Search

To evaluate whether an imaging classifier for radiology practice can improve lung nodule classification and follow-up.A machine learning classifier was developed and trained using imaging data from the National Lung Screening Trial (NSLT) to produce a malignancy risk score (malignancy Similarity Index [mSI]) for individual lung nodules. In addition to NLST cohorts, external cohorts were developed from a tertiary referral lung cancer screening program data set and an external nonscreening data set of all nodules detected on CT. Performance of the mSI combined with Lung-RADS was compared with Lung-RADS alone and the Mayo and Brock risk calculators.We analyzed 963 subjects and 1,331 nodules across these cohorts. The mSI was comparable in accuracy (area under the curve = 0.89) to existing clinical risk models (area under the curve = 0.86-0.88) and independently predictive in the NLST cohort of 704 nodules. When compared with Lung-RADS, the mSI significantly increased sensitivity across all cohorts (25%-117%), with significant increases in specificity in the screening cohorts (17%-33%). When used in conjunction with Lung-RADS, use of mSI would result in earlier diagnoses and reduced follow-up across cohorts, including the potential for early diagnosis in 42% of malignant NLST nodules from prior-year CT scans.A computer-assisted diagnosis software improved risk classification from chest CTs of screening and incidentally detected lung nodules compared with Lung-RADS. mSI added predictive value independent of existing radiological and clinical variables. These results suggest the generalizability and potential clinical impact of a tool that is straightforward to implement in practice.

Telerobotic ultrasound technology allows radiologists and sonographers to remotely provide ultrasound services in underserved areas. This study aimed to compare costs associated with using telerobotic ultrasound to provide ultrasound services in rural and remote communities to costs associated with alternate models.A cost-minimization approach was used to compare four ultrasound service delivery models: telerobotic ultrasound (Model 1), telerobotic ultrasound and an itinerant sonographer (Model 2), itinerant sonographer without telerobotic ultrasound (Model 3), and travel to another community for all exams (Model 4). In Models 1-3, travel was assumed when exams could not be successfully performed telerobotically or by an itinerant sonographer. A publicly funded healthcare payer perspective was used for the reference case and a societal perspective was used for a secondary non-reference case. Costs were based on the literature and experience using telerobotic ultrasound in Saskatchewan, Canada. Costs were expressed in 2020 Canadian dollars.Average cost per ultrasound exam was $342, $323, $368, and $478 for Models 1, 2, 3, and 4, respectively, from a publicly funded healthcare payer perspective, and $461, $355, $447, and $849, respectively, from a societal perspective. In one-way sensitivity analyses, Model 2 was the lowest cost from a payer perspective for communities with population2075 people, distance350 km from the nearest ultrasound facility, or47% of the population eligible for publicly funded medical transportation.Health systems may wish to consider solutions such as telerobotic ultrasound and itinerant sonographers to reduce healthcare costs and improve access to ultrasound in rural and remote communities.

Most artificial intelligence (AI) studies have focused primarily on adult imaging, with less attention to the unique aspects of pediatric imaging. The objectives of this study were to (1) identify all publicly available pediatric datasets and determine their potential utility and limitations for pediatric AI studies and (2) systematically review the literature to assess the current state of AI in pediatric chest radiograph interpretation. We searched PubMed, Web of Science and Embase to retrieve all studies from 1990 to 2021 that assessed AI for pediatric chest radiograph interpretation and abstracted the datasets used to train and test AI algorithms, approaches and performance metrics. Of 29 publicly available chest radiograph datasets, 2 datasets included solely pediatric chest radiographs, and 7 datasets included pediatric and adult patients. We identified 55 articles that implemented an AI model to interpret pediatric chest radiographs or pediatric and adult chest radiographs. Classification of chest radiographs as pneumonia was the most common application of AI, evaluated in 65% of the studies. Although many studies report high diagnostic accuracy, most algorithms were not validated on external datasets. Most AI studies for pediatric chest radiograph interpretation have focused on a limited number of diseases, and progress is hindered by a lack of large-scale pediatric chest radiograph datasets.

Patients living in many rural and remote areas do not have readily available access to ultrasound services because of a lack of sonographers and radiologists in these communities. The objective of this study was to determine the feasibility of using telerobotic ultrasound to establish a service delivery model to remotely provide access to diagnostic ultrasound in rural and remote communities.Telerobotic ultrasound clinics were developed in three remote communities more than 500 km away from our academic medical center. Sonographers remotely performed all ultrasound examinations using telerobotic ultrasound systems, and examinations were subsequently interpreted by radiologists at an academic medical center. Diagnostic performance was assessed by each interpreting radiologist using a standardized reporting form. Patient experience was assessed through quantitative and qualitative analysis of survey responses. Operational challenges and solutions were identified.Eighty-seven telerobotic ultrasound examinations were remotely performed and included in this study, with the most frequent examination types being abdominal (n = 35), first-trimester obstetrical (n = 26), and second-trimester complete obstetrical (n = 12). Across all examination types, 70% of telerobotic ultrasound examinations were sufficient for diagnosis, minimizing travel or reducing wait times for these patients. Ninety-five percent of patients would be willing to have another telerobotic ultrasound examination in the future. Operational challenges were related to technical infrastructure, human resources, and coordination between clinic sites.Telerobotic ultrasound can provide access to diagnostic ultrasound services to underserved rural and remote communities without regular ultrasound services, thereby reducing disparities in access to care and improving health equity.

