Objective
Although a clinical diagnosis, the standard initial imaging modality for patients with concern for pediatric community acquired pneumonia (pCAP) is a chest x‐ray (CXR), which has a ...relatively high false negative rate, exposes patients to ionizing radiation, and may not be available in resource limited settings. The primary objective of this meta‐analysis is to evaluate the accuracy of lung ultrasound (LUS) compared to CXR for the diagnosis of pCAP.
Methods
Data were collected via a systematic review of PubMed, EMBASE, and Web of Science with dates up to August 2017. Keywords and search terms were generated for pneumonia, lung ultrasound, and pediatric population. Two independent investigators screened s for inclusion. PRISMA was used for selecting appropriate studies. QUADAS was applied to these studies to assess quality for inclusion into the meta‐analysis. We collected data from included studies and calculated sensitivity, specificity, positive predictive value, and negative predictive values of CXR and LUS for the diagnosis of pCAP.
Results
Twelve studies including 1510 patients were selected for data extraction. LUS had a sensitivity of 95.5% (93.6‐97.1) and specificity of 95.3% (91.1‐98.3). CXR had a sensitivity of 86.8% (83.3‐90.0) and specificity of 98.2% (95.7‐99.6). Variations between the studies included ultrasound findings diagnostic of pneumonia, study setting (inpatient vs emergency department) and inclusion of CXR in the reference standard for pneumonia.
Conclusions
In our meta‐analysis, lung ultrasound had significantly better sensitivity with similar specificity when compared to chest x‐ray for the diagnosis of pediatric community acquired pneumonia.
To evaluate an artificial intelligence (AI)-assisted double reading system for detecting clinically relevant missed findings on routinely reported chest radiographs.
A retrospective study was ...performed in two institutions, a secondary care hospital and tertiary referral oncology centre. Commercially available AI software performed a comparative analysis of chest radiographs and radiologists' authorised reports using a deep learning and natural language processing algorithm, respectively. The AI-detected discrepant findings between images and reports were assessed for clinical relevance by an external radiologist, as part of the commercial service provided by the AI vendor. The selected missed findings were subsequently returned to the institution's radiologist for final review.
In total, 25,104 chest radiographs of 21,039 patients (mean age 61.1 years ± 16.2 SD; 10,436 men) were included. The AI software detected discrepancies between imaging and reports in 21.1% (5289 of 25,104). After review by the external radiologist, 0.9% (47 of 5289) of cases were deemed to contain clinically relevant missed findings. The institution's radiologists confirmed 35 of 47 missed findings (74.5%) as clinically relevant (0.1% of all cases). Missed findings consisted of lung nodules (71.4%, 25 of 35), pneumothoraces (17.1%, 6 of 35) and consolidations (11.4%, 4 of 35).
The AI-assisted double reading system was able to identify missed findings on chest radiographs after report authorisation. The approach required an external radiologist to review the AI-detected discrepancies. The number of clinically relevant missed findings by radiologists was very low.
The AI-assisted double reader workflow was shown to detect diagnostic errors and could be applied as a quality assurance tool. Although clinically relevant missed findings were rare, there is potential impact given the common use of chest radiography.
• A commercially available double reading system supported by artificial intelligence was evaluated to detect reporting errors in chest radiographs (n=25,104) from two institutions. • Clinically relevant missed findings were found in 0.1% of chest radiographs and consisted of unreported lung nodules, pneumothoraces and consolidations. • Applying AI software as a secondary reader after report authorisation can assist in reducing diagnostic errors without interrupting the radiologist's reading workflow. However, the number of AI-detected discrepancies was considerable and required review by a radiologist to assess their relevance.
Introduction : La pose des cathéters veineux centraux (CVC) est un geste fréquent en
réanimation. Elle n’est pas dénuée de complications dont le diagnostic reposait
longtemps sur la radiographie du ...thorax (RTX). Actuellement, l’échographie parait une
alternative intéressante.
Objectif : Rapporter l’impact de l’utilisation de l’échographie sur le délai
d’exclusion des complications mécaniques après pose des CVC.
Méthodes: Il s’agit d’une étude prospective, multicentrique, comparative et à double
aveugle. Ont été inclus les patients chez lesquels la pose d’un CVC a été décidée.
Après la pose, une RTX a été demandée et une échographie a été réalisée à la
recherche des signes de mauvais placement
et de pneumothorax. Les deux examens ont été interprétés par deux médecins
différents. Le critère de jugement principal entre le groupe échographie et le groupe
RTX était le temps « T1 » défini par le temps requis pour exclure les complications.
Résultats : 30 patients ont été inclus dans notre étude. La moyenne du temps T1
échographique était significativement inférieure à la moyenne
du temps T1 radiologique (p=0,000). Un seul cas de pneumothorax était observé. Il a
été détecté précocement par l’échographie. Pour les 29 autres malades, l’exclusion a
été échographique et radiologique. Par ailleurs, pas de complications de type mauvais
placement : l’exclusion était radiologique et échographique.
Conclusion : L’échographie est un outil plus rapide que la RTX dans l’exclusion des
complications mécaniques après pose des CVC. Elle garantit un examen aussi performant
que la RTX et non irradiant pour les patients de réanimation.
