis reciprocal rotation ratio only in FW. The walking trajectory deviation, calculated by normalizing the summation of medial-lateral swaying with 1/2 body height (%), was significantly larger (FW mean ± standard deviation: 20.4 ± 7.1% (median (M)/interquartile range (IQR): 19.3/14.4-25.2)in healthy vs. 43.9 ± 27. 3% (M/IQR: 36.9/21.3-56.9) in patients, p = 0.020)/(BW mean ± standard deviation: 19.2 ± 11.5% (M/IQR: 13.6/10.4-25.3) in healthy vs. 29.3 ± 6.4% (M/IQR: 27.7/26.5-34.4) in patients, p = 0.026), and the walking DVA was also significantly higher (LogMAR score in the patient group [FW LogMAR: rightDVA: mean ± standard deviation:0.127 ± 0.081 (M/IQR: 0.127/0.036-0.159) in healthy vs. 0.243 ± 0.101 (M/IQR: 0.247/0.143-0.337) in patients (p = 0.013) and leftDVA: 0.136 ± 0.096 (M/IQR: 0.127/0.036-0.176) in healthy vs. 0.258 ± 0.092 (M/IQR: 0.247/0.176-0.301) in patients (p = 0.016); BW LogMAR: rightDVA: mean ± standard deviation: 0.162 ± 0.097 (M/IQR: 0.159/0.097-0.273) in healthy vs. 0.281 ± 0.130 (M/IQR: 0.273/0.176-0.418) in patients(p = 0.047) and leftDVA: 0.156 ± 0.101 (M/IQR: 0.159/0.097-0.198) in healthy vs. 0.298 ± 0.153 (M/IQR: 0.2730/0.159-0.484) in patients (p = 0.038)]. Conclusions: Our sensor-based vestibular evaluation system provided a more functionally relevant assessment for the identification of BVH patients.
ABSTRACT
Artificial intelligence (AI) has been widely applied in the medical field and achieved enormous milestones in helping specialists to make diagnosis and remedy decisions, particularly in the field of eye diseases and ophthalmic screening. With the development of AI-based systems, the enormous hardware and software resources are required for optimal performance. In reality, there are many places on the planet where such resources are highly limited. Hence, the smartphone-based AI systems can be used to provide a remote control route to quickly screen eye diseases such as diabetic-related retinopathy or diabetic macular edema. However, the performance of such mobile-based AI systems is still uncharted territory. In this article, we discuss the issues of computing resource consumption and performance of the mobile device-based AI systems and highlight recent research on the feasibility and future potential of application of the mobile device-based AI systems in telemedicine.