One of the critical topics in medical low-dose Computed Tomography (CT) imaging is how best to maintain image quality. As the quality of images decreases with lowering the Xray radiation dose, improving image quality is extremely important and challenging. We have proposed a novel approach to denoise low-dose CT images. Our algorithm learns directly from an end-to-end mapping from the low-dose Computed Tomography images for denoising the normal-dose CT images. Our method is based on a deep convolutional neural network with rectified linear units. By learning various low-level to high-level features from a low-dose image the proposed algorithm is capable of creating a high-quality denoised image. We demonstrate the superiority of our technique by comparing the results with two other state-ofthe-art methods in terms of the peak signal to noise ratio, root mean square error, and a structural similarity index.

Deep learning techniques have emerged in de-noising low-dose computed tomography (CT) images to avoid the potential health risks of high ionizing radiation dose on patients. Although these post-processing methods display high quality denoised images, the denoising performance still has the potential to improve. The primary purpose of this work was to determine and analyze the most effective and efficient hybrid loss function in deep learning (DL)-based denoising network. Objective functions in deep learning algorithms are the main keys for optimizing the parameters of a network and can affect the quality of the denoised image significantly. Hence, this work examined the various combinations of the most common objective functions in CT denoising networks, namely L1 loss, per-pixel loss, perceptual loss, and structural dissimilarity loss. Further, a hyperparameter learning algorithm was also introduced to find the best scalable factors of the loss functions in each hybrid loss function combination. For simplic-ity, RED-CNN was used in this study to easily demonstrate the performance of the losses during the denoising process. Based on this experiment, the balance between these loss function via the gradient-based optimization algorithm could help in the generalizability prediction of designing future CT denoising networks.

With the increasing concern regarding the radiation exposure of patients undergoing computed tomography (CT) scans, researchers have been using deep learning techniques to improve the quality of denoised low-dose CT (LDCT) images. In this paper, a cascaded dilated residual network (ResNet) with integrated attention modules, specifically spatial- and channel- attention modules, is proposed. This experiment demonstrated how these attention modules improved the denoised CT image by testing a simple ResNet with and without the modules. Further, an investigation regarding the effectiveness of per-pixel loss, perceptual loss via VGG16-Net, and structural dissimilarity loss functions is also covered through an ablation experiment. By knowing how these loss functions affect the output denoised images, a combination of the these loss function is then proposed which aims to prevent edge over-smoothing, enhance textural details and finally, preserve structural details on the denoised images. Finally, a bench testing was also done by comparing the visual and quantitative results of the proposed model with the state-of-the-art models such as block matching 3D (BM3D), patch-GAN and dilated convolution with edge detection layer (DRL-E-MP) for accuracy.

We develop and evaluate a deep learning algorithm to classify multiple catheters on neonatal chest and abdominal radiographs. A convolutional neural network (CNN) was trained using a dataset of 777 neonatal chest and abdominal radiographs, with a split of 81%–9%–10% for training-validation-testing, respectively. We employed ResNet-50 (a CNN), pre-trained on ImageNet. Ground truth labelling was limited to tagging each image to indicate the presence or absence of endotracheal tubes (ETTs), nasogastric tubes (NGTs), and umbilical arterial and venous catheters (UACs, UVCs). The dataset included 561 images containing two or more catheters, 167 images with only one, and 49 with none. Performance was measured with average precision (AP), calculated from the area under the precision-recall curve. On our test data, the algorithm achieved an overall AP (95% confidence interval) of 0.977 (0.679–0.999) for NGTs, 0.989 (0.751–1.000) for ETTs, 0.979 (0.873–0.997) for UACs, and 0.937 (0.785–0.984) for UVCs. Performance was similar for the set of 58 test images consisting of two or more catheters, with an AP of 0.975 (0.255–1.000) for NGTs, 0.997 (0.009–1.000) for ETTs, 0.981 (0.797–0.998) for UACs, and 0.937 (0.689–0.990) for UVCs. Our network thus achieves strong performance in the simultaneous detection of these four catheter types. Radiologists may use such an algorithm as a time-saving mechanism to automate reporting of catheters on radiographs. SUPPLEMENTARY INFORMATION: The online version contains supplementary material available at 10.1007/s10278-021-00473-y.

