scholarly journals Object Detection and Instance Segmentation in Chest X-rays for Tuberculosis Screening

2021 ◽  
Vol 3 (1) ◽  
pp. 1-24
Author(s):  
Terence Griffin ◽  
Yu Cao ◽  
Benyuan Liu ◽  
Maria J. Brunette ◽  
Xinzi Sun
Keyword(s):  
Object Detection ◽  
False Positive Rate ◽  
X Rays ◽  
Tuberculosis Screening ◽  
Screening Process ◽  
Middle Income ◽  
Middle Income Countries ◽  
Positive Rate ◽  
Object Level ◽  
Instance Segmentation

Tuberculosis (TB) is a highly contagious disease leading to the deaths of approximately 2 million people annually. TB primarily affects the lungs and is spread through the air when people cough, sneeze, or spit. Providing healthcare professionals with better information, at a faster pace, is essential for combating this disease, especially in Low and Middle Income Countries (LMICs) with resource-constrained health systems. In this paper we describe how using convolution neural networks (CNNs) with an object level annotated dataset of chest X-rays (CXRs) allows us to identify the location of pulmonary issues indicative of TB. We compare the performance of Faster R-nobreakdash-CNN, Mask R-nobreakdash-CNN, Cascade versions of each, and SOLOv2, demonstrating reasonable results with a small dataset. We present a method to reduce the false positive rate by comparing the location of a detected object with the known location of areas where the detected class is likely to occur in the lung. Our results show that object detection and instance segmentation of CXRs can be achieved with a dataset of high-quality, object level annotations, and could be used as part of an automated TB screening process. This work has the potential to improve the speed of TB diagnosis in LMICs, if properly integrated into the healthcare system and adapted to existing clinical workflows and local regulations.

scholarly journals Toward Model-Generated Household Listing in Low- and Middle-Income Countries Using Deep Learning

2018 ◽  
Vol 7 (11) ◽  
pp. 448 ◽  
Author(s):  
Robert Chew ◽  
Kasey Jones ◽  
Jennifer Unangst ◽  
James Cajka ◽  
Justine Allpress ◽  
...  
Keyword(s):  
Sustainable Development ◽  
Object Detection ◽  
Household Surveys ◽  
Building Structures ◽  
Middle Income ◽  
The Sustainable Development ◽  
Detection Model ◽  
Middle Income Countries ◽  
Development Goals ◽  
Low And Middle Income

While governments, researchers, and NGOs are exploring ways to leverage big data sources for sustainable development, household surveys are still a critical source of information for dozens of the 232 indicators for the Sustainable Development Goals (SDGs) in low- and middle-income countries (LMICs). Though some countries’ statistical agencies maintain databases of persons or households for sampling, conducting household surveys in LMICs is complicated due to incomplete, outdated, or inaccurate sampling frames. As a means to develop or update household listings in LMICs, this paper explores the use of machine learning models to detect and enumerate building structures directly from satellite imagery in the Kaduna state of Nigeria. Specifically, an object detection model was used to identify and locate buildings in satellite images. In the test set, the model attained a mean average precision (mAP) of 0.48 for detecting structures, with relatively higher values in areas with lower building density (mAP = 0.65). Furthermore, when model predictions were compared against recent household listings from fieldwork in Nigeria, the predictions showed high correlation with household coverage (Pearson = 0.70; Spearman = 0.81). With the need to produce comparable, scalable SDG indicators, this case study explores the feasibility and challenges of using object detection models to help develop timely enumerated household lists in LMICs.

Computed Tomography in Partial Carpal Arthrodesis

1997 ◽  
Vol 22 (5) ◽  
pp. 653-655
Author(s):  
J. M. SOLER-MINOVES ◽  
J. GONZALEZ-USTES ◽  
R. PÉREZ ◽  
M. GIFREU ◽  
A. M. GALLART
Keyword(s):  
Computed Tomography ◽  
False Positive ◽  
False Positive Rate ◽  
False Negative ◽  
Plain Radiography ◽  
X Rays ◽  
Positive Rate

We carried out X-rays and computed tomography in 59 wrists in patients who had previous surgical intercarpal fusions. 1.2 mm thick axial images were obtained perpendicular to the axis of the joint. CT showed whether or not the carpal fusions were united. Compared with CT, plain radiography yielded a 25% false negative and 6% false positive rate. We conclude that CT is more useful than plain X-rays for evaluating partial carpal arthrodesis.