Background: Many invasive procedures are performed in the emergency room (ER), which have potential risks and complications. Due to limitations, especially with respect to size, portable X-ray ...devices are generally not used during such procedures. However, they have been miniaturized, enabling physicians to capture X-ray images by themselves..Methods: We developed a safe, compact, and lightweight X-ray unit and performed five invasive procedures in the ER. In all the procedures, a chest X-ray image was taken to confirm its utility.Results: Case 1 (central venous catheter placement): After needle and guidewire insertion and the placement of the catheter, the location of catheter could be confirmed. Case 2 (chest tube insertion): During the insertion of the chest tube into the pleural space, it was observed that the tip of the thoracic tube was at the appropriate location. Case 3 (percutaneous tracheostomy or cricothyroidotomy): After needle and guidewire insertion, it was visualized that the guidewire was in the right main bronchus and that the tube was inserted into the trachea. Case 4 (resuscitative endovascular aortic balloon of the aorta): The captured image revealed that the catheter was located in zone I before balloon inflation. Case 5 (Sengstaken–Blakemore tube): The image revealed that the balloon was located in the stomach.Conclusions: The devised portable X-ray unit could contribute medical safety during invasive procedures frequently performed in the ER.
•Two algorithms for detecting abnormal radiographic findings were proposed.•Findings included cardiomegaly, abnormal lung patterns, mediastinal shift, pleural effusion, and pneumothorax.•Both ...predictive models showed high performance in detection of abnormal findings.
This study evaluated the feasibility of bag-of-features (BOF) and convolutional neural networks (CNN) for computer-aided detection in distinguishing normal from abnormal radiographic findings. Computed thoracic radiographs of dogs were collected. For the purposes of this study, radiographic findings were used to distinguish between normal and abnormal in the following areas: (1) normal cardiac silhouette vs. cardiomegaly, (2) normal lung vs. abnormal lung patterns, (3) normal mediastinal position vs. mediastinal shift, (4) normal pleural space vs. pleural effusion, and (5) normal pleural space vs. pneumothorax. Images for training and testing the models consisted of ventrodorsal and lateral projection images of the same scale. The number of images used for each finding are as follow: 3142 for cardiomegaly (1571 normal and 1571 abnormal from 1143 dogs), 2086 for lung pattern (1043 normal and 1043 abnormal from 1247 dogs), 892 for mediastinal shift (446 normal and 446 abnormal from 387 dogs), 940 for pleural effusion (470 normal and 470 abnormal from 284 dogs), and 78 for pneumothorax (39 normal and 39 abnormal from 61 dogs). All data samples were divided so that 60% would be used for training the algorithms and 40% for testing the two models. The performance of the classifiers was evaluated by calculating the accuracy, sensitivity and specificity.
The accuracy of both models ranged from 79.6% to 96.9% in the testing set. CNN showed higher accuracy (CNN; 92.9–96.9% and BOF; 79.6–96.9%) and sensitivity (CNN; 92.1–100% and BOF; 74.1–94.8%) than BOF. In conclusion, both BOF and CNN have potential to be useful for improving work efficiency by double reading.
Objectives
To externally validate the performance of a commercial AI software program for interpreting CXRs in a large, consecutive, real-world cohort from primary healthcare centres.
Methods
A total ...of 3047 CXRs were collected from two primary healthcare centres, characterised by low disease prevalence, between January and December 2018. All CXRs were labelled as normal or abnormal according to CT findings. Four radiology residents read all CXRs twice with and without AI assistance. The performances of the AI and readers with and without AI assistance were measured in terms of area under the receiver operating characteristic curve (AUROC), sensitivity, and specificity.
Results
The prevalence of clinically significant lesions was 2.2% (68 of 3047). The AUROC, sensitivity, and specificity of the AI were 0.648 (95% confidence interval CI 0.630–0.665), 35.3% (CI, 24.7–47.8), and 94.2% (CI, 93.3–95.0), respectively. AI detected 12 of 41 pneumonia, 3 of 5 tuberculosis, and 9 of 22 tumours. AI-undetected lesions tended to be smaller than true-positive lesions. The readers’ AUROCs ranged from 0.534–0.676 without AI and 0.571–0.688 with AI (all
p
values < 0.05). For all readers, the mean reading time was 2.96–10.27 s longer with AI assistance (all
p
values < 0.05).
Conclusions
The performance of commercial AI in these high-volume, low-prevalence settings was poorer than expected, although it modestly boosted the performance of less-experienced readers. The technical prowess of AI demonstrated in experimental settings and approved by regulatory bodies may not directly translate to real-world practice, especially where the demand for AI assistance is highest.
Key Points
•
This study shows the limited applicability of commercial AI software for detecting abnormalities in CXRs in a health screening population.
•
When using AI software in a specific clinical setting that differs from the training setting, it is necessary to adjust the threshold or perform additional training with such data that reflects this environment well.
•
Prospective test accuracy studies, randomised controlled trials, or cohort studies are needed to examine AI software to be implemented in real clinical practice.
