Single patch super-resolution of histopathology whole slide images: a comparative study.

Purpose: The latest generation of scanners can digitize histopathology glass slides for computerized image analysis. These images contain valuable information for diagnostic and prognostic purposes. Consequently, the availability of high digital magnifications like 20 × and 40 × is commonly expected in scanning the slides. Thus, the image acquisition typically generates gigapixel high-resolution images, times as large as 100,000 × 100,000 pixels . Naturally, the storage and processing of such huge files may be subject to severe computational bottlenecks. As a result, the need for techniques that can operate on lower magnification levels but produce results on par with outcomes for high magnification levels is becoming urgent. Approach: Over the past decade, the digital solution of enhancing images resolution has been addressed by the concept of super resolution (SR). In addition, deep learning has offered state-of-the-art results for increasing the image resolution after acquisition. In this study, multiple deep learning networks designed for image SR are trained and assessed for the histopathology domain. Results: We report quantitative and qualitative comparisons of the results using publicly available cancer images to shed light on the benefits and challenges of deep learning for extrapolating image resolution in histopathology. Three pathologists evaluated the results to assess the quality and diagnostic value of generated SR images. Conclusions: Pixel-level information, including structures and textures in histopathology images, are learnable by deep networks; hence improving the resolution quantity of scanned slides is possible by training appropriate networks. Different SR networks may perform best for various cancer sites and subtypes.

A Similarity Measure of Histopathology Images by Deep Embeddings.

Histopathology digital scans are large-size images that contain valuable information at the pixel level. Contentbased comparison of these images is a challenging task. This study proposes a content-based similarity measure for highresolution gigapixel histopathology images. The proposed similarity measure is an expansion of cosine vector similarity to a matrix. Each image is divided into same-size patches with a meaningful amount of information (i.e., contained enough tissue). The similarity is measured by the extraction of patchlevel deep embeddings of the last pooling layer of a pre-trained deep model at four different magnification levels, namely, 1x, 2.5x, 5x, and 10x magnifications. In addition, for faster measurement, embedding reduction is investigated. Finally, to assess the proposed method, an image search method is implemented. Results show that the similarity measure represents the slide labels with a maximum accuracy of 93.18% for top-5 search at 5x magnification.

Fine-Tuning and training of densenet for histopathology image representation using TCGA diagnostic slides.

Feature vectors provided by pre-trained deep artificial neural networks have become a dominant source for image representation in recent literature. Their contribution to the performance of image analysis can be improved through fine-tuning. As an ultimate solution, one might even train a deep network from scratch with the domain-relevant images, a highly desirable option which is generally impeded in pathology by lack of labeled images and the computational expense. In this study, we propose a new network, namely KimiaNet, that employs the topology of the DenseNet with four dense blocks, fine-tuned and trained with histopathology images in different configurations. We used more than 240,000 image patches with 1000×1000 pixels acquired at 20× magnification through our proposed "high-cellularity mosaic" approach to enable the usage of weak labels of 7126 whole slide images of formalin-fixed paraffin-embedded human pathology samples publicly available through The Cancer Genome Atlas (TCGA) repository. We tested KimiaNet using three public datasets, namely TCGA, endometrial cancer images, and colorectal cancer images by evaluating the performance of search and classification when corresponding features of different networks are used for image representation. As well, we designed and trained multiple convolutional batch-normalized ReLU (CBR) networks. The results show that KimiaNet provides superior results compared to the original DenseNet and smaller CBR networks when used as feature extractor to represent histopathology images.

The Effects of Spiritual Self-Care Training on Caregiving Strain in Mothers of Mentally Retarded Children.

