Deep learning-based extended field of view computed tomography image reconstruction: influence of network design on image estimation outside the scan field of view

The problem of data truncation in Computed Tomography (CT) is caused by the missing data when the patient exceeds the Scan Field of View (SFOV) of a CT scanner. The reconstruction of a truncated scan produces severe truncation artifacts both inside and outside the SFOV. We have employed a deep learning-based approach to extend the field of view and suppress truncation artifacts. Thereby, our aim is to generate a good estimate of the real patient data and not to provide a perfect and diagnostic image even in regions beyond the SFOV of the CT scanner. This estimate could then be used as an input to higher order reconstruction algorithms [1]. To evaluate the influence of the network structure and layout on the results, three convolutional neural networks (CNNs), in particular a general CNN called ConvNet, an autoencoder, and the U-Net architecture have been investigated in this paper. Additionally, the impact of L1, L2, structural dissimilarity and perceptual loss functions on the neural network’s learning have been assessed and evaluated. The evaluation of data set comprising 12 truncated test patients demonstrated that the U-Net in combination with the structural dissimilarity loss showed the best performance in terms of image restoration in regions beyond the SFOV of the CT scanner. Moreover, this network produced the best mean absolute error, L1, L2, and structural dissimilarity evaluation measures on the test set compared to other applied networks. Therefore, it is possible to achieve truncation artifact removal using deep learning techniques.

An Integrated Deep Learning and Belief Rule Base Intelligent System to Predict Survival of COVID-19 Patient under Uncertainty

The novel Coronavirus-induced disease COVID-19 is the biggest threat to human health at the present time, and due to the transmission ability of this virus via its conveyor, it is spreading rapidly in almost every corner of the globe. The unification of medical and IT experts is required to bring this outbreak under control. In this research, an integration of both data and knowledge-driven approaches in a single framework is proposed to assess the survival probability of a COVID-19 patient. Several neural networks pre-trained models: Xception, InceptionResNetV2, and VGG Net, are trained on X-ray images of COVID-19 patients to distinguish between critical and non-critical patients. This prediction result, along with eight other significant risk factors associated with COVID-19 patients, is analyzed with a knowledge-driven belief rule-based expert system which forms a probability of survival for that particular patient. The reliability of the proposed integrated system has been tested by using real patient data and compared with expert opinion, where the performance of the system is found promising.

Longitudinal Interprofessional Family-Based Experience (LIFE): An Authentic Experiential Interprofessional Education Learning Framework

The Longitudinal Interprofessional Family-Based Experience (LIFE) was developed to address the need for longitudinal, experiential IPE opportunities that bring students together with real patient-family units with an intentional plan for multiple qualitative and quantitative evaluation measures. LIFE engaged 48 early learners from eight health science schools at a large midwestern university in ongoing team skill-based interactions coupled with real patient experiential learning over 11 weeks. Student teams were introduced and encouraged to apply the socio-ecological model (SEM) and social determinants of health (SDH) while collaboratively exploring the impact of the patient-family’s interface with the healthcare system and community during two consecutive patient-family interviews. A creative collaboration with the health system’s Office of Patient Experience, provided eight patients who had experienced chronic illness and treatment in the healthcare system, who engaged with the learners as both teachers as well as evaluators in this experience. LIFE is a framework model that has applicability and adaptability for designing, implementing, and sustaining experiential IPE. Initial summary data regarding outcomes for students are presented as well as considerations to increase accessible and sustainable authentic IPE experiences through untapped patient and community collaborations.

A simple cornea deformation model

In this paper, we present a cornea deformation model based on the idea of extending the ‘neutral axis’ model to two-dimensional deformations. Considering this simple model, assuming the corneal tissue to behave like a continuous, isotropic and non-compressible material, we are able to partially describe, e.g., the observed deviation in refractive power after lenticule extraction treatments. The model provides many input parameters of the patient and the treatment itself, leading to an individual compensation ansatz for different setups. The model is analyzed for a reasonable range of various parameters. A semi-quantitative comparison to real patient data is performed.

BurnoutWords - Detecting Burnout for a Clinical Setting

Burnout, a syndrome conceptualized as resulting from major workplace stress that has not been successfully managed, is a major problem of today's society, in particular in crisis times such as a global pandemic situation. Burnout detection is hard, because the symptoms often overlap with other diseases and syndromes. Typical clinical approaches are using inventories to assess burnout for their patients, even though free-text approaches are considered promising. In research of natural language processing (NLP) applied to mental health, often data from social media is used and not real patient data, which leads to some limitations for the application in clinical use cases. In this paper, we fill the gap and provide a dataset using extracts from interviews with burnout patients containing 216 records. We train a support vector machine (SVM) classifier to detect burnout in text snippets with an accuracy of around 80%, which is clearly higher than the random baseline of our setup. This provides the foundation for a next generation of clinical methods based on NLP.

Computer-Aided Diagnosis for Pneumoconiosis Staging Based on Multi-scale Feature Mapping

In this research, we explored a method of multi-scale feature mapping to pre-screen radiographs quickly and accurately in the aided diagnosis of pneumoconiosis staging. We utilized an open dataset and a self-collected dataset as research datasets. We proposed a multi-scale feature mapping model based on deep learning feature extraction technology for detecting pulmonary fibrosis and a discrimination method for pneumoconiosis staging. The diagnostic accuracy was evaluated using under the curve (AUC) of the receiver operating characteristic (ROC) curve. The AUC value of our model was 0.84, which showed the best performance compared with previous work on datasets. The diagnosis results indicated that our method was highly consistent with that of clinical experts on real patient. Furthermore, the AUC value obtained through categories I–IV on the testing dataset demonstrated that categories I (AUC = 0.86) and IV (AUC = 0.82) obtained the best performance and achieved the level of clinician categorization. Our research could be applied to the pre-screening and diagnosis of pneumoconiosis in the clinic.