Objectives To develop a lung nodule management strategy combining the Lung CT Screening Reporting and Data System (Lung-RADS) with an artificial intelligence (AI) malignancy risk score and determine its impact on follow-up investigations and associated costs in a baseline lung cancer screening population. Materials and Methods Secondary analysis was undertaken of a data set consisting of AI malignancy risk scores and Lung-RADS classifications from six radiologists for 192 baseline low-dose CT studies. Low-dose CT studies were weighted to model a representative cohort of 3,197 baseline screening patients. An AI risk score threshold was defined to match average sensitivity of six radiologists applying Lung-RADS. Cases initially Lung-RADS category 1 or 2 with a high AI risk score were upgraded to category 3, and cases initially category 3 or higher with a low AI risk score were downgraded to category 2. Follow-up investigations resulting from Lung-RADS and the AI-informed management strategy were determined. Investigation costs were based on the 2019 US Medicare Physician Fee Schedule. Results The AI-informed management strategy achieved sensitivity and specificity of 91% and 96%, respectively. Average sensitivity and specificity of six radiologists using Lung-RADS only was 91% and 66%, respectively. Using the AI-informed management strategy, 41 (0.2%) category 1 or 2 classifications were upgraded to category 3, and 5,750 (30%) category 3 or higher classifications were downgraded to category 2. Minimum net cost savings using the AI-informed management strategy was estimated to be $72 per patient screened. Conclusion Using an AI risk score combined with Lung-RADS at baseline lung cancer screening may result in fewer follow-up investigations and substantial cost savings.

Low-dose computed tomography (LDCT) has yet to get its full potential benefit due to excessive quantum noise. Although learning to restore an image using two noisy images in the Noise2Noise (N2N) model has shown good promise for different noise models, it does not perform well in LDCT. In this article, we have introduced a collaborative technique to train multiple N2N generators simultaneously and learn the image representation from LDCT images. We have presented three models using this collaborative N2N (CN) principle employing CN two generators (CN2G), CN three generators (CN3G), and hybrid CN3G (HCN3G). The CN3G model has shown better performance than the CN2G model in terms of denoised image quality at the expense of an additional LDCT image. The HCN3G model has taken the advantages of both these models by managing to train three collaborative generators using only two LDCT images by leveraging one previous work called blind source separation (BSS) with block matching 3-D (BM3D). To make the collaboration among different generators more efficient, we have introduced collaborative loss terms among the generators. All three methods have shown improved performance in terms of peak signal-to-noise ratio and structural similarity index metrics compared to similar benchmark methods.

Introduction Obstetrical ultrasound imaging is critical in identifying at-risk pregnancies and informing clinical management. The coronavirus disease 2019 (COVID-19) pandemic has exacerbated challenges in accessing obstetrical ultrasound for patients in underserved rural and remote communities where this service is not available. This prospective descriptive study describes our experience of providing obstetrical ultrasound services remotely using a telerobotic ultrasound system in a northern Canadian community isolated due to a COVID-19 outbreak. Methods A telerobotic ultrasound system was used to perform obstetrical ultrasound exams remotely in La Loche, Canada, a remote community without regular access to obstetrical ultrasound. Using a telerobotic ultrasound system, a sonographer 605 km away remotely controlled an ultrasound probe and ultrasound settings. Twenty-one exams were performed in a five-week period during a COVID-19 outbreak in the community, including limited first-, second- and third-trimester exams ( n = 11) and complete second-trimester exams ( n = 10). Participants were invited to complete a survey at the end of the telerobotic ultrasound exam describing their experiences with telerobotic ultrasound. Radiologists subsequently interpreted all exams and determined the adequacy of the images for diagnosis. Results Of 11 limited obstetrical exams, radiologists indicated images were adequate in nine (81%) cases, adequate with some reservations in one (9%) case and inadequate in one (9%) case. Of 10 second-trimester complete obstetrical exams, radiologists indicated images were adequate in two (20%) cases, adequate with some reservations in three (30%) cases and inadequate in five (50%) cases. Second-trimester complete obstetrical exams were limited due to a combination of body habitus, foetal lie and telerobotic technology. Discussion A telerobotic ultrasound system may be used to answer focused clinical questions such as foetal viability, dating and foetal presentation in a timely manner while minimising patient travel to larger centres and potential exposure to severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), during the COVID-19 pandemic.

Artificial intelligence (AI) presents a key opportunity for radiologists to improve quality of care and enhance the value of radiology in patient care and population health. The potential opportunity of AI to aid in triage and interpretation of conventional radiographs (X-ray images) is particularly significant, as radiographs are the most common imaging examinations performed in most radiology departments. Substantial progress has been made in the past few years in the development of AI algorithms for analysis of chest and musculoskeletal (MSK) radiographs, with deep learning now the dominant approach for image analysis. Large public and proprietary image data sets have been compiled and have aided the development of AI algorithms for analysis of radiographs, many of which demonstrate accuracy equivalent to radiologists for specific, focused tasks. This article describes (1) the basis for the development of AI solutions for radiograph analysis, (2) current AI solutions to aid in the triage and interpretation of chest radiographs and MSK radiographs, (3) opportunities for AI to aid in noninterpretive tasks related to radiographs, and (4) considerations for radiology practices selecting AI solutions for radiograph analysis and integrating them into existing IT systems. Although comprehensive AI solutions across modalities have yet to be developed, institutions can begin to select and integrate focused solutions which increase efficiency, increase quality and patient safety, and add value for their patients.