The Impact of Cash Transfers: A Review of the Evidence from Low- and Middle-income Countries

2018 ◽  
Vol 48 (03) ◽  
pp. 569-594 ◽  
Author(s):  
FRANCESCA BASTAGLI ◽  
JESSICA HAGEN-ZANKER ◽  
LUKE HARMAN ◽  
VALENTINA BARCA ◽  
GEORGINA STURGE ◽  
...  
Keyword(s):  
Evidence Base ◽  
Cash Transfers ◽  
Cash Transfer ◽  
Future Research ◽  
Unintended Effects ◽  
Screening Process ◽  
Middle Income ◽  
Middle Income Countries ◽  
The Impact ◽  
Low And Middle Income

AbstractThis article presents the findings of a review of the impact of non-contributory cash transfers on individuals and households in low- and middle-income countries, covering the literature of 15 years, from 2000 to 2015. Based on evidence extracted from 165 studies, retrieved through a systematic search and screening process, this article discusses the impact of cash transfers on 35 indicators covering six outcome areas: monetary poverty; education; health and nutrition; savings, investment and production; work; and empowerment. For most of the studies, cash transfers contributed to progress in the selected indicators in the direction intended by policymakers. Despite variations in the size and strength of the underlying evidence base by outcome and indicator, this finding is consistent across all outcome areas. The article also investigates unintended effects of cash transfer receipt, such as potential reductions in adult work effort and increased fertility, finding limited evidence for such unintended effects. Finally, the article highlights gaps in the evidence base and areas which would benefit from additional future research.

scholarly journals Reducing Diagnostic Errors in Musculoskeletal Trauma by Reviewing Non-Admission Orthopaedic Referrals in the Next-Day Trauma Meeting

2007 ◽  
Vol 89 (7) ◽  
pp. 692-695 ◽  
Author(s):  
H Sharma ◽  
S Bhagat ◽  
WJ Gaine
Keyword(s):  
False Positive Rate ◽  
False Negative ◽  
Diagnostic Errors ◽  
District General Hospital ◽  
X Rays ◽  
Out Of Hours ◽  
Accident And Emergency ◽  
Number Of Patients ◽  
Positive Rate ◽  
Musculoskeletal Trauma

INTRODUCTION Diagnostic errors in orthopaedics are usually caused by missing a fracture or misreading radiographs. The aim of this study was to document the pick-up rate of the wrong diagnoses by reviewing X-rays and casualty notes in the next-day trauma meeting. PATIENTS AND METHODS The casualty notes and radiographs of 503 patients were prospectively reviewed in the daily trauma meeting between August 2002 and December 2002 in a district general hospital. The relevant data were collected and analysed by a single assessor. RESULTS The false positive rate for making an orthopaedic diagnosis was 12.6% (i.e.) diagnosing a fracture, when none existed). The false negative (missing) rate was 4%, while 2.4% incidental findings were missed, or at least not documented, after reading the X-rays. There were 7.8% wrong diagnoses made. The majority of the patients were seen by the senior house officers. CONCLUSIONS The medicolegal significance of false negative diagnosis is obviously greater. In a busy emergency department, where a large number of patients are seen, there is a greater risk. This study shows the importance in a small-to-medium sized accident and emergency unit as well, where there is no senior cover available out-of-hours for final radiological interpretation. A morning trauma meeting which covers reviewing admitted patients as well as non-admission orthopaedic referrals has an effective risk management solution to early detection of missed and wrong diagnoses.

scholarly journals Determinants of Clean Birthing Practices in Low- and Middle-Income Countries: a Scoping Review

2019 ◽  
Author(s):  
Joanna Esteves Mills ◽  
Erin Flynn ◽  
Oliver Cumming ◽  
Robert Dreibelbis
Keyword(s):  
Healthcare Facility ◽  
Perinatal Period ◽  
Screening Process ◽  
Middle Income ◽  
Middle Income Countries ◽  
Effective Interventions ◽  
Multi Stage ◽  
Behavioural Determinants ◽  
Birthing Practices ◽  
Low And Middle Income

