[Application of a light-weighted convolutional neural network for automatic recognition of coal workers' pneumoconiosis in the early stage].

Objective: To construct and verify a light-weighted convolutional neural network (CNN), and explore its application value for screening the early stage (subcategory 0/1 and stage Ⅰ of pneumoconiosis) of coal workers' pneumoconiosis (CWP) from digital chest radiography (DR) . Methods: A total of 1225 DR images of coal workers who were examined at an Occupational Disease Prevention and Control Institute in Anhui Province from October 2018 to March 2021 were retrospectively collected. All DR images were collectively diagnosed by 3 radiologists with diagnostic qualifications and gave diagnostic results. There were 692 DR images with small opacity profusion 0/- or 0/0 and 533 DR images with small opacity profusion 0/1 to stage Ⅲ of pneumoconiosis. The original chest radiographs were preprocessed differently to generate four datasets, namely 16-bit grayscale original image set (Origin16), 8-bit grayscale original image set (Origin 8), 16-bit grayscale histogram equalized image set (HE16) and 8-bit grayscale histogram equalized image set (HE8). The light-weighted CNN, ShuffleNet, was applied to train the generated prediction model on the four datasets separately. The performance of the four models for pneumoconiosis prediction was evaluated on a test set containing 130 DR images using measures such as the receiver operating characteristic (ROC) curve, accuracy, sensitivity, specificity, and Youden index. The Kappa consistency test was used to compare the agreement between the model predictions and the physician diagnosed pneumoconiosis results. Results: Origin16 model achieved the highest ROC area under the curve (AUC=0.958), accuracy (92.3%), specificity (92.9%), and Youden index (0.8452) for predicting pneumoconiosis, with a sensitivity of 91.7%. And the highest consistency between identification and physician diagnosis was observed for Origin16 model (Kappa value was 0.845, 95%CI: 0.753-0.937, P<0.001). HE16 model had the highest sensitivity (98.3%) . Conclusion: The light-weighted CNN ShuffleNet model can efficiently identify the early stages of CWP, and its application in the early screening of CWP can effectively improve physicians' work efficiency.

Application of a light-weighted convolutional neural network for automatic recognition of coal workers' pneumoconiosis in the early stage / 中华劳动卫生职业病杂志

Objective: To construct and verify a light-weighted convolutional neural network (CNN), and explore its application value for screening the early stage (subcategory 0/1 and stage Ⅰ of pneumoconiosis) of coal workers' pneumoconiosis (CWP) from digital chest radiography (DR) . Methods: A total of 1225 DR images of coal workers who were examined at an Occupational Disease Prevention and Control Institute in Anhui Province from October 2018 to March 2021 were retrospectively collected. All DR images were collectively diagnosed by 3 radiologists with diagnostic qualifications and gave diagnostic results. There were 692 DR images with small opacity profusion 0/- or 0/0 and 533 DR images with small opacity profusion 0/1 to stage Ⅲ of pneumoconiosis. The original chest radiographs were preprocessed differently to generate four datasets, namely 16-bit grayscale original image set (Origin16), 8-bit grayscale original image set (Origin 8), 16-bit grayscale histogram equalized image set (HE16) and 8-bit grayscale histogram equalized image set (HE8). The light-weighted CNN, ShuffleNet, was applied to train the generated prediction model on the four datasets separately. The performance of the four models for pneumoconiosis prediction was evaluated on a test set containing 130 DR images using measures such as the receiver operating characteristic (ROC) curve, accuracy, sensitivity, specificity, and Youden index. The Kappa consistency test was used to compare the agreement between the model predictions and the physician diagnosed pneumoconiosis results. Results: Origin16 model achieved the highest ROC area under the curve (AUC=0.958), accuracy (92.3%), specificity (92.9%), and Youden index (0.8452) for predicting pneumoconiosis, with a sensitivity of 91.7%. And the highest consistency between identification and physician diagnosis was observed for Origin16 model (Kappa value was 0.845, 95%CI: 0.753-0.937, P<0.001). HE16 model had the highest sensitivity (98.3%) . Conclusion: The light-weighted CNN ShuffleNet model can efficiently identify the early stages of CWP, and its application in the early screening of CWP can effectively improve physicians' work efficiency.

Ultrasound imaging features of bronchial anthracofibrosis: A case-control study.

