Assessment of the comparative agreement between chest radiographs and CT scans in intensive care units

PURPOSE

Chest radiographs in critically ill patients can be difficult to interpret due to technical and clinical factors. We sought to determine the agreement of chest radiographs and CT scans, and the inter-observer variation of chest radiograph interpretation, in intensive care units (ICUs).

METHODS

Chest radiographs and corresponding thoracic computerised tomography (CT) scans (as reference standard) were collected from 45 ICU patients. All radiographs were analysed by 20 doctors (radiology consultants, radiology trainees, ICU consultants, ICU trainees) from 4 different centres, blinded to CT results. Specificity/sensitivity were determined for pleural effusion, lobar collapse and consolidation/atelectasis. Separately, Fleiss' kappa for multiple raters was used to determine inter-observer variation for chest radiographs.

RESULTS

The median sensitivity and specificity of chest radiographs for detecting abnormalities seen on CTs scans were 43.2% and 85.9% respectively. Diagnostic sensitivity for pleural effusion was significantly higher among radiology consultants but no specialty/experience distinctions were observed for specificity. Median inter-observer kappa coefficient among assessors was 0.295 ("fair").

CONCLUSIONS

Chest radiographs commonly miss important radiological features in critically ill patients. Inter-observer agreement in chest radiograph interpretation is only "fair". Consultant radiologists are least likely to miss thoracic radiological abnormalities. The consequences of misdiagnosis by chest radiographs remain to be determined.

Detection of pulmonary nodules in chest radiographs: novel cost function for effective network training with purely synthesized datasets

PURPOSE

Many large radiographic datasets of lung nodules are available, but the small and hard-to-detect nodules are rarely validated by computed tomography. Such difficult nodules are crucial for training nodule detection methods. This lack of difficult nodules for training can be addressed by artificial nodule synthesis algorithms, which can create artificially embedded nodules. This study aimed to develop and evaluate a novel cost function for training networks to detect such lesions. Embedding artificial lesions in healthy medical images is effective when positive cases are insufficient for network training. Although this approach provides both positive (lesion-embedded) images and the corresponding negative (lesion-free) images, no known methods effectively use these pairs for training. This paper presents a novel cost function for segmentation-based detection networks when positive-negative pairs are available.

METHODS

Based on the classic U-Net, new terms were added to the original Dice loss for reducing false positives and the contrastive learning of diseased regions in the image pairs. The experimental network was trained and evaluated, respectively, on 131,072 fully synthesized pairs of images simulating lung cancer and real chest X-ray images from the Japanese Society of Radiological Technology dataset.

RESULTS

The proposed method outperformed RetinaNet and a single-shot multibox detector. The sensitivities were 0.688 and 0.507 when the number of false positives per image was 0.2, respectively, with and without fine-tuning under the leave-one-case-out setting.

CONCLUSION

To our knowledge, this is the first study in which a method for detecting pulmonary nodules in chest X-ray images was evaluated on a real clinical dataset after being trained on fully synthesized images. The synthesized dataset is available at https://zenodo.org/records/10648433 .

Deep Learning-Based Localization and Detection of Malpositioned Nasogastric Tubes on Portable Supine Chest X-Rays in Intensive Care and Emergency Medicine: A Multi-center Retrospective Study

Malposition of a nasogastric tube (NGT) can lead to severe complications. We aimed to develop a computer-aided detection (CAD) system to localize NGTs and detect NGT malposition on portable chest X-rays (CXRs). A total of 7378 portable CXRs were retrospectively retrieved from two hospitals between 2015 and 2020. All CXRs were annotated with pixel-level labels for NGT localization and image-level labels for NGT presence and malposition. In the CAD system, DeepLabv3 + with backbone ResNeSt50 and DenseNet121 served as the model architecture for segmentation and classification models, respectively. The CAD system was tested on images from chronologically different datasets (National Taiwan University Hospital (National Taiwan University Hospital)-20), geographically different datasets (National Taiwan University Hospital-Yunlin Branch (YB)), and the public CLiP dataset. For the segmentation model, the Dice coefficients indicated accurate delineation of the NGT course (National Taiwan University Hospital-20: 0.665, 95% confidence interval (CI) 0.630-0.696; National Taiwan University Hospital-Yunlin Branch: 0.646, 95% CI 0.614-0.678). The distance between the predicted and ground-truth NGT tips suggested accurate tip localization (National Taiwan University Hospital-20: 1.64 cm, 95% CI 0.99-2.41; National Taiwan University Hospital-Yunlin Branch: 2.83 cm, 95% CI 1.94-3.76). For the classification model, NGT presence was detected with high accuracy (area under the receiver operating characteristic curve (AUC): National Taiwan University Hospital-20: 0.998, 95% CI 0.995-1.000; National Taiwan University Hospital-Yunlin Branch: 0.998, 95% CI 0.995-1.000; CLiP dataset: 0.991, 95% CI 0.990-0.992). The CAD system also detected NGT malposition with high accuracy (AUC: National Taiwan University Hospital-20: 0.964, 95% CI 0.917-1.000; National Taiwan University Hospital-Yunlin Branch: 0.991, 95% CI 0.970-1.000) and detected abnormal nasoenteric tube positions with favorable performance (AUC: 0.839, 95% CI 0.807-0.869). The CAD system accurately localized NGTs and detected NGT malposition, demonstrating excellent potential for external generalizability.