Aim The objective of this survey was to understand the global trends of imaging assessments in persons with haemophilia, focusing on point-of-care ultrasound (POCUS). Insights into the barriers impeding its widespread proliferation as a frontline imaging modality were obtained. Methods The survey opened in September of 2017 and closed in May of 2018. Haemophilia Treatment Centres (HTCs) treating both paediatric/adult patients were the population of interest. A REDCap survey of 25 questions was disseminated to 232 clinical staff in 26 countries. Results The majority of respondents (88.3%, 91/103) reported that POCUS is most useful to confirm or rule out a presumed acute joint bleed. European HTCs reported the highest routine use of POCUS at 59.5% (22/37) followed by HTCs in the "Other" countries of the world at 46.7% (7/15) and North American HTCs at 43.9% (25/57). At the time of the survey, physiotherapists were identified as the clinical staff who perform POCUS 52.8% (28/53) of the time, in contrast with nurses/nurse practitioners who represent only 5.7% (3/53) of users. The greatest perceived barriers to the implementation of POCUS are the lack of trained healthcare professionals who can perform POCUS at 69.2% (74/107) and the overall time commitment required at 68.2% (73/107). Conclusion Despite POCUS being used in 49.5% (54/109) of sampled HTCs, it is still utilized almost 30% less globally than full diagnostic ultrasound. A list of barriers has been identified to inform HTCs which challenges they will likely need to overcome should they choose to incorporate this imaging modality into their practice.

X-ray computed tomography (CT) is a non-invasive medical diagnostic tool that has raised public concerns due to the associated health risks of radiation dose to patients. Reducing the radiation dose leads to noise artifacts, making the low-dose CT images unreliable for diagnosis. Hence, low-dose CT (LDCT) image reconstruction techniques have offered a new research area. In this study, a deep neural network is proposed, specifically a residual network (ResNet) using dilated convolution, batch normalization, and rectified linear unit (ReLU) layers with fused spatial- and channel-attention modules to enhance the quality of LDCT images. The network is optimized using the integration of per-pixel loss, perceptual loss via VGG16-net, and dissimilarity index loss. Through an ablation experiment, these functions show that they could effectively prevent edge oversmoothing, improve image texture, and preserve the structural details. Finally, comparative experiments showed that the qualitative and quantitative results of the proposed network outperform state-of-the-art denoising models such as block-matching 3D filtering (BM3D), Markovian-based patch generative adversarial network (patch-GAN), and dilated residual network with edge detection (DRL-E-MP).

A model was developed for long term metformin tissue retention based upon temporally inclusive models of serum/plasma concentration (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ C $$\end{document}C) having power function tails called the gamma-Pareto type I convolution (GPC) model and was contrasted with biexponential (E2) and noncompartmental (NC) metformin models. GPC models of \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ C $$\end{document}C have a peripheral venous first arrival of drug-times parameter, early \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ C $$\end{document}C peaks and very slow washouts of \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ C $$\end{document}C. The GPC, E2 and NC models were applied to a total of 148 serum samples drawn from 20 min to 72 h following bolus intravenous metformin in seven healthy mongrel dogs. The GPC model was used to calculate area under the curve (AUC), clearance (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ CL $$\end{document}CL), and functions of time, f(t), for drug mass remaining (M), apparent volume of distribution (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$V_{d}$$\end{document}Vd), as well as \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$t_{1/2}\ f(t)$$\end{document}t1/2f(t) for \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ C $$\end{document}C, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ M $$\end{document}M and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$V_{d}$$\end{document}Vd. The GPC models of \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ C $$\end{document}C yielded metformin \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ CL $$\end{document}CL-values that were 84.8% of total renal plasma flow (RPF) as estimated from meta-analysis. The GPC \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ CL $$\end{document}CL-values were significantly less than the corresponding NC and E2 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ CL $$\end{document}CL-values of 104.7% and 123.7% of RPF, respectively. The GPC plasma/serum only model predicted 78.9% drug \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ M $$\end{document}M average urinary recovery at 72 h; similar to prior human urine drug \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ M $$\end{document}M collection results. The GPC model \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$t_{1/2}$$\end{document}t1/2 of \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ M $$\end{document}M, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ C $$\end{document}C and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$V_d$$\end{document}Vd, were asymptotically proportional to elapsed time, with a constant limiting \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$t_{1/2}$$\end{document}t1/2 ratio of M/C averaging 7.0 times, a result in keeping with prior simultaneous \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ C $$\end{document}C and urine \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ M $$\end{document}M collection studies and exhibiting a rate of apparent volume growth of \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$V_d$$\end{document}Vd that achieved limiting constant values. A simulated constant average drug mass multidosing protocol exhibited increased \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$V_d$$\end{document}Vd and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$t_{1/2}$$\end{document}t1/2 with elapsing time, effects that have been observed experimentally during same-dose multidosing. The GPC heavy-tailed models explained multiple documented phenomena that were unexplained with lighter-tailed models. Electronic supplementary material The online version of this article (10.1007/s10928-019-09666-z) contains supplementary material, which is available to authorized users.