To determine the ultrasound imaging characteristics of patients with bronchial anthracofibrosis (BAF) and identify clinical markers for prevention and treatment. We randomly selected 1243 participants (113 with BAF) who underwent bronchoscopy and received treatment at our institution between April 2018 and October 2019. BAF was classified as flat, deep-seated retracted, or black mucosal protruding based on microscopic findings. Ultrasound probes were used to determine the maximum thickness of the tube walls and submucosa. The average values of the submucosal and bony tissue areas in the BAF subtypes were compared. The BAF group included 13 participants with a history of tuberculosis (11.5%) and 57 participants with biofuel exposure (50.4%). The average exposure time was 17.4 ± 6.2 years; BAF accounted for 10% of the bronchoscopies performed. The maximum tube-wall thicknesses of the deep-seated retracted (17.3 ± 5.7) and black mucosal protruding (19.3 ± 5.4) groups were significantly greater than those of the flat group (12.5 ± 5.0; P < .05). The maximum thicknesses of the submucosa in the deep-retracted (9.8 ± 3.0) and black mucosal protruding (14.5 ± 5.0) groups were significantly greater than that of the flat group (6.6 ± 3.5; P < .05). The ratios of bone tissue in the flat and black mucosal protruding groups were 33.3 ± 9.3% and 34.9% ± 12.1%, respectively. The ratio in the deep-seated retracted group (65.2% ± 8.7%) was significantly reduced (P < .05). The flat group showed no significant change (P > .05). Differences in BAF airway remodeling among different subtypes may lead to varying clinical symptoms. Analyzing the characteristics of BAF airway remodeling and the regulatory pathway may provide new clues for treatment.

Use data augmentation for a deep learning classification model with chest X-ray clinical imaging featuring coal workers' pneumoconiosis.

PURPOSE: This paper aims to develop a successful deep learning model with data augmentation technique to discover the clinical uniqueness of chest X-ray imaging features of coal workers' pneumoconiosis (CWP). PATIENTS AND METHODS: We enrolled 149 CWP patients and 68 dust-exposure workers for a prospective cohort observational study between August 2021 and December 2021 at First Hospital of Shanxi Medical University. Two hundred seventeen chest X-ray images were collected for this study, obtaining reliable diagnostic results through the radiologists' team, and confirming clinical imaging features. We segmented regions of interest with diagnosis reports, then classified them into three categories. To identify these clinical features, we developed a deep learning model (ShuffleNet V2-ECA Net) with data augmentation through performances of different deep learning models by assessment with Receiver Operation Characteristics (ROC) curve and area under the curve (AUC), accuracy (ACC), and Loss curves. RESULTS: We selected the ShuffleNet V2-ECA Net as the optimal model. The average AUC of this model was 0.98, and all classifications of clinical imaging features had an AUC above 0.95. CONCLUSION: We performed a study on a small dataset to classify the chest X-ray clinical imaging features of pneumoconiosis using a deep learning technique. A deep learning model of ShuffleNet V2 and ECA-Net was successfully constructed using data augmentation, which achieved an average accuracy of 98%. This method uncovered the uniqueness of the chest X-ray imaging features of CWP, thus supplying additional reference material for clinical application.

Computer-Aided Diagnosis of Coal Workers' Pneumoconiosis in Chest X-ray Radiographs Using Machine Learning: A Systematic Literature Review.

Computer-aided diagnostic (CAD) systems can assist radiologists in detecting coal workers' pneumoconiosis (CWP) in their chest X-rays. Early diagnosis of the CWP can significantly improve workers' survival rate. The development of the CAD systems will reduce risk in the workplace and improve the quality of chest screening for CWP diseases. This systematic literature review (SLR) amis to categorise and summarise the feature extraction and detection approaches of computer-based analysis in CWP using chest X-ray radiographs (CXR). We conducted the SLR method through 11 databases that focus on science, engineering, medicine, health, and clinical studies. The proposed SLR identified and compared 40 articles from the last 5 decades, covering three main categories of computer-based CWP detection: classical handcrafted features-based image analysis, traditional machine learning, and deep learning-based methods. Limitations of this review and future improvement of the review are also discussed.

Prevalence and severity of abnormal lung function among US former coal miners with and without radiographic coal workers' pneumoconiosis.

