scholarly journals Assessment of the effect of a comprehensive chest radiograph deep learning model on radiologist reports and patient outcomes: a real-world observational study

BMJ Open ◽  
2021 ◽  
Vol 11 (12) ◽  
pp. e052902
Author(s):  
Catherine M Jones ◽  
Luke Danaher ◽  
Michael R Milne ◽  
Cyril Tang ◽  
Jarrel Seah ◽  
...  
Keyword(s):  
Real World ◽  
Patient Management ◽  
Secondary Outcome ◽  
Complete Agreement ◽  
Imaging Features ◽  
Chest X Ray ◽  
Reporting Environment ◽  
Deep Learning Model ◽  
Imaging Recommendations ◽  
Level Of Agreement

ObjectivesArtificial intelligence (AI) algorithms have been developed to detect imaging features on chest X-ray (CXR) with a comprehensive AI model capable of detecting 124 CXR findings being recently developed. The aim of this study was to evaluate the real-world usefulness of the model as a diagnostic assistance device for radiologists.DesignThis prospective real-world multicentre study involved a group of radiologists using the model in their daily reporting workflow to report consecutive CXRs and recording their feedback on level of agreement with the model findings and whether this significantly affected their reporting.SettingThe study took place at radiology clinics and hospitals within a large radiology network in Australia between November and December 2020.ParticipantsEleven consultant diagnostic radiologists of varying levels of experience participated in this study.Primary and secondary outcome measuresProportion of CXR cases where use of the AI model led to significant material changes to the radiologist report, to patient management, or to imaging recommendations. Additionally, level of agreement between radiologists and the model findings, and radiologist attitudes towards the model were assessed.ResultsOf 2972 cases reviewed with the model, 92 cases (3.1%) had significant report changes, 43 cases (1.4%) had changed patient management and 29 cases (1.0%) had further imaging recommendations. In terms of agreement with the model, 2569 cases showed complete agreement (86.5%). 390 (13%) cases had one or more findings rejected by the radiologist. There were 16 findings across 13 cases (0.5%) deemed to be missed by the model. Nine out of 10 radiologists felt their accuracy was improved with the model and were more positive towards AI poststudy.ConclusionsUse of an AI model in a real-world reporting environment significantly improved radiologist reporting and showed good agreement with radiologists, highlighting the potential for AI diagnostic support to improve clinical practice.

scholarly journals Role of efficient testing and contact tracing in mitigating the COVID-19 pandemic: a network modelling study

BMJ Open ◽  
2021 ◽  
Vol 11 (7) ◽  
pp. e045886
Author(s):  
Yiying Hu ◽  
Jianying Guo ◽  
Guanqiao Li ◽  
Xi Lu ◽  
Xiang Li ◽  
...  
Keyword(s):  
Real World ◽  
Case Fatality ◽  
Contact Tracing ◽  
Secondary Outcome ◽  
Time Interval ◽  
Fatality Rate ◽  
Real World Data ◽  
Network Modelling ◽  
World Data ◽  
Quantitative Evidence

ObjectivesThis study quantified how the efficiency of testing and contact tracing impacts the spread of COVID-19. The average time interval between infection and quarantine, whether asymptomatic cases are tested or not, and initial delays to beginning a testing and tracing programme were investigated.SettingWe developed a novel individual-level network model, called CoTECT (Testing Efficiency and Contact Tracing model for COVID-19), using key parameters from recent studies to quantify the impacts of testing and tracing efficiency. The model distinguishes infection from confirmation by integrating a ‘T’ compartment, which represents infections confirmed by testing and quarantine. The compartments of presymptomatic (E), asymptomatic (I), symptomatic (Is), and death with (F) or without (f) test confirmation were also included in the model. Three scenarios were evaluated in a closed population of 3000 individuals to mimic community-level dynamics. Real-world data from four Nordic countries were also analysed.Primary and secondary outcome measuresSimulation result: total/peak daily infections and confirmed cases, total deaths (confirmed/unconfirmed by testing), fatalities and the case fatality rate. Real-world analysis: confirmed cases and deaths per million people.Results(1) Shortening the duration between Is and T from 12 to 4 days reduces infections by 85.2% and deaths by 88.8%. (2) Testing and tracing regardless of symptoms reduce infections by 35.7% and deaths by 46.2% compared with testing only symptomatic cases. (3) Reducing the delay to implementing a testing and tracing programme from 50 to 10 days reduces infections by 35.2% and deaths by 44.6%. These results were robust to sensitivity analysis. An analysis of real-world data showed that tests per case early in the pandemic are critical for reducing confirmed cases and the fatality rate.ConclusionsReducing testing delays will help to contain outbreaks. These results provide policymakers with quantitative evidence of efficiency as a critical value in developing testing and contact tracing strategies.

scholarly journals Accessory pathway analysis using a multimodal deep learning model