Abstract Background Infection is a leading cause of maternal and newborn mortality in low- and middle-income countries (LMIC). Clean birthing practices are fundamental to infection prevention efforts, but these are inadequate in LMIC. This scoping study reviews the literature on studies that describe determinants of clean birthing practices of healthcare workers or mothers during the perinatal period in LMIC. Methods We reviewed literature published between January 2000 and February 2018 providing information on behaviour change interventions, behaviours or behavioural determinants during the perinatal period in LMIC. Following a multi-stage screening process, we extracted key data manually from studies. We mapped identified determinants according to the COM-B behavioural framework, which posits that behaviour is shaped by three categories of determinants – capability, opportunity and motivation. Results 78 studies were included: 47 observational studies and 31 studies evaluating an intervention. 51% had a household or community focus, 28% had a healthcare facility focus and 21% focused on both. We identified 31 determinants of clean birthing practices. Determinants related to clean birthing practices as a generalised set of behaviours featured in 50 studies; determinants related specifically to one or more of six predefined behaviours – commonly referred to as “the six cleans” – featured in 31 studies. Determinants of hand hygiene (n=13) and clean cord care (n=11) were most commonly reported. Reported determinants across all studies clustered around psychological capability (knowledge) and physical opportunity (access to resources). However, greater heterogeneity in reported behavioural determinants was found across studies investigating specific clean birthing practices compared to those studying clean birthing as a generalised set of behaviours. Conclusions Efforts to combine clean birthing practices into a single suite of behaviours – such as the “six cleans”– may simplify policy and advocacy efforts. However, each clean practice has a unique set of determinants and understanding what drives or hinders the adoption of these individual practices is critical to designing more effective interventions to improve hygiene behaviours and neonatal and maternal health outcomes in LMIC. Current understanding in this regard remains limited. More theory-grounded formative research is required to understand motivators and social influences across different contexts.

The Identification and Evaluation of Methodologies to Assess the Quality of mHealth apps in High, Low & Middle-Income Countries: a Rapid Review (Preprint)

2021 ◽  
Author(s):  
Fionn Woulfe ◽  
Philip Kayode Fadahunsi ◽  
Simon Smith ◽  
Griphin Baxter Chirambo ◽  
Emma Larsson ◽  
...  
Keyword(s):  
Quality Assessment ◽  
Healthcare Professionals ◽  
Rapid Review ◽  
Screening Process ◽  
Middle Income ◽  
Middle Income Countries ◽  
Low Middle Income Countries ◽  
Mhealth Apps ◽  
Low And Middle Income

BACKGROUND There has been a rapid growth in the availability and use of mobile health (mHealth) apps around the world in recent years. However, consensus regarding an accepted standard to assess the quality of such apps does not exist. Differing interpretations of quality add to this problem. Consequently, it has become increasingly difficult for healthcare professionals to distinguish apps of high quality from those of lower quality. This exposes both patients and healthcare professionals to unnecessary risk. Despite progress, limited understanding of contributions by those in low- and middle- income countries (LMIC) on this topic exists. As such, the applicability of quality assessment methodologies in LMIC settings remains unexplored. OBJECTIVE The objectives of this rapid review are to; 1) Identify current methodologies within the literature to assess the quality of mHealth apps. 2) Understand what aspects of quality these methodologies address. 3) Determine what input has been made by authors from LMICs. 4) Examine the applicability of such methodologies in low- and middle- income settings. METHODS The review is registered with Prospero (CRD42020205149). A search of PubMed, EMBASE, Web of Science and Scopus was performed for papers relating to mHealth app quality assessment methodologies, published in English between 2005 and the 28th of December, 2020. A thematic and descriptive analysis of methodologies and papers was performed. RESULTS Electronic database searches identified 841 papers. After the screening process, 53 papers remained for inclusion; 6 proposed novel methodologies which could be used to evaluate mHealth apps of diverse medical areas of interest; 8 proposed methodologies which could be used to assess apps concerned with a specific medical focus; 39 used methodologies developed by other published authors to evaluate the quality of various groups of mHealth apps. Authors of 3 papers were solely affiliated to institutes in LMICs. A further 8 papers had at least one co-author affiliated to an institute in a LMIC. CONCLUSIONS Quality assessment of mHealth apps is complex in nature and at times, subjective. Despite growing research on this topic, to date an all-encompassing, appropriate means for evaluating the quality of mHealth apps does not exist. There has been engagement with authors affiliated to institutes in LMICs, however limited consideration of current generic methodologies for application in a LMIC settings have been identified.

