Knowledge, attitude and practices on diabetes, hypertension and diabetic retinopathy and the factors that motivate screening for diabetes and diabetic retinopathy in a pyramidal model of eye health care.

INTRODUCTION: To report the results of a knowledge, attitude and practice (KAP) study related to diabetes mellitus (DM), hypertension and diabetic retinopathy (DR) of patient populations in India at different levels (Tertiary (T), Secondary (S) and Primary (P)) of a pyramidal model of eye health care. METHODS: In total, 202 participants, composed of equal numbers of diabetic and non-diabetic patients at a Tertiary urban facility (T), a Secondary rural facility (S) and a Primary (P) community-screening program, were surveyed on their knowledge, knowledge sources, attitudes, practices and factors that motivate use of eye health services. RESULTS: People with diabetes had a higher mean knowledge and attitude score about DM, hypertension and DR (67.3% T, 59.4% S, 47.0% P) than non-diabetics (41.8% T, 29.0% S, 23.5% P; p<0.001). Awareness of DR was more 65.3% among diabetics compared with 22.0% among non-diabetics at all locations. Most participants in all locations were aware of hypertension (84.0% T, 65.3% S, 52.9% P), but few knew it could affect the eyes (30.0% T, 12.2% S, 13.7% P) or be associated with diabetic complications (30.0% T, 32.7% S, 21.8% P). Many participants had never previously had a dilated eye examination (2% T, 40% S, 50% P). Participants were motivated to visit an eye facility for a routine checkup (70.6%), poor vision (22.6%) or a glucose/blood pressure test (17.7%) at a Primary-level facility and for follow-up or poor vision at the other facilities (28% and 42% Tertiary, 50% and 30% Secondary). CONCLUSION: Practice-oriented education and advertising of facilities tailored for the relevant populations at each level of an eye health pyramid and continuation of fundus, glucose and blood pressure screening programs can help in creating awareness about diabetes, hypertension and diabetic retinopathy.

High-Throughput Classification of Radiographs Using Deep Convolutional Neural Networks.

The study aimed to determine if computer vision techniques rooted in deep learning can use a small set of radiographs to perform clinically relevant image classification with high fidelity. One thousand eight hundred eighty-five chest radiographs on 909 patients obtained between January 2013 and July 2015 at our institution were retrieved and anonymized. The source images were manually annotated as frontal or lateral and randomly divided into training, validation, and test sets. Training and validation sets were augmented to over 150,000 images using standard image manipulations. We then pre-trained a series of deep convolutional networks based on the open-source GoogLeNet with various transformations of the open-source ImageNet (non-radiology) images. These trained networks were then fine-tuned using the original and augmented radiology images. The model with highest validation accuracy was applied to our institutional test set and a publicly available set. Accuracy was assessed by using the Youden Index to set a binary cutoff for frontal or lateral classification. This retrospective study was IRB approved prior to initiation. A network pre-trained on 1.2 million greyscale ImageNet images and fine-tuned on augmented radiographs was chosen. The binary classification method correctly classified 100 % (95 % CI 99.73-100 %) of both our test set and the publicly available images. Classification was rapid, at 38 images per second. A deep convolutional neural network created using non-radiological images, and an augmented set of radiographs is effective in highly accurate classification of chest radiograph view type and is a feasible, rapid method for high-throughput annotation.

Anti-adipocyte antibody response in patients with non-alcoholic fatty liver disease.

BACKGROUND AND AIM: A significant number of autoantibodies have been reported in patients with non-alcoholic fatty liver disease (NAFLD) patients. In the present study, our aim was to assess the role of disease and cell-specific antibodies, namely anti-adipocyte antibodies (anti-AdAb) in patients with NAFLD and non-alcoholic steatohepatitis (NASH). METHODS: Flow cytometry was used to detect the presence of anti-AdAb (immunoglobulin M [IgM] and immunoglobulin G [IgG]) in sera from patients with biopsy-proven NAFLD (n = 98) and in controls (n = 49) without liver disease. Univariate and multivariate analysis was performed to draw associations between anti-AdAb IgM and IgG levels and the different clinical variables. RESULTS: Patients with NAFLD had significantly higher levels of anti-AdAb IgM and significantly lower levels of AdAb IgG when compared with controls (P = 0.002 and P < 0.001, respectively). Patients with NASH had significantly higher levels of anti-AdAb IgM when compared with non-NASH NAFLD patients, P = 0.04. In multivariate analysis, anti-AdAb IgM was independently associated with a higher risk for NASH (odds ratio[OR]: 2.90 [confidence interval (CI) 1.18-7.16], P = 0.02). Anti-AdAb IgM was also found to be independently associated with portal inflammation in patients with NAFLD (OR: 3.01 [CI 1.15-7.90 P = 0.02]). CONCLUSIONS: Anti-AdAb IgM was independently associated with NAFLD and NASH while anti-AdAb IgG was found to be protective against NAFLD. Anti-AdAb IgM was found specifically to be associated with the inflammatory processes in NAFLD. These findings indicate that the anti-AdAb IgM and IgG may play an immunomodulatory role in the pathogenesis of NAFLD and NASH.

Comparison of convolutional neural networks for detecting large vessel occlusion on computed tomography angiography.

PURPOSE: Artificial intelligence diagnosis and triage of large vessel occlusion may quicken clinical response for a subset of time-sensitive acute ischemic stroke patients, improving outcomes. Differences in architectural elements within data-driven convolutional neural network (CNN) models impact performance. Foreknowledge of effective model architectural elements for domain-specific problems can narrow the search for candidate models and inform strategic model design and adaptation to optimize performance on available data. Here, we study CNN architectures with a range of learnable parameters and which span the inclusion of architectural elements, such as parallel processing branches and residual connections with varying methods of recombining residual information. METHODS: We compare five CNNs: ResNet-50, DenseNet-121, EfficientNet-B0, PhiNet, and an Inception module-based network, on a computed tomography angiography large vessel occlusion detection task. The models were trained and preliminarily evaluated with 10-fold cross-validation on preprocessed scans (n = 240). An ablation study was performed on PhiNet due to superior cross-validated test performance across accuracy, precision, recall, specificity, and F1 score. The final evaluation of all models was performed on a withheld external validation set (n = 60) and these predictions were subsequently calibrated with sigmoid curves. RESULTS: Uncalibrated results on the withheld external validation set show that DenseNet-121 had the best average performance on accuracy, precision, recall, specificity, and F1 score. After calibration DenseNet-121 maintained superior performance on all metrics except recall. CONCLUSIONS: The number of learnable parameters in our five models and best-ablated PhiNet directly related to cross-validated test performance-the smaller the model the better. However, this pattern did not hold when looking at generalization on the withheld external validation set. DenseNet-121 generalized the best; we posit this was due to its heavy use of residual connections utilizing concatenation, which causes feature maps from earlier layers to be used deeper in the network, while aiding in gradient flow and regularization.