2021 ◽  
Vol 11 (1) ◽  
Author(s):  
Makoto Nishimori ◽  
Kunihiko Kiuchi ◽  
Kunihiro Nishimura ◽  
Kengo Kusano ◽  
Akihiro Yoshida ◽  
...  
Keyword(s):  
Deep Learning ◽  
Electrophysiological Study ◽  
Accessory Pathway ◽  
Learning Model ◽  
X Rays ◽  
X Ray ◽  
Wpw Syndrome ◽  
Chest X Ray ◽  
Deep Learning Model ◽  
Ecg Data

AbstractCardiac accessory pathways (APs) in Wolff–Parkinson–White (WPW) syndrome are conventionally diagnosed with decision tree algorithms; however, there are problems with clinical usage. We assessed the efficacy of the artificial intelligence model using electrocardiography (ECG) and chest X-rays to identify the location of APs. We retrospectively used ECG and chest X-rays to analyse 206 patients with WPW syndrome. Each AP location was defined by an electrophysiological study and divided into four classifications. We developed a deep learning model to classify AP locations and compared the accuracy with that of conventional algorithms. Moreover, 1519 chest X-ray samples from other datasets were used for prior learning, and the combined chest X-ray image and ECG data were put into the previous model to evaluate whether the accuracy improved. The convolutional neural network (CNN) model using ECG data was significantly more accurate than the conventional tree algorithm. In the multimodal model, which implemented input from the combined ECG and chest X-ray data, the accuracy was significantly improved. Deep learning with a combination of ECG and chest X-ray data could effectively identify the AP location, which may be a novel deep learning model for a multimodal model.

scholarly journals Machine learning applied on chest x-ray can aid in the diagnosis of COVID-19: a first experience from Lombardy, Italy

2021 ◽  
Vol 5 (1) ◽  
Author(s):  
Isabella Castiglioni ◽  
Davide Ippolito ◽  
Matteo Interlenghi ◽  
Caterina Beatrice Monti ◽  
Christian Salvatore ◽  
...  
Keyword(s):  
Deep Learning ◽  
Area Under The Curve ◽  
Clinical Information ◽  
Training Dataset ◽  
Independent Dataset ◽  
X Ray ◽  
Learning Classifier ◽  
Chest X Ray ◽  
Polymerase Chain ◽  
Deep Learning Model

Abstract Background We aimed to train and test a deep learning classifier to support the diagnosis of coronavirus disease 2019 (COVID-19) using chest x-ray (CXR) on a cohort of subjects from two hospitals in Lombardy, Italy. Methods We used for training and validation an ensemble of ten convolutional neural networks (CNNs) with mainly bedside CXRs of 250 COVID-19 and 250 non-COVID-19 subjects from two hospitals (Centres 1 and 2). We then tested such system on bedside CXRs of an independent group of 110 patients (74 COVID-19, 36 non-COVID-19) from one of the two hospitals. A retrospective reading was performed by two radiologists in the absence of any clinical information, with the aim to differentiate COVID-19 from non-COVID-19 patients. Real-time polymerase chain reaction served as the reference standard. Results At 10-fold cross-validation, our deep learning model classified COVID-19 and non-COVID-19 patients with 0.78 sensitivity (95% confidence interval [CI] 0.74–0.81), 0.82 specificity (95% CI 0.78–0.85), and 0.89 area under the curve (AUC) (95% CI 0.86–0.91). For the independent dataset, deep learning showed 0.80 sensitivity (95% CI 0.72–0.86) (59/74), 0.81 specificity (29/36) (95% CI 0.73–0.87), and 0.81 AUC (95% CI 0.73–0.87). Radiologists’ reading obtained 0.63 sensitivity (95% CI 0.52–0.74) and 0.78 specificity (95% CI 0.61–0.90) in Centre 1 and 0.64 sensitivity (95% CI 0.52–0.74) and 0.86 specificity (95% CI 0.71–0.95) in Centre 2. Conclusions This preliminary experience based on ten CNNs trained on a limited training dataset shows an interesting potential of deep learning for COVID-19 diagnosis. Such tool is in training with new CXRs to further increase its performance.

scholarly journals Imaging features and differential diagnoses of non-neoplastic diffuse mediastinal diseases

2020 ◽  
Vol 11 (1) ◽  
Author(s):  
Flavian Tabotta ◽  
Gilbert R. Ferretti ◽  
Helmut Prosch ◽  
Samia Boussouar ◽  
Anne-Laure Brun ◽  
...  
Keyword(s):  
Early Stage ◽  
Specific Treatment ◽  
Imaging Features ◽  
Mediastinal Diseases ◽  
Pictorial Review ◽  
Wide Range ◽  
Chest X Ray ◽  
Delayed Recognition ◽  
Patient’S History ◽  
Contiguous Spread