Image denoising of Low-dose computed tomography (LDCT) images has continues to receive attention in the research community due to ongoing concerns about high-dose radiation exposure of patients for diagnosis. The use of low radiation CT image, however, could lead to inaccurate diagnosis due to the presence of noise. Deep learning techniques are being integrated into denoising methods to address this problem. In this paper, a General Adversarial Network (GAN) composed of boosting fusion of spatial and channel attention modules is proposed. These modules are embedded in the denoiser to address the limitations of other GAN-based denoising models that tend to only focus on the local processing and neglect the dependencies of creating feature maps with spatial- and channel- wise image characteristics. This study aims to preserve structural details of LDCT images by applying boosting attention modules, prevents edge over-smoothing by integrating perceptual loss via VGG16 pre-trained network, and finally, improves the computational efficiency by taking advantage of deep learning techniques and GPU parallel computation.

CT machines can be tuned in order to reduce the radiation dose used for imaging, yet reducing the radiation dose results in noisy images which are not suitable in clinical practice. In order for low dose CT to be used effectively in practice this issue must be addressed. Generative Adversarial Networks (GAN) have been used widely in computer vision research and have proven themselves as a powerful tool for producing images with high perceptual quality. In this work we use a cascade of two neural networks, the first is a Generative Adversarial Network and the second is a Deep Convolutional Neural Network. The first network generates a denoised sample which is then fine-tuned by the second network via residue learning. We show that our cascaded method outperforms related works and more effectively reconstructs fine structural details in low contrast regions of the image.

Objective: Ultrasound is one of the most commonly used imaging modalities, though some populations face barriers in accessing ultrasound services, potentially resulting in disparities in utilization. The objective of this study was to assess the association between sociodemographic and geographic factors and non-obstetrical ultrasound utilization in the province of Saskatchewan, Canada. Methods: All non-obstetrical ultrasound exams performed from 2014 to 2018 in Saskatchewan, Canada were retrospectively identified from province-wide databases. Univariate and multivariate Poisson regression analyses were performed to assess the association between ultrasound utilization and sex, age, First Nations status, Charlson Comorbidity Index, urban vs. rural residence, geographic remoteness, and neighborhood income. Results: A total of 1,324,846 individuals (5,857,044 person-years) were included in the analysis. Female sex (adjusted incidence rate ratio [aIRR], 2.20; 95% confidence interval [CI], 2.19-2.22), age (aIRR, 4.97; 95% CI, 4.90-5.05 for ≥57 years vs. 10 vs. 0; 95% CI, 3.78-5.03), and higher neighborhood income (aIRR, 1.04; 95% CI, 1.02-1.05 for highest vs. lowest quintile) were associated with higher rates of ultrasound utilization. Individuals who were status First Nations (aIRR, 0.91; 95% CI, 0.90-0.92) or resided in geographically remote areas (aIRR, 0.87 for most vs. least remote; 95% CI, 0.83-0.91) had lower rates of ultrasound utilization. Individuals who lived in a rural area also had lower rates of ultrasound utilization (aIRR, 0.93; 95% CI, 0.92-0.94). Conclusion: Substantial disparities exist in non-obstetrical ultrasound utilization among individuals in low-income neighborhoods, status First Nations individuals, and individuals in rural and remote communities.

Purpose of the Program: Nîsohkamâtowak, the Cree word for Helping Each Other, is an initiative to close gaps in kidney health care for First Nations and Métis patients, their families, and communities in northern Saskatchewan. Nîsohkamâtowak emerged from a collaboration between the Kidney Health Community Program and First Nations and Métis Health Services to find ways to deliver better care and education to First Nations and Métis people living with kidney disease while acknowledging Truth and Reconciliation and the Calls to Action. Sources of Information: This article describes how traditional Indigenous protocols and storytelling were woven into the Nîsohkamâtowak events, gathering of patient and family voices in writing and video format, and how this work led to a collaborative co-designed process that incorporates the Truth and Reconciliation: Calls to Action into kidney care and the benefits we have seen so far. The teachings of the 4 Rs—respect, reciprocity, responsibility, and relevance, were critical to ensuring that Nîsohkamâtowak reports and learning were shared with participants and the communities represented in this initiative. Methods: Group discussions and sharing circles were facilitated in several locations throughout northern and central Saskatchewan. Main topics of discussion were traditional medicines, residential schools impact, community and peer supports for kidney disease patients, and cultural safety education for health care providers. Key Findings: The general themes selected for improvement were education, support within the local community, traditional practices and cultural competency, and delivery of services. To address these gaps in kidney care, the following objectives were co-created with First Nations and Métis patients, families, and communities for Kidney Health to provide culturally appropriate education and resources, to ensure appropriate follow-up support to include strengthening connections to communities and other health authorities, to incorporate traditional practices into program design, and to ensure appropriate service delivery across the spectrum of care with a focus on screening and referral, which is strongly linked to coordination of care with local health centers. Implications: As a result of this work, the Kidney Health Community Program restructured the delivery of services and continues to work with Nîsohkamâtowak advisors on safety initiatives and chronic kidney disease awareness, prevention, and management in their respective communities. The Truth and Reconciliation and Calls to Action are honored to close the gaps in kidney care. Limitations: Nîsohkamâtowak is a local Kidney Health initiative that has the good fortune of having dedicated funding and staff to carry out this work. The findings may be unique to the First Nations and Métis communities and people who shared their stories. Truth and Reconciliation is an ongoing commitment that must be nurtured. Although not part of this publication, the effects of COVID-19 have made it difficult to further advance the Calls to Action, with more limited staff resources and the inability to meet in person as in the past.