Diagnostic value of cardiothoracic ratio in patients with non-ischaemic cardiomyopathy: comparison to cardiovascular magnetic resonance imaging

PURPOSE

To determine the reliability of the cardiothoracic ratio (CTR) as a simple method to assess the cardiac size and function in patients with non-ischemic cardiomyopathy (NICM).

METHODS

In a sample of 91 patients (66 patients with diagnosed non-ischemic cardiomyopathy and 25 controls) we calculated the CTR on a posteroanterior chest radiograph and ventricular and atrial size based on accepted cardiovascular magnetic resonance (CMR) imaging values. Left and right ventricular ejection fraction was also calculated. The CTR and cardiac chamber size were compared between patients with NICM and healthy individuals. The distinction between normal and increased cardiac chamber size was made using published normal CMR reference values stratified by age and gender.

RESULTS

CTR values were higher in the NICM group (50.7±5.5 % Vs. 45.3±4.7 %, p<0.001). Likewise, LVEDVi, LV indexed mass, LA indexed volume, LA indexed area, and RA indexed area were higher, and LVEF and RVEF were lower in patients with non-ischemic cardiomyopathy (p < 0.05). In patients with non-ischemic cardiomyopathy, the greatest correlation between CTR and CMR values was with LVEDVi (ρ=0.4, p < 0.001), LA indexed volume (ρ=0.5, p < 0.001), LA indexed area (ρ=0.5, p < 0.001) and RA indexed area (ρ=0.4, p < 0.001). However, the correlation strength was only moderate.

CONCLUSION

Despite patients with NICM had higher CTR values than the control group, a substantial proportion of these patients showed normal CTRs (<50 %). This fact limits the usefulness of CTR to reliably predict NICM. Correlation between CTR and heart chamber dilation on CMR was only weak to moderate.

Asian Intensive Reader of Pneumoconiosis program: examination for certification during 2008-2020

This study examined physicians' participation and performance in the examinations administered by the Asian Intensive Reader of Pneumoconiosis (AIR Pneumo) program from 2008 to 2020 and compared radiograph readings of physicians who passed with those who failed the examinations. Demography of the participants, participation trends, pass/fail rates, and proficiency scores were summarized; differences in reading the radiographs for pneumoconiosis of physicians who passed the examinations and those who failed were evaluated. By December 2020, 555 physicians from 20 countries had taken certification examinations; the number of participants increased in recent years. Reported background specialty training and work experience varied widely. Passing rate and mean proficiency score for participants who passed were 83.4% and 77.6 ± 9.4 in certification, and 76.8% and 88.1 ± 4.5 in recertification examinations. Compared with physicians who passed the examinations, physicians who failed tended to classify test radiographs as positive for pneumoconiosis and read a higher profusion; they likely missed large opacities and pleural plaques and had a lower accuracy in recognizing the shape of small opacities. Findings suggest that physicians who failed the examination tend to over-diagnose radiographs as positive for pneumoconiosis with higher profusion and have difficulty in correctly identifying small opacity shape.

Utility of artificial intelligence for detection of pneumothorax on chest radiopgraphs done after transthoracic percutaneous transthoracic biopsy guided by computed tomography

OBJETIVE

To assess the ability of an artificial intelligence software to detect pneumothorax in chest radiographs done after percutaneous transthoracic biopsy.

MATERIAL AND METHODS

We included retrospectively in our study adult patients who underwent CT-guided percutaneous transthoracic biopsies from lung, pleural or mediastinal lesions from June 2019 to June 2020, and who had a follow-up chest radiograph after the procedure. These chest radiographs were read to search the presence of pneumothorax independently by an expert thoracic radiologist and a radiodiagnosis resident, whose unified lecture was defined as the gold standard, and the result of each radiograph after interpretation by the artificial intelligence software was documented for posterior comparison with the gold standard.

RESULTS

A total of 284 chest radiographs were included in the study and the incidence of pneumothorax was 14.4%. There were no discrepancies between the two readers' interpretation of any of the postbiopsy chest radiographs. The artificial intelligence software was able to detect 41/41 of the present pneumothorax, implying a sensitivity of 100% and a negative predictive value of 100%, with a specificity of 79.4% and a positive predictive value of 45%. The accuracy was 82.4%, indicating that there is a high probability that an individual will be adequately classified by the software. It has also been documented that the presence of Port-a-cath is the cause of 8 of the 50 of false positives by the software.

CONCLUSIONS

The software has detected 100% of cases of pneumothorax in the postbiopsy chest radiographs. A potential use of this software could be as a prioritisation tool, allowing radiologists not to read immediately (or even not to read) chest radiographs classified as non-pathological by the software, with the confidence that there are no pathological cases.