OBJECTIVES: Examination of lung function abnormalities among coal miners has historically focused on actively working miners. This likely underestimates the true burden of chronic respiratory disease. The objective of this study was to characterise patterns and severity of lung function impairment among a population of former coal miners. METHODS: Cross-sectional data from 2568 former coal miners evaluated at eight US Black Lung clinics in a 12-month period were retrospectively analysed for patterns of prebronchodilator spirometric abnormality and severity of lung function impairment. Spirometry data from a subset of former miners with chest radiographs were analysed based on the presence and severity of coal workers' pneumoconiosis (CWP). RESULTS: Abnormal spirometry was identified in 56.6% of subjects. The age-standardised prevalence of airflow obstruction among miners aged ≥45 years was 18.9% overall and 12.2% among never smokers. Among 1624 subjects who underwent chest radiography, the prevalence and severity of abnormal spirometry increased with worsening radiographic category for pneumoconiosis. Of never-smoking former miners without radiographic CWP, 39.0% had abnormal spirometry; 25.1% had abnormally low forced expiratory volume in 1 s (FEV1), and 17.1% had moderate to severe FEV1 impairment. CONCLUSIONS: Abnormal spirometry is common among former coal miners. While ever-smoking former miners had higher rates of airflow obstruction, never-smoking former miners also demonstrated clinically significant airflow obstruction, including those without radiographic pneumoconiosis. These findings demonstrate the importance of recognising physiological as well as imaging manifestations of coal mine dust lung diseases in former miners.

Early radiographic pneumoconiosis is associated with impaired exercise gas exchange among coal miners with normal resting spirometry.

RATIONALE: We sought to determine if radiographic pneumoconiosis predicts abnormal gas exchange during exercise in coal mine workers with preserved resting lung function. METHODS: We analyzed data from former coal miners seen between 2006 and 2014 in a single clinic specializing in black lung evaluations. We limited the analysis to those with normal resting spirometry and an A-a gradient at peak exercise ≥10 mmHg. We used multivariable logistic regression to estimate predictors of A-a gradient widened to >150% of the reference value. We focused on chest radiographs consistent with pneumoconiosis, taking into account higher silica exposure mining activities and years underground, and adjusting for cigarette smoking, obesity, and coronary artery disease. RESULTS: Of 5507 miners, we analyzed data for 742 subjects with normal spirometry and all key clinical variables available, of whom 372 (50.1%) had radiographic evidence of pneumoconiosis. All but 21 had small opacity profusion of less than 2/1. The median A-a gradient at peak exercise was 108% of reference value (interquartile range, 81%-141%). In the multivariable analysis, radiographic pneumoconiosis was associated with increased odds of widened A-a gradient (odds ratio [OR], 2.47; 95% confidence interval [CI], 1.7-3.7). Limited to 660 subjects with normal diffusing capacity for carbon monoxide, the odds were similarly increased (OR, 3.20; 95% CI, 1.5-3.6). DISCUSSION: Among coal miners with preserved resting lung function, radiographic evidence of early pneumoconiosis more than doubled the odds of abnormal exercise physiology. Impairment in pneumoconiosis occurs in early disease and may only be evident on exercise testing.

Association between Financial Conflicts of Interest and International Labor Office Classifications for Black Lung Disease.

Rationale: The U.S. Department of Labor administers the Federal Black Lung Program (FBLP), an administrative system charged with managing claims by coal miners for workers' compensation for totally disabling coal mine dust lung disease. Specific case reports have raised concern that financial conflicts of interest (COIs) may systematically bias physicians when they are classifying chest X-rays (CXRs) for the absence, presence, and severity of pneumoconiosis. Objectives: To evaluate the direction and magnitude of association between financial COIs of physicians participating in the FBLP and international standards for the classification of radiographs of pneumoconiosis. Methods: An epidemiologic assessment of black lung claims filed to the FBLP from 2000 to 2013 was conducted to determine physician classifications of radiographs. FBLP court decisions from 2002 to 2019 (n = 7,656) were used to evaluate financial COIs of each physician. The main outcome measures used were classifications of radiographs for the absence of pneumoconiosis (small opacity classifications of 0/0 or 0/1), simple pneumoconiosis (small opacity classifications of 1/0 through 3/+), and progressive massive fibrosis (PMF) (large opacities with classifications of A, B, or C). Results: Of 63,780 radiograph classifications made by 264 physicians, 31.4% were read positive for simple pneumoconiosis and 3.6% were read as having PMF. There were 52 physicians who classified CXRs as having no evidence of pneumoconiosis in 99%+ of their readings and 18 physicians who classified CXRs as positive for simple pneumoconiosis in 99%+ of their readings. The adjusted odds of a negative classification of pneumoconiosis was 1.46 (95% confidence interval [CI], 1.44-1.47) per 10% increase in the proportion of court records demonstrating that a physician was hired by the employer. Per 10% increase in court records indicating a physician was hired by the miner/claimant, the adjusted odds ratio for classifying simple pneumoconiosis was 1.51 (95% CI, 1.49-1.52), and the adjusted odds ratio for finding PMF was 1.28 (95% CI, 1.26-1.30). Conclusions: There was a strong association between source of payment and radiograph classification, suggesting the importance of eliminating financial COIs in what should be an objective determination of eligibility for Black Lung Workers' compensation benefits.