Introduction For persons with hemophilia, optimization of joint outcomes is an important unmet need. The aim of this initiative was to determine use of ultrasound in evaluating arthropathy in persons with hemophilia, and to move toward consensus among hemophilia care providers regarding the preferred ultrasound protocols for global adaptation. Methods A global survey of hemophilia treatment centers was conducted that focused on understanding how and why ultrasound was being used and endeavored to move toward consensus definitions of both point‐of‐care musculoskeletal ultrasound (POC‐MSKUS) and full diagnostic ultrasound, terminology to describe structures being assessed by ultrasound, and how these assessments should be interpreted. Next, an in‐person meeting of an international group of hemophilia health care professionals and patient representatives was held, with the objective of achieving consensus regarding the acquisition and interpretation of POC‐MSKUS and full diagnostic ultrasound for use in the assessment of musculoskeletal (MSK) pathologies in persons with hemophilia. Results The recommendations were that clear definitions of the types of ultrasound examinations should be adopted and that a standardized ultrasound scoring/measurement system should be developed, tested, and implemented. The scoring/measurement system should be tiered to allow for a range of complexity yet maintain the ability for comparison across levels. Conclusion Ultrasound is an evolving technology increasingly used for the assessment of MSK outcomes in persons with hemophilia. As adoption increases globally for clinical care and research, it will become increasingly important to establish clear guidelines for image acquisition, interpretation, and reporting to ensure accuracy, consistency, and comparability across groups.

Low-dose CT denoising is a challenging task that has been studied by many researchers. Some studies have used deep neural networks to improve the quality of low-dose CT images and achieved fruitful results. In this paper, we propose a deep neural network that uses dilated convolutions with different dilation rates instead of standard convolution helping to capture more contextual information in fewer layers. Also, we have employed residual learning by creating shortcut connections to transmit image information from the early layers to later ones. To further improve the performance of the network, we have introduced a non-trainable edge detection layer that extracts edges in horizontal, vertical, and diagonal directions. Finally, we demonstrate that optimizing the network by a combination of mean-square error loss and perceptual loss preserves many structural details in the CT image. This objective function does not suffer from over smoothing and blurring effects causing by per-pixel loss and grid-like artifacts resulting from perceptual loss. The experiments show that each modification to the network improves the outcome while changing the complexity of the network, minimally.

Computer-aided diagnosis (CAD) has already been widely used in medical image processing. We recently make another trial to implement convolutional neural network (CNN) on the classification of pulmonary nodules of thoracic CT images. The biggest challenge in medical image classification with the help of CNN is the difficulty of acquiring enough samples, and overfitting is a common problem when there are not enough images for training. Transfer learning has been verified as reasonable in dealing with such problems with an acceptable loss value. We use the classic LeNet-5 model to classify pulmonary nodules of thoracic CT images, including benign and malignant pulmonary nodules, and different malignancies of the malignant nodules. The CT images are obtained from Lung Image Database Consortium and Image Database Resource Initiative (LIDC-IDRI) where both pulmonary nodule scanning and nodule annotations are available. These images are labeled and stored in a medical images knowledge base (KB), which is designed and implemented in our previous work. We implement the 10-folder cross validation (CV) to testify the robustness of the classification model we trained. The result demonstrates that the transfer learning of the LeNet-5 is good for classifying pulmonary nodules of thoracic CT images, and the average values of Top-1 accuracy are 97.041% and 96.685% respectively. We believe that our work is beneficial and has potential for practical diagnosis of lung nodules.

PurposeTo develop and evaluate the reliability of an explicit set of parameters and criteria for simple bone cysts (SBCs) and evaluate the reliability of single versus serial chronological reading methodsMethodsRadiographic criteria were developed based on the literature and expert consensus. A single anteroposterior/lateral radiograph from 32 subjects with SBC were evaluated by three radiologists. A second reading was then conducted using revised criteria including a visual schematic. In the third reading the same images were assessed but radiologists had access to images from two additional time points. Inter-rater reliability was assessed after each reading using kappa (κ) and percentage agreement for categorical and binary parameters and intra-class correlation coefficient (ICC) for continuous parameters.ResultsParameters that were revised with more explicit definitions including the visual schematic demonstrated consistent or improved inter-rater reliability with the exception of continuous cortical rim present and cyst location in the metaphysis and mid-diaphysis. Cortical rim displayed only slight reliability throughout (κ= -0.008 to 0.16). All other categorical parameters had a percentage agreement above 0.8 or a moderate (κ= 0.41 to 0.60), substantial (κ = 0.61 to 0.80) or almost perfect inter-rater reliability (κ = 0.81 to 1.0) in at least one reading. All continuous parameters demonstrated excellent inter-rater reliability (ICC > 0.75) in at least one reading with the exception of scalloping (ICC = 0.37 to 0.70). Inter-rater reliability values did not indicate an obviously superior method of assessment between single and serial chronological readings.ConclusionExplicit criteria for SBC parameters used in their assessment demonstrated improved and substantial inter-rater reliability. Inter-rater reliability did not differ between single and serial chronological readings.Level of EvidenceNot Applicable