Deep Learning-based Diagnosis of Pulmonary Tuberculosis on Chest X-ray in the Emergency Department: A Retrospective Study

Prompt and correct detection of pulmonary tuberculosis (PTB) is critical in preventing its spread. We aimed to develop a deep learning-based algorithm for detecting PTB on chest X-ray (CXRs) in the emergency department. This retrospective study included 3498 CXRs acquired from the National Taiwan University Hospital (NTUH). The images were chronologically split into a training dataset, NTUH-1519 (images acquired during the years 2015 to 2019; n = 2144), and a testing dataset, NTUH-20 (images acquired during the year 2020; n = 1354). Public databases, including the NIH ChestX-ray14 dataset (model training; 112,120 images), Montgomery County (model testing; 138 images), and Shenzhen (model testing; 662 images), were also used in model development. EfficientNetV2 was the basic architecture of the algorithm. Images from ChestX-ray14 were employed for pseudo-labelling to perform semi-supervised learning. The algorithm demonstrated excellent performance in detecting PTB (area under the receiver operating characteristic curve [AUC] 0.878, 95% confidence interval [CI] 0.854-0.900) in NTUH-20. The algorithm showed significantly better performance in posterior-anterior (PA) CXR (AUC 0.940, 95% CI 0.912-0.965, p-value < 0.001) compared with anterior-posterior (AUC 0.782, 95% CI 0.644-0.897) or portable anterior-posterior (AUC 0.869, 95% CI 0.814-0.918) CXR. The algorithm accurately detected cases of bacteriologically confirmed PTB (AUC 0.854, 95% CI 0.823-0.883). Finally, the algorithm tested favourably in Montgomery County (AUC 0.838, 95% CI 0.765-0.904) and Shenzhen (AUC 0.806, 95% CI 0.771-0.839). A deep learning-based algorithm could detect PTB on CXR with excellent performance, which may help shorten the interval between detection and airborne isolation for patients with PTB.

Development of Pericardial Fat Count Images Using a Combination of Three Different Deep-Learning Models: Image Translation Model From Chest Radiograph Image to Projection Image of Three-Dimensional Computed Tomography

RATIONALE AND OBJECTIVES

Pericardial fat (PF)-the thoracic visceral fat surrounding the heart-promotes the development of coronary artery disease by inducing inflammation of the coronary arteries. To evaluate PF, we generated pericardial fat count images (PFCIs) from chest radiographs (CXRs) using a dedicated deep-learning model.

MATERIALS AND METHODS

We reviewed data of 269 consecutive patients who underwent coronary computed tomography (CT). We excluded patients with metal implants, pleural effusion, history of thoracic surgery, or malignancy. Thus, the data of 191 patients were used. We generated PFCIs from the projection of three-dimensional CT images, wherein fat accumulation was represented by a high pixel value. Three different deep-learning models, including CycleGAN were combined in the proposed method to generate PFCIs from CXRs. A single CycleGAN-based model was used to generate PFCIs from CXRs for comparison with the proposed method. To evaluate the image quality of the generated PFCIs, structural similarity index measure (SSIM), mean squared error (MSE), and mean absolute error (MAE) of (i) the PFCI generated using the proposed method and (ii) the PFCI generated using the single model were compared.

RESULTS

The mean SSIM, MSE, and MAE were 8.56 × 10-1, 1.28 × 10-2, and 3.57 × 10-2, respectively, for the proposed model, and 7.62 × 10-1, 1.98 × 10-2, and 5.04 × 10-2, respectively, for the single CycleGAN-based model.

CONCLUSION

PFCIs generated from CXRs with the proposed model showed better performance than those generated with the single model. The evaluation of PF without CT may be possible using the proposed method.

Performance of artificial intelligence-based software for the automatic detection of lung lesions on chest radiographs of patients with suspected lung cancer

PURPOSE

This study aimed to evaluate the performance of the commercially available artificial intelligence-based software CXR-AID for the automatic detection of pulmonary nodules on the chest radiographs of patients suspected of having lung cancer.

MATERIALS AND METHODS

This retrospective study included 399 patients with clinically suspected lung cancer who underwent CT and chest radiography within 1 month between June 2020 and May 2022. The candidate areas on chest radiographs identified by CXR-AID were categorized into target (properly detected areas) and non-target (improperly detected areas) areas. The non-target areas were further divided into non-target normal areas (false positives for normal structures) and non-target abnormal areas. The visibility score, characteristics and location of the nodules, presence of overlapping structures, and background lung score and presence of pulmonary disease were manually evaluated and compared between the nodules detected or undetected by CXR-AID. The probability indices calculated by CXR-AID were compared between the target and non-target areas.

RESULTS

Among the 450 nodules detected in 399 patients, 331 nodules detected in 313 patients were visible on chest radiographs during manual evaluation. CXR-AID detected 264 of these 331 nodules with a sensitivity of 0.80. The detection sensitivity increased significantly with the visibility score. No significant correlation was observed between the background lung score and sensitivity. The non-target area per image was 0.85, and the probability index of the non-target area was lower than that of the target area. The non-target normal area per image was 0.24. Larger and more solid nodules exhibited higher sensitivities, while nodules with overlapping structures demonstrated lower detection sensitivities.

CONCLUSION

The nodule detection sensitivity of CXR-AID on chest radiographs was 0.80, and the non-target and non-target normal areas per image were 0.85 and 0.24, respectively. Larger, solid nodules without overlapping structures were detected more readily by CXR-AID.

Frequency of abnormal findings on chest radiograph after positive PPD in children and adolescents in an urban setting in the United States

BACKGROUND

Chest radiographs (CXR) for tuberculosis (TB) screening in children are valuable in high-burden settings. However, less certain in low prevalence contexts. In the United States, positive PPD is sufficient to treat for "latent" TB, or TB infection in asymptomatic patients.

OBJECTIVE

We sought to determine frequency of abnormal CXR findings after a positive purified protein derivative (PPD) test at a tertiary pediatric center in the United States.