scholarly journals RDSNet: A New Deep Architecture forReciprocal Object Detection and Instance Segmentation

2020 ◽  
Vol 34 (07) ◽  
pp. 12208-12215 ◽  
Author(s):  
Shaoru Wang ◽  
Yongchao Gong ◽  
Junliang Xing ◽  
Lichao Huang ◽  
Chang Huang ◽  
...  
Keyword(s):  
Object Detection ◽  
Source Code ◽  
Localization Accuracy ◽  
Structure Information ◽  
Stream Structure ◽  
Deep Architecture ◽  
Effectiveness And Efficiency ◽  
Bounding Boxes ◽  
Object Level ◽  
Instance Segmentation

Object detection and instance segmentation are two fundamental computer vision tasks. They are closely correlated but their relations have not yet been fully explored in most previous work. This paper presents RDSNet, a novel deep architecture for reciprocal object detection and instance segmentation. To reciprocate these two tasks, we design a two-stream structure to learn features on both the object level (i.e., bounding boxes) and the pixel level (i.e., instance masks) jointly. Within this structure, information from the two streams is fused alternately, namely information on the object level introduces the awareness of instance and translation variance to the pixel level, and information on the pixel level refines the localization accuracy of objects on the object level in return. Specifically, a correlation module and a cropping module are proposed to yield instance masks, as well as a mask based boundary refinement module for more accurate bounding boxes. Extensive experimental analyses and comparisons on the COCO dataset demonstrate the effectiveness and efficiency of RDSNet. The source code is available at https://github.com/wangsr126/RDSNet.

scholarly journals Initial Experience in Developing AI Algorithms in Medical Imaging Based on Annotations Derived From an E-Learning Platform

2021 ◽  
Author(s):  
Maurice Henkel ◽  
Hanns-Christian Breit ◽  
Patricia Wiesner ◽  
Jakob Wasserthal ◽  
Victor Parmar ◽  
...  
Keyword(s):  
Medical Imaging ◽  
False Positive Rate ◽  
X Rays ◽  
Algorithm Development ◽  
The Public ◽  
Learning Platform ◽  
Testing Dataset ◽  
Positive Rate ◽  
E Learning ◽  
Localization Performance

Development of supervised AI algorithms requires a large amount of labeled images. Image labelling is both time-consuming and expensive. Therefore, we explored the value of e-learning derived annotations for AI algorithm development in medical imaging. Methods We have developed an e-learning platform that involves image-based single click labelling as part of the educational learning process. Ten radiology residents, as part of their residency training, trained the recognition of pneumothorax on 1161 chest X-rays in posterior-anterior projection. Using this data, multiple AI algorithms for detecting pneumothorax were developed. Classification and localization performance of the models was tested on an independent internal testing dataset and on the public NIH ChestX-ray14 dataset. Results The AI models F1 scores on the internal and the NIH dataset were 0.87 and 0.44, respectively. Sensitivity was 0.85 and 0.80 for classification and specificity 0.96 and 0.48 for classification. F1 scores were 0.72 and 0.66, sensitivity 0.72 and 0.72. False positive rate was 0.36 and 0.32 for localisation. Conclusion Our results demonstrated that e-learning derived annotations are a valuable data source for algorithm development. Further work is needed to include additional parameters such as user performance, consensus of diagnosis, and quality control in the development pipeline.

scholarly journals Economic Burden of Viral Acute Respiratory Infections in Upper-Middle-Income Countries: Protocol For A Systematic Review