Obstetrical ultrasound imaging is an important part of prenatal care, though not all patients have readily available access to ultrasound services. This study aimed to assess the association between sociodemographic and geographic factors and (1) having a second trimester complete obstetrical ultrasound and (2) overall obstetrical ultrasound utilization.All pregnancies and obstetrical ultrasound exams billed from 2014-2018 in Saskatchewan, Canada were identified from province-wide databases. Generalized estimating equation (GEE) models with binomial and Poisson distributions were used to identify factors associated with having a second trimester ultrasound and overall obstetrical ultrasound utilization, respectively.80,536 pregnancies from 57,881 individuals were included. Of 57,186 pregnancies carried to ≥23 weeks, a second trimester ultrasound was performed in 50,180 (87.7%). Patients living in rural areas (adjusted odds ratio [aOR], 0.70; 95% confidence interval [CI], 0.63-0.77; p0.0001), remote areas (aOR, 0.35 for greatest vs. least remoteness level; 95% CI, 0.32-0.39; p0.0001), and status First Nations individuals (aOR, 0.50; 95% CI, 0.46-0.53; p0.0001) were less likely to have a second trimester ultrasound. Patients living in higher income neighbourhoods (aOR, 1.86 for highest vs. lowest quintile; 95% CI, 1.62-2.13; p0.0001) were more likely to have a second trimester ultrasound. GEE Poisson regression analysis demonstrated these same factors, except rural residence, were associated with overall obstetrical ultrasound utilization.Substantial disparities in obstetrical ultrasound utilization exist among patients in remote geographic areas, Indigenous peoples, and patients in low income neighbourhoods. Addressing barriers which these demographic groups face in accessing ultrasound imaging is critical to ensure health equity.

Ultrasound imaging is an essential component of healthcare services. This study sought to explore perceptions of access, and factors which shape access, to ultrasound imaging in two northern, remote, Indigenous communities in Canada. Using interpretive description as a methodological approach and a multi-dimensional conceptualisation of access to care as a theoretical framework, 15 semi-structured interviews were conducted in the northern Canadian communities of Stony Rapids and Black Lake, Saskatchewan. All participants had an obstetrical or non-obstetrical ultrasound exam performed in the past 10 years. Interviews were audio recorded and interview transcripts were analysed using constant comparative analysis. Geographic isolation from imaging facilities was a central barrier to participants accessing ultrasound imaging. Other barriers became apparent when participants had to travel for ultrasound, including fear of air travel, isolation from family, financial means, and unfamiliarity with larger cities. Barriers such as family and work responsibilities were exacerbated by the barrier of geography. Participants overcame these barriers as they were motivated by potential diagnostic benefits of ultrasound imaging. This study highlights disparities in access to ultrasound for northern, remote, Indigenous populations. Future efforts to improve access to imaging should consider barriers of distance to imaging facilities and strategies to bridge these barriers.

Lung cancer is a leading cause of cancer death in Canada, and accurate, early diagnosis are critical to improving clinical outcomes. Artificial Intelligence (AI)-based imaging analytics are a promising healthcare innovation that aim to improve the accuracy and efficiency of lung cancer diagnosis. Maximizing their clinical potential while mitigating their risks and limitations will require focused leadership informed by interdisciplinary expertise and system-wide insight. We convened a knowledge exchange workshop with diverse Saskatchewan health system leaders and stakeholders to explore issues surrounding the use of AI in diagnostic imaging for lung cancer, including implementation opportunities, challenges, and priorities. This technology is anticipated to improve patient outcomes, reduce unnecessary healthcare spending, and increase knowledge. However, health system leaders must also address the needs for robust data, financial investment, effective communication and collaboration between healthcare sectors, privacy and data protections, and continued interdisciplinary research to achieve this technology’s potential benefits.

In cardiac perfusion MRI, the low imaging speed limits the spatiotemporal resolution. To improve the imaging speed and hence the spatiotemporal resolution, Compressed sensing (CS) has been utilized. The purpose of this paper is to improve reconstruction of under sampled data based on k-space data acquisition using a Golden Radial trajectory, and then applying CS reconstruction and a nonrigid temporal registration algorithm. We propose a multiresolution scheme in which coarser reconstructions serve as initialization of the finer ones. An initial, low temporal resolution sequence is obtained by a regular CS reconstruction. From the quantitative evaluation, we determined that Signal to Error Ratio (SER) differences are within the range of 5-6 dB. Our experiments showed that the proposed algorithm can reduce motion artifacts and temporal blurring.