Objective
To investigate the feasibility of a deep learning–based detection (DLD) system for multiclass lesions on chest radiograph, in comparison with observers.
Methods
A total of 15,809 chest ...radiographs were collected from two tertiary hospitals (7204 normal and 8605 abnormal with nodule/mass, interstitial opacity, pleural effusion, or pneumothorax). Except for the test set (100 normal and 100 abnormal (nodule/mass, 70; interstitial opacity, 10; pleural effusion, 10; pneumothorax, 10)), radiographs were used to develop a DLD system for detecting multiclass lesions. The diagnostic performance of the developed model and that of nine observers with varying experiences were evaluated and compared using area under the receiver operating characteristic curve (AUROC), on a per-image basis, and jackknife alternative free-response receiver operating characteristic figure of merit (FOM) on a per-lesion basis. The false-positive fraction was also calculated.
Results
Compared with the group-averaged observations, the DLD system demonstrated significantly higher performances on image-wise normal/abnormal classification and lesion-wise detection with pattern classification (AUROC, 0.985 vs. 0.958;
p
= 0.001; FOM, 0.962 vs. 0.886;
p
< 0.001). In lesion-wise detection, the DLD system outperformed all nine observers. In the subgroup analysis, the DLD system exhibited consistently better performance for both nodule/mass (FOM, 0.913 vs. 0.847;
p
< 0.001) and the other three abnormal classes (FOM, 0.995 vs. 0.843;
p
< 0.001). The false-positive fraction of all abnormalities was 0.11 for the DLD system and 0.19 for the observers.
Conclusions
The DLD system showed the potential for detection of lesions and pattern classification on chest radiographs, performing normal/abnormal classifications and achieving high diagnostic performance.
Key Points
•
The DLD system was feasible for detection with pattern classification of multiclass lesions on chest radiograph.
•
The DLD system had high performance of image-wise classification as normal or abnormal chest radiographs (AUROC, 0.985) and showed especially high specificity (99.0%).
•
In lesion-wise detection of multiclass lesions, the DLD system outperformed all 9 observers (FOM, 0.962 vs. 0.886; p < 0.001).
To determine associations between antibody serologic tests and tracheobronchial lymphadenopathy (TBL) in dogs with pulmonary coccidioidomycosis and identify variables associated with time to ...resolution of TBL.
32 client owned dogs with newly diagnosed pulmonary coccidioidomycosis from October 2020 to February 2021.
Prospective cohort study. Thoracic radiographs and anti-Coccidioides spp antibody serology were performed at baseline and once every 3 months until remission or for a maximum of 12 months. Radiographic tracheobronchial lymph node height, length, and area were measured and recorded as ratios via comparison with the length of the T4 vertebral body (LT4) and length of the manubrium. Severity of TBL was also subjectively categorized as mild, moderate, or severe.
Tracheobronchial lymphadenopathy was identified in 81% (26/32; 95% CI, 64% to 93%) of dogs. There was no relevant association between TBL presence or severity and antibody serology results. Tracheobronchial lymphadenopathy resolved in 72% (n = 18) of dogs at the 3-month evaluation. The median time to resolution of TBL after initiation of fluconazole was 96 days (range, 72 to 386 days). Univariate analysis identified increasing TBL severity (hazard ratio, 0.40; 95% CI, 0.19 to 0.84; P = .02) and length:LT4 ratio (hazard ratio, 0.41; 95% CI, 0.20 to 0.82; P = .01) as variables associated with reduced probability of resolution of TBL.
Antibody serologic test results are not clinically useful to predict TBL presence or severity in dogs with pulmonary coccidioidomycosis, and larger tracheobronchial lymph nodes are more likely to take longer to resolve. Resolution of TBL occurs in most dogs within 3 to 6 months after fluconazole administration.
The goal of this review was to systematize the evidence on pulmonary ultrasound (PU) use in diagnosis, monitorization or hospital discharge criteria for patients with coronavirus disease 2019 ...(COVID-19). Evidence on the use of PU for diagnosis and monitorization of or as hospital discharge criteria for COVID-19 patients confirmed to have COVID-19 by reverse transcription polymerase chain reaction (RT-PCR) between December 1, 2019 and July 5, 2020 was compared with evidence obtained with thoracic radiography (TR), chest computed tomography (CT) and RT-PCR. The type of study, motives for use of PU, population, type of transducer and protocol, results of PU and quantitative or qualitative correlation with TR and/or chest CT and/or RT-PCR were evaluated. A total of 28 articles comprising 418 patients were involved. The average age was 50 y (standard deviation: 25.1 y), and there were 395 adults and 23 children. One hundred forty-three were women, 13 of whom were pregnant. The most frequent result was diffuse, coalescent and confluent B-lines. The plural line was irregular, interrupted or thickened. The presence of subpleural consolidation was noduliform, lobar or multilobar. There was good qualitative correlation between TR and chest CT and a quantitative correlation with chest CT of r = 0.65 (p < 0.001). Forty-four patients were evaluated only with PU. PU is a useful tool for diagnosis and monitorization and as criteria for hospital discharge for patients with COVID-19.