METHOD

A retrospective evaluation was conducted of patients (0-18 years) with a CXR after a positive PPD (e.g., known exposure, employment, migratory requirements or before immunosuppression) between 2011 and 2021. Clinical information, demographics, and reason for PPD were recorded from health record. CXRs were evaluated using initial report and by a pediatric radiologist with special interest in TB and 8 years of experience.

RESULT

Of 485 patients, median [interquartile range (IQR)] age 8.5[3.3-14.4], abnormal CXRs were described in 5 (1%). Most common reasons for PPD included: close contact with someone with TB or with high risk for TB. Most patients 373 (76.9%) received treatment for latent TB, and 111 (22.9%) no treatment. One patient (0.2%) received treatment for active disease. Radiographic findings included isolated lymphadenopathy (n = 2), consolidation (n = 1), pleural fluid/thickening (n = 1) and a patient with lymphadenopathy and a calcified nodule (n = 1).

CONCLUSION

In our experience, prevalence of chest radiographs findings for patients with positive PPD was very low. Moreover, no cases of severe disease were seen and those with abnormal findings would not merit treatment change under current WHO guidelines.

Comparison of chest radiograph and lung ultrasound in children with acute respiratory failure

PURPOSE

Chest x-ray (CXR) is the standard imaging used to evaluate children in acute respiratory distress and failure. Our objective was to compare the lung-imaging techniques of CXR and lung ultrasound (LUS) in the evaluation of children with acute respiratory failure (ARF) to quantify agreement and to determine which technique identified a higher frequency of pulmonary abnormalities.

METHODS

This was a secondary analysis of a prospective observational study evaluating the sensitivity and specificity of LUS in children with ARF from 12/2018 to 02/2020 completed at the University of Wisconsin-Madison (USA). Children > 37.0 weeks corrected gestational age and ≤ 18 years of age admitted to the PICU with ARF were evaluated with LUS. We compared CXR and LUS completed within 6 h of each other. Kappa statistics (k) adjusted for maximum attainable agreement (k/kmax) were used to quantify agreement between imaging techniques and descriptive statistics were used to describe the frequency of abnormalities.

RESULTS

Eighty-eight children had LUS completed, 32 with concomitant imaging completed within 6 h are included. There was fair agreement between LUS and CXR derived diagnoses with 58% agreement (k/kmax = 0.36). Evaluation of imaging patterns included: normal, 57% agreement (k = 0.032); interstitial pattern, 47% agreement (k = 0.003); and consolidation, 65% agreement (k = 0.29). CXR identified more imaging abnormalities than LUS.

CONCLUSIONS

There is fair agreement between CXR and LUS-derived diagnoses in children with ARF. Given this, clinicians should consider the benefits and limitations of specific imaging modalities when evaluating children with ARF. Additional studies are necessary to further define the role of LUS in pediatric ARF given the small sample size of our study.

Deep learning-based automatic detection for pulmonary nodules on chest radiographs: The relationship with background lung condition, nodule characteristics, and location

PURPOSE

Computer-aided diagnosis (CAD), which assists in the interpretation of chest radiographs, is becoming common. However, few studies have evaluated the benefits and pitfalls of CAD in the real world. This study aimed to evaluate the independent performance of commercially available deep learning-based automatic detection (DLAD) software, EIRL Chest X-ray Lung Nodule, in a cohort that included patients with background pulmonary abnormalities often encountered in clinical situations.

METHODS

Patients with clinically suspected lung cancer for whom chest radiography was performed within a month before or after CT scan between June 2020 and May 2022 in our institution were enrolled. The reference standard was created using a bounding box annotated by two radiologists with reference to the CT. The visibility score, characteristics, location of the pulmonary nodules, presence of overlapping structures or pulmonary disease, and background lung score were manually determined.

RESULTS

We included 388 patients. The DLAD software detected 222 of the 322 nodules visible on manual evaluation, with a sensitivity of 0.689 and a false-positive rate of 0.168. The detectability of the DLAD software was significantly lower for small and subsolid and nodules with overlapping structures. The visibility score and sensitivity of detection by the DLAD software were positively correlated. The relationship between the background lung score and detection by the DLAD software was unclear.

CONCLUSION

The standalone performance of DLAD in detecting pulmonary nodules exhibited a sensitivity of 0.689 and a false-positive rate of 0.168. Understanding the characteristics of DLAD is crucial when interpreting chest radiographs with the assistance of the DLAD.

Dual center validation of deep learning for automated multi-label segmentation of thoracic anatomy in bedside chest radiographs

BACKGROUND AND OBJECTIVES

Bedside chest radiographs (CXRs) are challenging to interpret but important for monitoring cardiothoracic disease and invasive therapy devices in critical care and emergency medicine. Taking surrounding anatomy into account is likely to improve the diagnostic accuracy of artificial intelligence and bring its performance closer to that of a radiologist. Therefore, we aimed to develop a deep convolutional neural network for efficient automatic anatomy segmentation of bedside CXRs.

METHODS

To improve the efficiency of the segmentation process, we introduced a "human-in-the-loop" segmentation workflow with an active learning approach, looking at five major anatomical structures in the chest (heart, lungs, mediastinum, trachea, and clavicles). This allowed us to decrease the time needed for segmentation by 32% and select the most complex cases to utilize human expert annotators efficiently. After annotation of 2,000 CXRs from different Level 1 medical centers at Charité - University Hospital Berlin, there was no relevant improvement in model performance, and the annotation process was stopped. A 5-layer U-ResNet was trained for 150 epochs using a combined soft Dice similarity coefficient (DSC) and cross-entropy as a loss function. DSC, Jaccard index (JI), Hausdorff distance (HD) in mm, and average symmetric surface distance (ASSD) in mm were used to assess model performance. External validation was performed using an independent external test dataset from Aachen University Hospital (n = 20).