Abstract Acute or chronic non-neoplastic diffuse mediastinal diseases have multiple causes, degrees of severity, and a wide range of management. Some situations require emergency care while others do not need specific treatment. Although the diagnosis may be suspected on chest X-ray, it is mainly based on CT. A delayed recognition is not uncommonly observed. Some findings may prompt the radiologist to look for specific associated injuries or lesions. This pictorial review will successively describe the various non-neoplastic causes of diffuse mediastinal diseases with their typical findings and major differentials. First, pneumomediastinum that can be provoked by extra- or intra-thoracic triggers requires the knowledge of patient’s history or recent occurrences. Absence of any usual etiological factor should raise suspicion of cocaine inhalation in young individuals. Next, acute mediastinitis may be related to post-operative complications, esophageal perforation, or contiguous spread of odontogenic or retropharyngeal infections. The former diagnosis is not an easy task in the early stage, owing to the similarities of imaging findings with those of normal post-operative appearance during the first 2–3 weeks. Finally, fibrosing mediastinitis that is linked to an excessive fibrotic reaction in the mediastinum with variable compromise of mediastinal structures, in particular vascular and airway ones. Differential diagnosis includes tumoral and inflammatory infiltrations of the mediastinum.

scholarly journals Evolving Deep Learning Convolutional Neural Networks for Early COVID-19 Detection in Chest X-ray Images

Mathematics ◽  
2021 ◽  
Vol 9 (9) ◽  
pp. 1002
Author(s):  
Mohammad Khishe ◽  
Fabio Caraffini ◽  
Stefan Kuhn
Keyword(s):  
Deep Learning ◽  
Early Detection ◽  
Iterative Process ◽  
High Accuracy ◽  
X Ray ◽  
Starting Point ◽  
Reliable Model ◽  
Minimum Number ◽  
Chest X Ray ◽  
Deep Learning Model

This article proposes a framework that automatically designs classifiers for the early detection of COVID-19 from chest X-ray images. To do this, our approach repeatedly makes use of a heuristic for optimisation to efficiently find the best combination of the hyperparameters of a convolutional deep learning model. The framework starts with optimising a basic convolutional neural network which represents the starting point for the evolution process. Subsequently, at most two additional convolutional layers are added, at a time, to the previous convolutional structure as a result of a further optimisation phase. Each performed phase maximises the the accuracy of the system, thus requiring training and assessment of the new model, which gets gradually deeper, with relevant COVID-19 chest X-ray images. This iterative process ends when no improvement, in terms of accuracy, is recorded. Hence, the proposed method evolves the most performing network with the minimum number of convolutional layers. In this light, we simultaneously achieve high accuracy while minimising the presence of redundant layers to guarantee a fast but reliable model. Our results show that the proposed implementation of such a framework achieves accuracy up to 99.11%, thus being particularly suitable for the early detection of COVID-19.

Childhood Fever: Correlation of Diagnosis With Temperature Response to Acetaminophen

PEDIATRICS ◽  
1987 ◽  
Vol 80 (3) ◽  
pp. 315-318
Author(s):  
M. Douglas Baker ◽  
Patricia D. Fosarelli ◽  
Richard O. Carpenter
Keyword(s):  
Patient Management ◽  
Temperature Response ◽  
Pediatric Emergency ◽  
Bacterial Disease ◽  
X Ray ◽  
Chest X Ray ◽  
Antipyretic Treatment ◽  
Fever Response ◽  
Childhood Fever

Many people believe that temperature response to antipyretics in febrile children varies according to diagnosis. To evaluate the validity of this premise, we prospectively studied the temperature response to acetaminophen of febrile children who came to an urban pediatric emergency and walk-in facility. The study group consisted of 1,559 patients between the ages of 8 weeks and 6 years whose temperatures when seen were greater than 38.4°C and who had not received antipyretic treatment within the previous four hours. Acetaminophen (15 mg/kg) was administered to each child and repeat temperatures were taken one and two hours later. Patient management was unaffected by the study, and physicians were unaware of the repeat temperature measurements. Telephone follow-up was conducted with the parents of each child within five days of the initial visit. Children with cultures positive for bacterial disease or chest x-ray films positive for pneumonia had slightly greater one- and two-hour temperature decreases compared with children with other diagnoses. Although statistically significant, we do not consider these differences in response to be clinically useful. We conclude that fever response to acetaminophen is not a clinically useful indicator by which to differentiate the causes of febrile illnesses in young children.

scholarly journals Sustained Virologic Response Rates of Direct Acting Antivirals in HIV Coinfected Hepatitis C Patients and Its Effect on Liver Fibrosis