The value of endobronchial ultrasound-guided transbronchial needle aspiration, 18-fluorodeoxyglucose positron emission tomography/computed tomography, and ultrasonography imaging techniques in the diagnosis of mediastinal and/or hilar malignant, anthracotic, and other benign lymph nodes.

INTRODUCTION: Endobronchial ultrasound-guided transbronchial needle aspiration (EBUS-TBNA) is a simple, reliable, minimally invasive and effective procedure. However, a surgical technique may be required, if the results are negative. Therefore, there is a need for new studies to increase the diagnostic value of EBUS-TBNA and provide additional information to guide the biopsy in performing the procedure. Here, we aimed to investigate the diagnostic value of EBUS-TBNA and 18-fluorodeoxyglucose positron emission tomography/computed tomography (18F-FDG PET/CT) in diagnosis of hilar and/or mediastinal lymph nodes (LNs). It was also aimed to determine the contributions of real-time ultrasonography (USG) images of LNs to distinguishing between the malignant and benign LNs during EBUS-TBNA, and in the diagnosis of anthracotic LNs. MATERIAL AND METHOD: In the retrospective study including 545 patients, 1068 LNs were sampled by EBUS-TBNA between January 2015 and February 2020. EBUS-TBNA, 18-FDG PET/CT and images of USG were investigated in the diagnosis of mediastinal and/or hilar malignant, anthracotic and other benign LNs. RESULTS: The sensitivity, specificity, positive predictive value and negative predictive value of EBUS-TBNA were found as 79.5, 98.1, 89.5, and 91.7%, respectively. Mean maximum standardized uptake value (SUVmax) values of 18F-FDG PET/CT were 6.31±4.3 in anthracotic LNs and 5.07 ±â€Š2.53 in reactive LNs. Also, mean SUVmax of malignant LNs was 11.02 ±â€Š7.30 and significantly higher than that of benign LNs. In differentiation of malignant-benign tumors, considering the cut off value of 18F-FDG PET/CT SUVmax as 2.72, the sensitivity and specificity was 99.3 and 11.7%, but given the cut off value as 6.48, the sensitivity, specificity, positive predictive value and negative predictive value was found as 76.5, 64, 20.49, and 78.38% for benign LNs, respectively. Compared LNs as to internal structure and contour features, malignant LNs had most often irregular contours and heterogeneous density. Anthracotic, reactive and other benign LNs were most frequently observed as regular contours and homogeneous density. The difference between malignant and benign LNs was significant. CONCLUSION: EBUS can contribute to the differential diagnosis of malignant, anthracotic and other benign LNs. Such contributions can guide clinician bronchoscopists during EBUS-TBNA. The triple modality of EBUS-TBNA, 18FDG PET/CT, and USG may increase the diagnostic value in hilar and mediastinal lymphadenopathies.

Biomass Fuel Exposure and Lungs: Unraveling the Imaging Conundrum.

Prolonged exposure to biomass fuel smoke is a proven irritant, known to aggravate chronic lung diseases. Of the myriad spectrum of thoracic manifestations associated with inhalation of biomass fuel smoke, bronchial anthracofibrosis is a recently described entity characterized by bronchial narrowing and visible anthracotic pigmentation on bronchoscopy. Common imaging features include bronchostenosis, peribronchial soft tissue with or without calcification along with peribronchial lymph nodes. Its close similarity to endobronchial tuberculosis and bronchogenic carcinoma in clinical presentation and imaging poses a diagnostic challenge and hence underlines the importance of knowledge about this entity. This review aims to summarize the key imaging features of bronchial anthracofibrosis while also briefly discussing the spectrum of thoracic manifestations including distinct entities associated with biomass fuel smoke exposure.

Progression of coal workers' pneumoconiosis absent further exposure.