RESULTS

The final training, validation, and testing dataset consisted of 1900/50/50 segmentation masks for each anatomical structure. Our model achieved a mean DSC/JI/HD/ASSD of 0.93/0.88/32.1/5.8 for the lung, 0.92/0.86/21.65/4.85 for the mediastinum, 0.91/0.84/11.83/1.35 for the clavicles, 0.9/0.85/9.6/2.19 for the trachea, and 0.88/0.8/31.74/8.73 for the heart. Validation using the external dataset showed an overall robust performance of our algorithm.

CONCLUSIONS

Using an efficient computer-aided segmentation method with active learning, our anatomy-based model achieves comparable performance to state-of-the-art approaches. Instead of only segmenting the non-overlapping portions of the organs, as previous studies did, a closer approximation to actual anatomy is achieved by segmenting along the natural anatomical borders. This novel anatomy approach could be useful for developing pathology models for accurate and quantifiable diagnosis.

A unilateral whiteout lung in child with multisystem inflammatory syndrome associated with COVID-19 due to SARS-CoV-2: one case report of a boy

BACKGROUND

Multisystem inflammatory syndrome in children (MIS-C) is a relatively new and rare complication of COVID-19. This complication seems to develop after the infection rather than during the acute phase of COVID-19. This report aims to describe a case of MIS-C in an 8-year-old Thai boy who presented with unilateral lung consolidation. Unilateral whiteout lung is not a common pediatric chest radiograph finding in MIS-C, but this is attributed to severe acute respiratory failure.

CASE PRESENTATION

An 8-year-old boy presented with persistent fever for seven days, right cervical lymphadenopathy, and dyspnea for 12 h. The clinical and biochemical findings were compatible with MIS-C. Radiographic features included total opacity of the right lung and CT chest found consolidation and ground-glass opacities of the right lung. He was treated with intravenous immunoglobulin and methylprednisolone, and he dramatically responded to the treatment. He was discharged home in good condition after 8 days of treatment.

CONCLUSION

Unilateral whiteout lung is not a common pediatric chest radiographic finding in MIS-C, but when it is encountered, a timely and accurate diagnosis is required to avoid delays and incorrect treatment. We describe a pediatric patient with unilateral lung consolidation from the inflammatory process.

Clinical and chest radiographic features of missed lung cancer and their association with patient outcomes

PURPOSE

To examine clinical and chest radiographic features of missed lung cancer (MLC) and explore their association with patient outcomes.

METHODS

We retrospectively reviewed chest radiographs obtained at least six months before lung cancer (LC) diagnosis in 95 patients to identify the first positive chest radiograph showing MLC. We assessed chest radiographic features of MLC and their association with patient outcomes.

RESULTS

Seventy-five (78.9%) patients (39 men, 36 women; mean age, 64.5 ± 10.5 years) had MLC. The median diagnostic delay was 31.3 months (6.6-128.0 months). The median MLC size was 16 mm (5-57 mm), and 54.7%, 68.0%, and 74.7% of MLC were in the left lung, the middle/lower zones, and the outer two-thirds of the lung, respectively. MLC exhibited a round/oval shape, partly/poorly defined margin, irregular/spiculated border, a density less than the aortic knob, and anatomical superimposition in 57.3%, 77.3%, 61.3%, 85.3%, and 88.0% of cases, respectively. Thirty-five (46.7%) patients had stage III + IV LC at diagnosis. Thirty-one (41.3%) patients died. MLC in the inner one-third of the lung, exhibiting a density equal to/greater than the aortic knob, or superimposed by midline structures was significantly associated with stage III + IV LC at diagnosis. The 3-year all-cause mortality significantly increased when MLC was in the upper zone, superimposed by pulmonary vessels, superimposed by pulmonary vessels plus ribs, or superimposed by pulmonary vessels plus in the inner one-third of the lung.

CONCLUSION

MLC with some radiographic features pertaining to their location, density, and superimposed structures was found to portend a worse outcome.

Quantification of pleural effusions by two-dimensional transthoracic echocardiography

PURPOSE

There is lack of validated methods for quantifying the size of pleural effusion from standard transthoracic (TTE) windows. The purpose of this study is to determine whether pleural effusion (Peff) measured from routine two-dimensional (2D) TTE views correlate with chest radiograph (CXR).

MATERIALS AND METHODS

We retrospectively identified all inpatients who underwent a TTE and CXR within 2 days in a large tertiary care center. Peff was measured on TTE from parasternal long axis (PLAX), apical four-chamber (A4C), and subcostal views and on CXR. Logistic regression models were used determine optimal cut points to predict moderate or greater Peff.

