According to recent studies, patients with COVID-19 have different feature characteristics on chest X-ray (CXR) than those with other lung diseases. This study aimed at evaluating the layer depths and degree of fine-tuning on transfer learning with a deep convolutional neural network (CNN)-based COVID-19 screening in CXR to identify efficient transfer learning strategies. The CXR images used in this study were collected from publicly available repositories, and the collected images were classified into three classes: COVID-19, pneumonia, and normal. To evaluate the effect of layer depths of the same CNN architecture, CNNs called VGG-16 and VGG-19 were used as backbone networks. Then, each backbone network was trained with different degrees of fine-tuning and comparatively evaluated. The experimental results showed the highest AUC value to be 0.950 concerning COVID-19 classification in the experimental group of a fine-tuned with only 2/5 blocks of the VGG16 backbone network. In conclusion, in the classification of medical images with a limited number of data, a deeper layer depth may not guarantee better results. In addition, even if the same pre-trained CNN architecture is used, an appropriate degree of fine-tuning can help to build an efficient deep learning model.

Otolaryngological diagnoses, such as otitis media, are traditionally performed using endoscopy, wherein diagnostic accuracy can be subjective and vary among clinicians. The integration of objective tools, like artificial intelligence (AI), could potentially improve the diagnostic process by minimizing the influence of subjective biases and variability. We systematically reviewed the AI techniques using medical imaging in otolaryngology. Relevant studies related to AI-assisted otitis media diagnosis were extracted from five databases: Google Scholar, PubMed, Medline, Embase, and IEEE Xplore, without date restrictions. Publications that did not relate to AI and otitis media diagnosis or did not utilize medical imaging were excluded. Of the 32identified studies, 26 used tympanic membrane images for classification, achieving an average diagnosis accuracy of 86% (range: 48.7–99.16%). Another three studies employed both segmentation and classification techniques, reporting an average diagnosis accuracy of 90.8% (range: 88.06–93.9%). These findings suggest that AI technologies hold promise for improving otitis media diagnosis, offering benefits for telemedicine and primary care settings due to their high diagnostic accuracy. However, to ensure patient safety and optimal outcomes, further improvements in diagnostic performance are necessary.

Background: We investigated whether fasting glucose (FG) variability could predict the risk of dementia.Methods: This cohort study analyzed data from Koreans with diabetes after at least three health examinations by the Korean National Health Insurance Corporation between 2005 and 2010, which included at least one examination between 2009 and 2010. A total of 769,554 individuals were included, excluding those aged <40 years and those with dementia. FG variability was measured using the variability independent of the mean (FG-VIM). The incidence of dementia was defined by the International Classification of Diseases 10th Revision codes and prescription of anti-dementia medication and was subdivided into Alzheimer’s disease (AD) and vascular dementia (VD).Results: During the 6.9-year follow-up, 54,837, 41,032, and 6,892 cases of all-cause dementia, AD, and VD, respectively, were identified. Cox proportional regression analyses showed that as the FG-VIM quartile increased, the risk of dementia serially increased after adjustment for metabolic factors, income status, and diabetes-related characteristics, including the mean FG. Participants in FG-VIM quartile 4 showed a 18%, 19%, and 17% higher risk for all-cause dementia, AD, and VD, respectively, than those in quartile 1; this particularly included non-obese patients with a longer duration of diabetes, high FG levels, dyslipidemia, and those taking glucose-lowering medications. Conversely, the baseline FG status and dementia showed a U-shaped association.Conclusion: Increased FG variability over 5 years can predict the risk of dementia in individuals with diabetes in Korea. This finding was more pronounced in patients with less favorable metabolic profiles.

This study aimed to develop a diabetes prediction model. The model performance was compared with logistic regression, and the decision tree Chi-square automatic interaction detection (CHAID) was used to predict diabetes. In total, 3233 patients were included in the analysis. Of these, 589 patients with diabetes and 2644 patients without diabetes were included after analyzing the study sample from the Korean Genome and Epidemiology Study (KoGES)-8 data. In this study, Diabetes Mellitus (DM) diagnosis prediction was compared with logistic regression and prediction through machine learning (ML) using the CHAID decision classification tree. We performed statistical analysis using the CHAID method with International Business Machine (IBM) statistical program SPSS®. We performed logistic regression analysis to predict the classification of diabetes accurately, and the total classification accuracy of the analysis was 81.7%, and the CHAID decision tree classification accuracy was 81.8%. A diabetes diagnosis decision tree was created, which included seven terminal nodes and three depth levels. This analysis showed that a blood pressure problem and hospital visits were the most decisive variables at the time of classification, and two risk levels were created for diabetes diagnosis. The suggested method is a valuable tool for predicting diabetes. Patients who visit the hospital because of blood pressure problems are more likely to develop diabetes than under-treating hyperlipidemia. The diabetes prediction model can help doctors make decisions by detecting the possibility of diabetes early; however, it is impossible to diagnose diabetes using only the model without the doctor’s opinion.

A diabetic wound presents a severe risk of infections and other complications because of its slow healing. Evaluating the pathophysiology during wound healing is imperative for wound care, necessitating a proper diabetic wound model and assay for monitoring. The adult zebrafish is a rapid and robust model for studying human cutaneous wound healing because of its fecundity and high similarities to human wound repair. OCTA as an assay can provide three-dimensional (3D) imaging of the tissue structure and vasculature in the epidermis, enabling monitoring of the pathophysiologic alterations in the zebrafish skin wound. We present a longitudinal study for assessing the cutaneous wound healing of the diabetic adult zebrafish model using OCTA, which is of importance for the diabetes research using the alternative animal models. We used non-diabetic (n = 9) and type 1 diabetes mellitus (DM) adult zebrafish models (n = 9). The full-thickness wound was generated on the fish skin, and the wound healing was monitored with OCTA for 15 days. The OCTA results demonstrated significant differences between diabetic and non-diabetic wound healing, involving delayed tissue remodeling and impaired angiogenesis for the diabetic wound, leading to slow wound recovery. The adult zebrafish model and OCTA technique may benefit long-term metabolic disease studies using zebrafish for drug development.

An early and accurate diagnosis of Candida albicans is critical for the rapid antifungal treatment of candidemia, a mortal bloodstream infection. This study demonstrates viscoelastic microfluidic techniques for continuous separation, concentration, and subsequent washing of Candida cells in the blood. The total sample preparation system contains two-step microfluidic devices: a closed-loop separation and concentration device and a co-flow cell-washing device. To determine the flow conditions of the closed-loop device, such as the flow rate factor, a mixture of 4 and 13 μm particles was used. Candida cells were successfully separated from the white blood cells (WBCs) and concentrated by 74.6-fold in the sample reservoir of the closed-loop system at 800 μL/min with a flow rate factor of 3.3. In addition, the collected Candida cells were washed with washing buffer (deionized water) in the microchannels with an aspect ratio of 2 at a total flow rate of 100 μL/min. Finally, Candida cells at extremely low concentrations (Ct > 35) became detectable after the removal of WBCs, the additional buffer solution in the closed-loop system (Ct = 30.3 ± 1.3), and further removal of blood lysate and washing (Ct = 23.3 ± 1.6).