The Effectiveness of Training on Auscultation of Heart with a Simulator of Cardiology in Medical Students

The purpose of this study was to determine whether the training module with a simulator of cardiology improves auscultation skills in medical students. Methods. Medical students of the third year after completing the module of the cardiovascular system of the discipline “Propaedeutics of internal diseases, passed a two-hour or four-hour training module in clinical auscultation with retesting immediately after the intervention and in the fourth year. The control group consisted of fourth-year medical students who had no intervention. Results. The diagnostic accuracy in two-hour training was 45.9% vs 35.3% in four-hour training p <.001. The use of a cardio simulator significantly increased the accurate detection of mitral regurgitation immediately after training on a simulator (more than 73%) p <.001. The next academic year, regression was observed in the diagnostic accuracy of mitral insufficiency in the intervention group after six months of observation by 4%. The auscultation skills of students at the bedside of real patients did not increase after training on a simulator: the accuracy of diagnosis of the auscultatory picture of the defect was equally low in the intervention group and the control group (35.0% vs 30.8%, p = 0.651). Conclusions. Two-hour training was more effective than four-hour training. After training on cardiac auscultation using a patient’s cardiological simulator, the accuracy rate was low in a situation close to the clinical conditions and a clinic on a real patient.

Enhancement-constrained acceleration: A robust reconstruction framework in breast DCE-MRI

In patients with dense breasts or at high risk of breast cancer, dynamic contrast enhanced MRI (DCE-MRI) is a highly sensitive diagnostic tool. However, its specificity is highly variable and sometimes low; quantitative measurements of contrast uptake parameters may improve specificity and mitigate this issue. To improve diagnostic accuracy, data need to be captured at high spatial and temporal resolution. While many methods exist to accelerate MRI temporal resolution, not all are optimized to capture breast DCE-MRI dynamics. We propose a novel, flexible, and powerful framework for the reconstruction of highly-undersampled DCE-MRI data: enhancement-constrained acceleration (ECA). Enhancement-constrained acceleration uses an assumption of smooth enhancement at small time-scale to estimate points of smooth enhancement curves in small time intervals at each voxel. This method is tested in silico with physiologically realistic virtual phantoms, simulating state-of-the-art ultrafast acquisitions at 3.5s temporal resolution reconstructed at 0.25s temporal resolution (demo code available here). Virtual phantoms were developed from real patient data and parametrized in continuous time with arterial input function (AIF) models and lesion enhancement functions. Enhancement-constrained acceleration was compared to standard ultrafast reconstruction in estimating the bolus arrival time and initial slope of enhancement from reconstructed images. We found that the ECA method reconstructed images at 0.25s temporal resolution with no significant loss in image fidelity, a 4x reduction in the error of bolus arrival time estimation in lesions (p < 0.01) and 11x error reduction in blood vessels (p < 0.01). Our results suggest that ECA is a powerful and versatile tool for breast DCE-MRI.

On network suppression of multidrug-resistant pathogen spread

In the paper we present a system of SIS type equations coupled by impulses at fixed times that describe the transfer of patients in the healthcare system represented by a graph of healthcare facilities and corresponding communities. The first aim for this considerations is to provide rigorous mathematical analysis of a general theoretical model, which is then used to model transmission of hospital acquired multidrug-resistant bacteria infections based on real patient hospital records provided by German insurance company – AOK Lower Saxony. Starting from the existence and the asymptotic behaviour, together with specification of parameter R, we propose sufficient conditions guaranteeing network suppression of infection. Furthermore, conditions derived analytically and proposed numerical procedure are used to indicate healthcare facilities that are most prone to the high prevalence bacteria spread in the healthcare system and to ensure the stability of disease-free steady state of the system.

Polycaprolactone Composite Micro/Nanofibrous Material as an Alternative to Restricted Access Media for Direct Extraction and Separation of Non-Steroidal Anti-Inflammatory Drugs from Human Serum Using Column-Switching Chromatography

Application of the poly-ɛ-caprolactone composite sorbent consisting of the micro- and nanometer fibers for the on-line extraction of non-steroidal anti-inflammatory drugs from a biological matrix has been introduced. A 100 μL human serum sample spiked with ketoprofen, naproxen, sodium diclofenac, and indomethacin was directly injected in the extraction cartridge filled with the poly-ɛ-caprolactone composite sorbent. This cartridge was coupled with a chromatographic instrument via a six-port switching valve allowing the analyte extraction and separation within a single analytical run. The 1.5 min long extraction step isolated the analytes from the proteinaceous matrix was followed by their 13 min HPLC separation using Ascentis Express RP-Amide (100 × 4.6 mm, 5 µm) column. The recovery of all analytes from human serum tested at three concentration levels ranged from 70.1% to 118.7%. The matrix calibrations were carried out in the range 50 to 20,000 ng mL−1 with correlation coefficients exceeding 0.996. The detection limit was 15 ng mL−1, and the limit of quantification corresponded to 50 ng mL−1. The developed method was validated and successfully applied for the sodium diclofenac determination in real patient serum. Our study confirmed the ability of the poly-ɛ-caprolactone composite sorbent to remove the proteins from the biological matrix, thus serving as an alternative to the application of restricted-access media.