RESULTS

In 200 patients (mean age 69.3 ± 14.3 years, 49.5% female), we found statistically significant associations between Peff size assessed by all TTE views and CXR, with weak to moderate correlation (PLAX length: 0.21 (95% CI [0.05, 0.35]); PLAX depth: 0.21 (95% CI [0.05, 0.35]); A4C left: 0.31 (95% CI [0.13, 0.46]); A4C right: 0.39 (95% CI [0.17, 0.57]); subcostal: 0.38 (95% CI [0.07, 0.61]). The best TTE thresholds for predicting moderate or greater left-sided Peff on CXR was PLAX length left >  = 8.6 cm (sensitivity 78%, specificity 54%, PPV 26%, and NPV 92%). The best TTE thresholds for predicting moderate or greater right-sided Peff on CXR was A4C right >  = 2.6 cm (sensitivity 87%, specificity 60%, PPV 37%, and NPV 94%).

CONCLUSIONS

We identified statistically significant associations with Peff size measured on TTE and CXR. The predictive ability of TTE to identify moderate or large pleural effusion is limited.

Utility of chest radiograph severity scoring in emergency department for predicting outcomes in COVID-19: A study of 1275 patients

OBJECTIVE

To measure the reliability and reproducibility of a chest radiograph severity score (CSS) in prognosticating patient's severity of disease and outcomes at the time of disease presentation in the emergency department (ED) with coronavirus disease 2019 (COVID-19).

MATERIALS AND METHODS

We retrospectively studied 1275 consecutive RT-PCR confirmed COVID-19 adult patients presenting to ED from March 2020 through June 2020. Chest radiograph severity score was assessed for each patient by two blinded radiologists. Clinical and laboratory parameters were collected. The rate of admission to intensive care unit, mechanical ventilation or death up to 60 days after the baseline chest radiograph were collected. Primary outcome was defined as occurrence of ICU admission or death. Multivariate logistic regression was performed to evaluate the relationship between clinical parameters, chest radiograph severity score, and primary outcome.

RESULTS

CSS of 3 or more was associated with ICU admission (78 % sensitivity; 73.1 % specificity; area under curve 0.81). CSS and pre-existing diabetes were independent predictors of primary outcome (odds ratio, 7; 95 % CI: 3.87, 11.73; p < 0.001 & odds ratio, 2; 95 % CI: 1-3.4, p 0.02 respectively). No significant difference in primary outcome was observed for those with history of hypertension, asthma, chronic kidney disease or coronary artery disease.

CONCLUSION

Semi-quantitative assessment of CSS at the time of disease presentation in the ED predicted outcomes in adults of all age with COVID-19.

Chest X-ray-based opportunistic screening of sarcopenia using deep learning

BACKGROUND

Early detection and management of sarcopenia is of clinical importance. We aimed to develop a chest X-ray-based deep learning model to predict presence of sarcopenia.

METHODS

Data of participants who visited osteoporosis clinic at Severance Hospital, Seoul, South Korea, between January 2020 and June 2021 were used as derivation cohort as split to train, validation and test set (65:15:20). A community-based older adults cohort (KURE) was used as external test set. Sarcopenia was defined based on Asian Working Group 2019 guideline. A deep learning model was trained to predict appendicular lean mass (ALM), handgrip strength (HGS) and chair rise test performance from chest X-ray images; then the machine learning model (SARC-CXR score) was built using the age, sex, body mass index and chest X-ray predicted muscle parameters along with estimation uncertainty values.

RESULTS

Mean age of the derivation cohort (n = 926; women n = 700, 76%; sarcopenia n = 141, 15%) and the external test (n = 149; women n = 95, 64%; sarcopenia n = 18, 12%) cohort was 61.4 and 71.6 years, respectively. In the internal test set (a hold-out set, n = 189, from the derivation cohort) and the external test set (n = 149), the concordance correlation coefficient for ALM prediction was 0.80 and 0.76, with an average difference of 0.18 ± 2.71 and 0.21 ± 2.28, respectively. Gradient-weight class activation mapping for deep neural network models to predict ALM and HGS commonly showed highly weight pixel values at bilateral lung fields and part of the cardiac contour. SARC-CXR score showed good discriminatory performance for sarcopenia in both internal test set [area under the receiver-operating characteristics curve (AUROC) 0.813, area under the precision-recall curve (AUPRC) 0.380, sensitivity 0.844, specificity 0.739, F1-score 0.540] and external test set (AUROC 0.780, AUPRC 0.440, sensitivity 0.611, specificity 0.855, F1-score 0.458). Among SARC-CXR model features, predicted low ALM from chest X-ray was the most important predictor of sarcopenia based on SHapley Additive exPlanations values. Higher estimation uncertainty of HGS contributed to elevate the predicted risk of sarcopenia. In internal test set, SARC-CXR score showed better discriminatory performance than SARC-F score (AUROC 0.813 vs. 0.691, P = 0.029).

CONCLUSIONS

Chest X-ray-based deep leaning model improved detection of sarcopenia, which merits further investigation.

Image augmentation and automated measurement of endotracheal-tube-to-carina distance on chest radiographs in intensive care unit using a deep learning model with external validation

BACKGROUND

Chest radiographs are routinely performed in intensive care unit (ICU) to confirm the correct position of an endotracheal tube (ETT) relative to the carina. However, their interpretation is often challenging and requires substantial time and expertise. The aim of this study was to propose an externally validated deep learning model with uncertainty quantification and image segmentation for the automated assessment of ETT placement on ICU chest radiographs.