2020 ◽  
Vol 3 (2) ◽  
pp. 01-07
Author(s):  
Dora Lebron
Keyword(s):  
Hepatitis C ◽  
Real World ◽  
Sustained Virologic Response ◽  
Virologic Response ◽  
Hiv Care ◽  
Secondary Outcome ◽  
Direct Acting Antivirals ◽  
Fibrosis Regression ◽  
Direct Acting ◽  
End Stage

Background: Hepatitis C virus (HCV) is an important cause of chronic hepatitis with necroinflammation and fibrosis resulting in end stage liver disease and hepatocellular carcinoma. Direct acting antivirals (DAAs) are newer agents that directly interfere with the HCV lifecycle and result in high rates of sustained virologic response (SVR). We evaluated if treatment with DAAs in a real-world setting is as successful in HCV/HIV coinfected patients as it is in HCV monoinfected patients, and if some degree of fibrosis regression can be observed after completion of therapy in both groups. Methods: We retrospectively reviewed data from HCV monoinfected and HCV/HIV coinfected patients who received treatment from 2014-2016 at the East Carolina University Infectious Diseases clinic. The primary outcome was to compare completion and sustained virologic response (SVR) rate at either 12 or 24 weeks between HCV monoinfected patients and HCV/HIV coinfected patients. The secondary outcome was to assess regression of fibrosis at either 12 or 24 weeks after completion of therapy, defined as one METAVIR stage improvement in their FibroSure™, a noninvasive biochemical test to estimate the stage of fibrosis. Results: There were 41 patients in each group. Compared to the coinfected group, patient no show rate was higher in the monoinfected group (p=0.0346). In the HCV monoinfected group, 25 (93%) achieved either SVR 12 or 24. Two patients were non-compliant and had detectable viral load on evaluation at week 12. In the HCV/HIV coinfected group, 37 patients achieved SVR (p=0.0039). One patient in the coinfected group did not complete therapy but achieved SVR. In terms of fibrosis, 12/18 (67%) in the monoinfected group demonstrated improvement in at least 1 Metavir stage and 6/18 (33%) had no change. In the coinfected group, 8/16 (50%) patients demonstrated an improvement in FibroSure™ stage, 5/16 (31%) had no change, and 3/16 (19%) had worsening fibrosis according to FibroSure™ stage, (p=0.4867). Conclusions: In this small, real-world cohort, HCV/HIV coinfected patients treated with DAAs had higher completion and SVR rates than HCV monoinfected patients. Treatment failures in the monoinfected group were all linked to non-adherence, whereas, more coinfected patients achieved SVR, likely related to the fact that they were regularly engaged in routine HIV care. Fibrosis regression based on FibroSure™ was observed more in monoinfected patients than those with coinfection. Although not statistically significant, at least 50% of the patients in each group had regression of fibrosis.

scholarly journals Omalizumab Treatment for Severe Atopic Asthma in a Real World Montréal Cohort

2018 ◽  
Vol 16 (1) ◽  
Author(s):  
David Haile-Meskale ◽  
Ron Olivenstein ◽  
Toby Mcgovern ◽  
Cathy Fugere ◽  
James G. Martin
Keyword(s):  
Real World ◽  
Immunoglobulin E ◽  
Smoking Status ◽  
Elevated Serum ◽  
Corticosteroid Treatment ◽  
Post Treatment ◽  
Atopic Asthma ◽  
Secondary Outcome ◽  
Blood Eosinophilia ◽  
Pre Treatment

Background: Individuals with severe atopic asthma are poorly controlled with standard treatments, including corticosteroids. A humanized monoclonal antibody binding immunoglobulin E (IgE), omalizumab, is approved to treat patients that are managed poorly despite optimal therapy and that have elevated serum levels of IgE.  Objective: The purpose of this study was to determine omalizumab’s effectiveness in a real-world setting. The primary outcome was the number of exacerbations of asthma requiring oral corticosteroid treatment in the 2 year pre-treatment period compared to 2 years post-treatment. The secondary outcome was cumulative dose of prednisone used before and after treatment. Other outcomes that were measured included: reduction in maintenance therapy, change in spirometry (FEV1) data, the stratification of patient population based on smoking status, and average exacerbation number and prednisone use as a function of IgE level and blood eosinophilia count.   Methods: Patient data were retrieved (n=41) through the hospital records of patients treated at the Montreal Chest Institute of the McGill University Health Center. Data were gathered and analyzed for the 2 years before the treatment start date and compared to data 2 years after.  Results:There was a significant reduction in average exacerbation number from 6.4  pre-treatment to 3.2  post-treatment (p=0.003). There was also a reduction in cumulative prednisone use from 2504mg to 1423mg (p=0.04) following the institution of omalizumab treatment. There was no correlation between either the initial IgE levels and blood eosinophilia and the reduction in exacerbations  Conclusion: Omalizumab was effective in reducing exacerbation number and prednisone use for patients with severe refractory asthma.  

Sign in / Sign up

Export Citation Format

Share Document