Low Dose CT Denoising is an open research problem which aims to reduce the risks of radiation exposure to patients. Recently researchers have used deep learning to denoise images with unknown noise distributions with promising results. However, approaches that use mean-squared-error (MSE) tend to over smooth the image resulting in loss of fine structural details in low contrast regions of the image. These regions are often crucial for diagnosis and must be preserved in order to maintain the diagnostic value of the image. In this work we show that using a new objective function which combines MSE, perceptual loss and structural dissimilarity (DSSIM) can effectively denoise low dose CT images while preserving fine structural details in low contrast regions. Further, we show that using a dilated residual network with fewer parameters outperforms a traditional deep convolutional neural network.

Access to sonographers and sonologists is limited in many communities around the world. Telerobotic sonography (robotic ultrasound) is a new technology to increase access to sonography, providing sonographers and sonologists the ability to manipulate an ultrasound probe from a distant location and remotely perform ultrasound examinations. This narrative review discusses the development of telerobotic ultrasound systems, clinical studies evaluating the feasibility and diagnostic accuracy of telerobotic sonography, and emerging use of telerobotic sonography in clinical settings. Telerobotic sonography provides an opportunity to provide real-time ultrasound examinations to underserviced rural and remote communities to increase equity in the delivery of diagnostic imaging.

Low Dose CT Denoising research aims to reduce the risks of radiation exposure to patients. Recently researchers have used deep learning to denoise low dose CT images with promising results. However, approaches that use mean-squared-error (MSE) tend to over smooth the image resulting in loss of fine structural details in low contrast regions of the image. These regions are often crucial for diagnosis and must be preserved in order for Low dose CT to be used effectively in practice. In this work we use a cascade of two neural networks, the first of which aims to reconstruct normal dose CT from low dose CT by minimizing perceptual loss, and the second which predicts the difference between the ground truth and prediction from the perceptual loss network. We show that our method outperforms related works and more effectively reconstructs fine structural details in low contrast regions of the image.

Ultrasound can penetrate deep into tissues and interact with human tissue via thermal and mechanical mechanisms. The ability to focus an ultrasound beam and its energy onto millimeter-size targets was a significant milestone in the development of therapeutic applications of focused ultrasound. Focused ultrasound can be used as a non-invasive thermal ablation technique for tumor treatment and is being developed as an option to standard oncologic therapies. High-intensity focused ultrasound has now been used for clinical treatment of a variety of solid malignant tumors, including those in the pancreas, liver, kidney, bone, prostate, and breast, as well as uterine fibroids and soft-tissue sarcomas. Magnetic resonance imaging and Ultrasound imaging can be combined with high intensity focused ultrasound to provide real-time imaging during ablation. Magnetic resonance guided focused ultrasound represents a novel non-invasive method of treatment that may play an important role as an alternative to open neurosurgical procedures for treatment of a number of brain disorders. This paper briefly reviews the underlying principles of HIFU and presents current applications, outcomes, and complications after treatment. Recent applications of Focused ultrasound for tumor treatment, drug delivery, vessel occlusion, histotripsy, movement disorders, and vascular, oncologic, and psychiatric applications are reviewed, along with clinical challenges and potential future clinical applications of HIFU.

Low-dose CT imaging is a valid approach to reduce patients’ exposure to X-ray radiation. However, reducing X-ray current increases noise and artifacts in the reconstructed CT images. Deep neural networks have been successfully employed to remove noise from low-dose CT images. This study proposes two novel techniques to boost the performance of a neural network with minimal change in the complexity. First, a non-trainable edge detection layer is proposed that extracts four edge maps from the input image. The layer improves quantitative metrics (PSNR and SSIM) and helps to predict a CT image with more precise boundaries. Next, a joint function of mean-square error and perceptual loss is employed to optimize the network. Using the perceptual loss helps to preserve structural detail; however, it adds check-board artifacts to the output. The proposed joint objective function takes advantage of the benefits offered by each loss. It improves the over-smoothing problem caused by mean-square error and the check-board artifacts caused by perceptual loss.

Decreasing the absorbed dosage by patient in x-ray imaging along with keeping image quality is one of the long-term goals of medical imaging field. Using low-dose images, instead of normal-dose images, can decrease the absorbed dosage; however, it also decreases the image quality due to quantum noise. In this paper, combination of Fast-ICA and adaptive Type-2 Fuzzy filter is utilized for filtering a group of low-dose images. Five different phantoms are used for investigating various effect of denoising, such as retaining slice geometry, high resolution, low-contrast, uniformity and bead geometry regions. Due to few numbers of images (8 images for each phantom), using deep learning method is not practical. The main novelty is attempting to convert the shot noise distribution to salt and pepper and denoising mapped image using fast independent component analysis. Concisely, the average and standard deviation of PSNR and SSIM of the proposed algorithm on five phantoms are 36.0 ± 2.7 dB and 0.83 ± 0.2, respectively, which shows a significant improvement comparing to the similar benchmark methods.