INTRODUCTION: Care for a mentally retarded child induces a lot of problems for the mother and leads her to care giving strain and ignorning her self-care. Spiritual health will co-ordinate all aspects of human life and is necessary for coping with diseases in mother of mentally retarded children. AIM: To evaluate the effects of spiritual self-care training on care giving strain in mothers of mentally retarded children. MATERIALS AND METHODS: The present study, is a before and after type quasi-experimental research based on which 60 mothers of mentally retarded children who were hospitalized in Elahi Rehabilitation Center in Quchan City, were selected using convenience sampling and were randomly assigned to intervention and control groups. Data was collected by demographic characteristic questionnaire and care giving strain questionnaire that were filled by groups before, immediately and two weeks after spiritual self-care training. Data was analysed using SPSS version 20. RESULTS: According to the results, there was no significant difference between the mean score of care giving strain in intervention and control groups before and immediately after the intervention. However, among the members of the intervention group the score of mother care giving strain decreased an average of 87.21% within two weeks after the intervention, which was statistically significant over time (p=0.001). The score of mothers in the control group increased an average of 5% over time which was not statistically significant (p=0.4). The observed differences between these groups were also statistically significant even after controlling the effects of such intervening factors as marital status, children age and the years of caring for children (p=0.001). CONCLUSION: Spiritual self-care training can decrease care giving strain in mothers of mentally retarded children. Therefore, strengthening their spiritual beliefs and backgrounds, mothers can greatly reduce the strain caused by care giving problems of mentally retarded children.

Epidemiology of diabetes mellitus, pre-diabetes, undiagnosed and uncontrolled diabetes and its predictors in general population aged 15 to 75 years: A community-based study (KERCADRS) in southeastern Iran.

BACKGROUND: The goal of this research was to measure the age-sex standardized prevalence of pre-diabetes (pre-DM) and diabetes (DM), and the effectiveness of diabetes management (using HbA1C as the indicator) in an urban area in Iran. METHODS: Using a randomized cluster household survey, we recruited 5900 individuals whose age ranged from 15 to 75 from Kerman for assessing coronary artery disease risk factors (KERCADRS) including diabetes. In 2010 and 2011, all of the participants were interviewed by trained staff for medical history and physical activities, and were then examined for blood pressure and anthropometric measures. Venus blood sample was also collected for fasting plasma glucose and HbA1c. RESULTS: The age-sex standardized prevalence of pre-diabetes, diagnosed and undiagnosed was 18.7%, 6.3% and 2.7%, respectively. Diabetes increased by age (from 14.7% in the 15-24 years old group to 28.4% in the 65-75 years old group), particularly after 40 years. Occasional opium users had the highest prevalence of Pre-DM (34.6%). Seventy-nine percent of the depressed and 75.5% of the anxious participants with diagnosed-DM were identified as uncontrolled-DM. More than 60% of diagnosed diabetic cases had impaired HbA1c. Overweight and obesity (adjusted odds ratio (AOR) 1.6) and low physical activity (AOR 1.5) were the most preventable risk factors associated with diabetes. CONCLUSION: Considerable prevalence of diabetes, susceptibility in progressing to diabetes and uncontrolled diabetes among individuals living in Kerman, suggested ineffective prevention and treatment of diabetes in urban areas in Iran. Successful experience regarding primary health-care in rural areas should be expanded to urban settings.

The promotion of axon extension in vitro using polymer-templated fibrin scaffolds.

Biomaterial nerve cuffs are a clinical alternative to autografts and allografts as a means to repair segmental peripheral nerve defects. However, existing clinical biomaterial constructs lack true incorporation of physical guidance cues into their design. In both two- and three-dimensional systems, it is known that substrate geometry directly affects rates of axon migration. However, the ability to incorporate these cues into biomaterial scaffolds of sufficient porosity to promote robust nerve regeneration in three-dimensional systems is a challenge. We have developed fibrin constructs fabricated by a sacrificial templating approach, yielding scaffolds with multiple 10-250 µm diameter conduits depending on the diameter of the template fibers. The resulting scaffolds contained numerous, highly aligned conduits, had porosity of ∼ 80%, and showed mechanical properties comparable to native nerve (150-300 kPa Young's modulus). We studied the effects of the conduit diameters on the rate of axon migration through the scaffold to investigate if manipulation of this geometry could be used to ultimately promote more rapid bridging of the scaffold. All diameters studied led to axon migration, but in contrast to effects of fiber diameters in other systems, the rate of axon migration was independent of conduit diameter in these templated scaffolds. However, aligned conduits did support more rapid axon migration than non-aligned, tortuous controls.