Application of artificial intelligence-driven endoscopic screening and diagnosis of gastric cancer

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.

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.

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.

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.

Background: Optical coherence tomography (OCT) is considered as a sensitive and noninvasive tool to evaluate the macular lesions. In patients with diabetes mellitus (DM), the existence of diabetic macular edema (DME) can cause significant vision impairment and further intravitreal injection (IVI) of anti–vascular endothelial growth factor (VEGF) is needed. However, the increasing number of DM patients makes it a big burden for clinicians to manually determine whether DME exists in the OCT images. The artificial intelligence (AI) now enormously applied to many medical territories may help reduce the burden on clinicians. Methods: We selected DME patients receiving IVI of anti-VEGF or corticosteroid at Taipei Veterans General Hospital in 2017. All macular cross-sectional scan OCT images were collected retrospectively from the eyes of these patients from January 2008 to July 2018. We further established AI models based on convolutional neural network architecture to determine whether the DM patients have DME by OCT images. Results: Based on the convolutional neural networks, InceptionV3 and VGG16, our AI system achieved a high DME diagnostic accuracy of 93.09% and 92.82%, respectively. The sensitivity of the VGG16 and InceptionV3 models was 96.48% and 95.15%., respectively. The specificity was corresponding to 86.67% and 89.63% for VGG16 and InceptionV3, respectively. We further developed an OCT-driven platform based on these AI models. Conclusion: We successfully set up AI models to provide an accurate diagnosis of DME by OCT images. These models may assist clinicians in screening DME in DM patients in the future.