METHODS

The CarinaNet model was constructed by applying transfer learning to the RetinaNet model using an internal dataset of ICU chest radiographs. The accuracy of the model in predicting the position of the ETT tip and carina was externally validated using a dataset of 200 images extracted from the MIMIC-CXR database. Uncertainty quantification was performed using the level of confidence in the ETT-carina distance prediction. Segmentation of the ETT was carried out using edge detection and pixel clustering.

RESULTS

The interrater agreement was 0.18 cm for the ETT tip position, 0.58 cm for the carina position, and 0.60 cm for the ETT-carina distance. The mean absolute error of the model on the external test set was 0.51 cm for the ETT tip position prediction, 0.61 cm for the carina position prediction, and 0.89 cm for the ETT-carina distance prediction. The assessment of ETT placement was improved by complementing the human interpretation of chest radiographs with the CarinaNet model.

CONCLUSIONS

The CarinaNet model is an efficient and generalizable deep learning algorithm for the automated assessment of ETT placement on ICU chest radiographs. Uncertainty quantification can bring the attention of intensivists to chest radiographs that require an experienced human interpretation. Image segmentation provides intensivists with chest radiographs that are quickly interpretable and allows them to immediately assess the validity of model predictions. The CarinaNet model is ready to be evaluated in clinical studies.

Inter-observer agreement and accuracy in classifying radiographs for pneumoconiosis among Asian physicians taking AIR Pneumo certification examination

This study examined inter-observer agreement and diagnostic accuracy in classifying radiographs for pneumoconiosis among Asian physicians taking the AIR Pneumo examination. We compared agreement and diagnostic accuracy for parenchymal and pleural lesions across residing countries, specialty training, and work experience using data on 93 physicians. Physicians demonstrated fair to good agreement with kappa values 0.30 (95% CI: 0.20-0.40), 0.29 (95% CI: 0.23-0.36), 0.59 (95% CI: 0.52-0.67), and 0.65 (95% CI: 0.55-0.74) in classifying pleural plaques, small opacity shapes, small opacity profusion, and large opacities, respectively. Kappa values among Asian countries ranging from 0.25 to 0.55 (pleural plaques), 0.47 to 0.73 (small opacity profusion), and 0.55 to 0.69 (large opacity size). The median Youden's J index (interquartile range) for classifying pleural plaque, small opacity, and large opacity was 61.1 (25.5), 76.8 (29.3), and 88.9 (23.3), respectively. Radiologists and recent graduates showed superior performance than other groups regarding agreement and accuracy in classifying all types of lesions. In conclusion, Asian physicians taking the AIR Pneumo examination were better at classifying parenchymal lesions than pleural plaques using the ILO classification. The degree of agreement and accuracy was different among countries and was associated with background specialty training.

Strategies for recognizing pneumonia look-alikes

Community-acquired pneumonia is a common diagnosis in children. Among the many children whose symptoms and/or chest X-ray is consistent with community-acquired pneumonia, it can be difficult to distinguish the rare cases of differential diagnoses that require specific management. The aim of this educational article is to provide clinicians with a series of questions to ask themselves in order to detect a possible differential diagnosis of pneumonia in children. The value of this approach is illustrated by 13 real clinical cases in which a child was misdiagnosed as having lobar pneumonia. What is Known: • When a lobar pneumonia is diagnosed, an appropriate antibiotic treatment leads to the resolution of the clinical signs in most cases. • However, several diseases can be look-alikes for pneumonia and mislead the practitioner. What is New: • This article provides a new approach to identify differential diagnoses of pneumonia in children. • It is illustrated by 13 real-life situations of children misdiagnosed as having pneumonia.

Artificial Intelligence-Based Identification of Normal Chest Radiographs: A Simulation Study in a Multicenter Health Screening Cohort

Objective

This study aimed to investigate the feasibility of using artificial intelligence (AI) to identify normal chest radiography (CXR) from the worklist of radiologists in a health-screening environment.

Materials and Methods

This retrospective simulation study was conducted using the CXRs of 5887 adults (mean age ± standard deviation, 55.4 ± 11.8 years; male, 4329) from three health screening centers in South Korea using a commercial AI (Lunit INSIGHT CXR3, version 3.5.8.8). Three board-certified thoracic radiologists reviewed CXR images for referable thoracic abnormalities and grouped the images into those with visible referable abnormalities (identified as abnormal by at least one reader) and those with clearly visible referable abnormalities (identified as abnormal by at least two readers). With AI-based simulated exclusion of normal CXR images, the percentages of normal images sorted and abnormal images erroneously removed were analyzed. Additionally, in a random subsample of 480 patients, the ability to identify visible referable abnormalities was compared among AI-unassisted reading (i.e., all images read by human readers without AI), AI-assisted reading (i.e., all images read by human readers with AI assistance as concurrent readers), and reading with AI triage (i.e., human reading of only those rendered abnormal by AI).

Results

Of 5887 CXR images, 405 (6.9%) and 227 (3.9%) contained visible and clearly visible abnormalities, respectively. With AI-based triage, 42.9% (2354/5482) of normal CXR images were removed at the cost of erroneous removal of 3.5% (14/405) and 1.8% (4/227) of CXR images with visible and clearly visible abnormalities, respectively. In the diagnostic performance study, AI triage removed 41.6% (188/452) of normal images from the worklist without missing visible abnormalities and increased the specificity for some readers without decreasing sensitivity.

Conclusion

This study suggests the feasibility of sorting and removing normal CXRs using AI with a tailored cut-off to increase efficiency and reduce the workload of radiologists.