2020 ◽  
Author(s):  
Cesar Ramos Rocha-Filho ◽  
Aline Pereira Rocha ◽  
Felipe Sebastiao de Assis Reis ◽  
Ana Carolina Pereira Nunes Pinto ◽  
Gabriel Sodre Ramalho ◽  
...  
Keyword(s):  
Systematic Review ◽  
Economic Burden ◽  
Burden Of Illness ◽  
Total Direct Cost ◽  
Economic Modelling ◽  
Screening Process ◽  
Cross Sectional ◽  
Population Total ◽  
Middle Income ◽  
Middle Income Countries

Objective: To synthesize the available data on the economic burden of Coronavirus Diseases 2019 (COVID-19), Severe Acute Respiratory Syndrome (SARS), Middle East Respiratory Syndrome (MERS), Influenza-Like Illness (ILI), Respiratory Syncytial Virus (RSV)-related Acute Respiratory Infection (ARI), and Parainfluenza Virus type III (PIV3)-related ARI in Upper-Middle-Income Countries (UMIC), highlighting its major causes and comparing direct costs among nations. Study design: Systematic review, following the recommendations proposed in the Cochrane Handbook, but with some adaptations from previous economic studies. Review question: Is there any economic cost of viral ARI in UMIC? Types of studies to be included: Partial economic evaluation, such as Cost-of-Illness (COI) studies and burden of illness/diseases, database analysis, observational reports (cross-sectional studies, and prospective and retrospective cohort), and economic modelling studies that discuss one of the viral ARI in UMIC. No year of publication filter or language limit will be applied. Search databases: MEDLINE, EMBASE, LILACS, CINAHL, EconLit, CRD Library, MedRxiv, and Research Square. Moreover, hand searches of the bibliographies of included studies and relevant reviews identified during the screening process will be undertaken to identify any additional relevant study for inclusion in our review. Synthesis of results: Qualitative analysis. We will focus on the overall economic burden of the diseases on health systems and population; total direct cost; the contribution of different cost components to the economic burden (e.g. pharmacological therapy, hospitalization); comparative assessments of costs analysis across geographical location and time horizon; and current research gaps. Moreover, we intend to identify, when presented, prevalence and incidence rates of each disease. PROSPERO registration number: CRD42020225757.

scholarly journals A Generative Approach Toward Precision Antimicrobial Peptide Design

2020 ◽  
Author(s):  
Jonathon B. Ferrell ◽  
Jacob M. Remington ◽  
Colin M. Van Oort ◽  
Mona Sharafi ◽  
Reem Aboushousha ◽  
...  
Keyword(s):  
False Positive Rate ◽  
Membrane Binding ◽  
Molecular Simulations ◽  
Sampling Technique ◽  
Cost Effective ◽  
Umbrella Sampling ◽  
Generative Model ◽  
Screening Process ◽  
Knowledge Based ◽  
Positive Rate

AbstractAntimicrobial peptides (AMPs) are peptides with promising applications for healthcare, veterinary, and agriculture industries. Despite prior success in AMP design using physics- or knowledge-based approaches, there is still a critical need to create new methodologies to design peptides with a low false positive rate and high AMP activity and selectivity. Toward this goal, we invented a cost-effective approach which utilizes a generative model to produce AMP-like sequences and molecular simulations to select peptides based on their structures and interactions. For a proof of concept, we curated a dataset that comprises 500,000 non-AMP peptide sequences and 8,000 labeled AMP sequences to train the generative model, which generated novel and diverse AMP candidates to potentially target a wide variety of microbes. Following a screening process to select peptides that are cationic and likely helical, we assessed 12 candidates by simulating their membrane-binding tendency to a lipid bilayer model. With the umbrella sampling technique, we determined the free energy change during transfer from the solution to the membrane environments for each peptide. Accordingly, we selected the six peptides with the best membrane-binding tendency, synthesized them, and characterized through spectroscopies and biological assays. Three novel peptides were validated with activity to inhibit bacterial growth. In aggregate, the combination of AMP generator and molecular simulations afford an enhanced accuracy in AMP design. Towards future precision AMP design, our methodology and results demonstrate the viability to design novel AMP-like peptides to target selected pathogens and mechanisms.

Sign in / Sign up

Export Citation Format

Share Document