Lung ultrasound (LUS) in pulmonary tuberculosis: correlation with chest CT and X-ray findings

AIMS

The aim is to describe lung ultrasound (LUS) findings in a cohort of patients with suspected pulmonary tuberculosis (PTB) and compare them with computed tomography (CT) and chest x-ray (CXR) findings in order to evaluate the potentiality of LUS in TB diagnosis.

METHODS

In this prospective study, 82 subjects with suspected TB were enrolled after being evaluated with CXR and chest CT. LUS was performed by blinded radiologists within 3 days after admission. A semiquantitative index was used: score 1 (lesions that extend for about 1-15% of the affected zone), score 2 (15-40%) and score 3 (40-100%).

RESULTS

Microbiological analysis confirmed TB diagnosis in 58/82 (70.7%). CT was positive in all patients, LUS in 79/82 (96.3%) CXR in 78/82 (95.1%) and adding LUS and CXR in 100%. In PTB patients we found a great number of lungs zones with micronodules and with total findings than non-TPB patients (p < 0.05). Overall LUS sensitivity was 80%, greater for micronodules (82%) and nodules (95%), lower for consolidation with air bronchogram (72%) and cavitations (33%). We reported 5 complicated pleural effusion at LUS, only 1 in CT. CXR overall sensitivity was 81%. Adding CXR and LUS findings we reported a sensitivity of 90%.

CONCLUSIONS

LUS could be considered a valid, non-invasive and cost-effective diagnostic tool especially in world regions where CT were not available, also in addiction with CXR.

TRIAL REGISTRATION

This study was approved by the Ethics Committee of our University Hospital (rif. CEAVC 14,816).

Enhancing deep learning based classifiers with inpainting anatomical side markers (L/R markers) for multi-center trials

BACKGROUND AND OBJECTIVE

The protocol for placing anatomical side markers (L/R markers) in chest radiographs varies from one hospital or department to another. However, the markers have strong signals that can be useful for deep learning-based classifier to predict diseases. We aimed to enhance the performance of a deep learning-based classifiers in multi-center datasets by inpainting the L/R markers.

METHODS

The L/R marker was detected with using the EfficientDet detection network; only the detected regions were inpainted using a generative adversarial network (GAN). To analyze the effect of the inpainting in detail, deep learning-based classifiers were trained using original images, marker-inpainted images, and original images clipped using the min-max value of the marker-inpainted images. Binary classification, multi-class classification, and multi-task learning with segmentation and classification were developed and evaluated. Furthermore, the performances of the network on internal and external validation datasets were compared using DeLong's test for two correlated receiver operating characteristic (ROC) curves in binary classification and Stuart-Maxwell test for marginal homogeneity in multi-class classification and multi-task learning. In addition, the qualitative results of activation maps were evaluated using the gradient-class activation map (Grad-CAM).

RESULTS

Marker-inpainting preprocessing improved the classification performances. In the binary classification based on the internal validation, the area under the curves (AUCs) and accuracies were 0.950 and 0.900 for the model trained on the min-max clipped images and 0.911 and 0.850 for the model trained on the original images, respectively (P-value=0.006). In the external validation, the AUCs and accuracies were 0.858 and 0.677 for the model trained using the inpainted images and 0.723 and 0.568 for the model trained using the original images (P-value<0.001), respectively. In addition, the models trained using the marker inpainted images showed the best performance in multi-class classification and multi-task learning. Furthermore, the activation maps obtained using the Grad-CAM improved with the proposed method. The 5-fold validation results also showed improvement trend according to the preprocessing strategies.

CONCLUSIONS

Inpainting an L/R marker significantly enhanced the classifier's performance and robustness, especially in internal and external studies, which could be useful in developing a more robust and accurate deep learning-based classifier for multi-center trials. The code for detection is available at: https://github.com/mi2rl/MI2RLNet. And the code for inpainting is available at: https://github.com/mi2rl/L-R-marker-inpainting.

Bone suppression on pediatric chest radiographs via a deep learning-based cascade model

BACKGROUND AND OBJECTIVE

Bone suppression images (BSIs) of chest radiographs (CXRs) have been proven to improve diagnosis of pulmonary diseases. To acquire BSIs, dual-energy subtraction (DES) or a deep-learning-based model trained with DES-based BSIs have been used. However, neither technique could be applied to pediatric patients owing to the harmful effects of DES. In this study, we developed a novel method for bone suppression in pediatric CXRs.

METHODS

First, a model using digitally reconstructed radiographs (DRRs) of adults, which were used to generate pseudo-CXRs from computed tomography images, was developed by training a 2-channel contrastive-unpaired-image-translation network. Second, this model was applied to 129 pediatric DRRs to generate the paired training data of pseudo-pediatric CXRs. Finally, by training a U-Net with these paired data, a bone suppression model for pediatric CXRs was developed.

RESULTS

The evaluation metrics were peak signal to noise ratio, root mean absolute error and structural similarity index measure at soft-tissue and bone region of the lung. In addition, an expert radiologist scored the effectiveness of BSIs on a scale of 1-5. The obtained result of 3.31 ± 0.48 indicates that the BSIs show homogeneous bone removal despite subtle residual bone shadow.

CONCLUSION

Our method shows that the pixel intensity at soft-tissue regions was preserved, and bones were well subtracted; this can be useful for detecting early pulmonary disease in pediatric CXRs.