OBJECTIVES: The natural history of coal workers' pneumoconiosis (CWP) after cessation of exposure remains poorly understood. METHODS: We characterised the development of and progression to radiographic progressive massive fibrosis (PMF) among former US coal miners who applied for US federal benefits at least two times between 1 January 2000 and 31 December 2013. International Labour Office classifications of chest radiographs (CXRs) were used to determine initial and subsequent disease severity. Multivariable logistic regression models were used to identify major predictors of disease progression. RESULTS: A total of 3351 former miners applying for benefits without evidence of PMF at the time of their initial evaluation had subsequent CXRs. On average, these miners were 59.7 years of age and had 22 years of coal mine employment. At the time of their first CXR, 46.7% of miners had evidence of simple CWP. At the time of their last CXR, 111 miners (3.3%) had radiographic evidence of PMF. Nearly half of all miners who progressed to PMF did so in 5 years or less. Main predictors of progression included younger age and severity of simple CWP at the time of initial CXR. CONCLUSIONS: This study provides further evidence that radiographic CWP may develop and/or progress absent further exposure, even among miners with no evidence of radiographic pneumoconiosis after leaving the industry. Former miners should undergo regular medical surveillance because of the risk for disease progression.

[Analysis on characteristic of patients with stage I coal worker's pneumoconiosis].

Objective: To compare the clinical and imaging characteristics of patients with stage I coal worker's pneumoconiosis (CWP) . Methods: All 347 cases of coal workers' pneumoconiosis diagnosed in the Third Hospital of Peking University from January 2014 to December 2018 were included in the study. According to different working posts, the subjects were divided into three categories: mining, tunneling and mixing workers. Dust exposure duration, initial dust exposure age, diagnosis age, latency, small shadow shape and lung regions distribution in X-ray chest film of different categories of CWP patients were analyzed. Results: Among the 347 patients, 216 were mining workers (62.2%) , 77 were tunneling workers (22.2%) and 54 were mixing workers (15.6%) . The dust exposure duration of mining, tunneling and mixing workers were (14.5±7.0) , (16.3±8.2) and (19.0±8.8) years, respectively. There are statistically significant differences in dust exposure duration between different categories of workers (P<0.05) . There were no significant difference in the age of diagnosis, initial dust exposure age and the latency between different categories of workers (P>0.05) . The X-ray films of mining, tunneling and mixing workers showed small round shadow, accounting for 50.9% (110/216) , 96.1% (74/77) and 96.3% (52/54) respectively. 48.1% (104/216) of the mining workers and 38.9% (21/54) of mixing workers, the distribution of small shadow in chest X-ray films reached middle and lower lung regions, while in the 48.1% (37/77) of the tunneling workers, the distribution of small shadow in chest X-ray films reached lower lung regions. There were differences in above indicators among workers with different categories (P<0.05) . Conclusion: The dust exposure duration, the shape and the distribution of lung area on chest X-ray films are different in stage I CWP patients of different occupational categories.

The utility of elastography during EBUS-TBNA in a population with a high prevalence of anthracosis.

BACKGROUND: Ultrasound elastography, is a pioneer sonographic modality that is conducted during endobronchial ultrasound-guided transbronchial needle aspiration (EBUS-TBNA) in order to increase the accuracy of sampling location. The current study aims to evaluate the usefulness of elastography during EBUS-TBNA in a population with a high prevalence of anthracosis. METHODS: This prospective single-blinded study was performed on 69 lymph nodes (LNs) of patients with mediastinal lymphadenopathy undergoing EBUS-TBNA and EBUS-elastography from October 2017 to July 2018. The stiffness level of the tissue was translated into a color to demonstrate the hardness of tissue. Blue and total areas of each section were measured to calculate the hardness of each LN. RESULTS: Sixty-nine LNs were evaluated by elastography. Twenty percent of LNs were malignant. There was a statistical difference between malignant and non-malignant nodes based on color dominancy (P = 0.032). However, with the exclusion of anthracosis nodes from the analysis, the difference was more significant (P < 0.001). Moreover, when the blue dominancy was used as the predictor of malignancy or anthracosis, the results showed a significant correlation (P < 001). CONCLUSION: The usefulness of elastography in selecting the hardest area of tissue that is appropriate for diagnosing diseases has been proven previously. Since in countries with a high prevalence of anthracosis, blue color achieved using elastography predicts either malignancy or anthracosis so, cases with blue dominancy of LNs in elastography and the white color in the EBUS-TBNA indicate anthracosis-caused calcification should be reconsidered.

What does coal mine dust lung disease look like? A radiological review following re-identification in Queensland.

INTRODUCTION: Coal mine dust lung disease (CMDLD), including the pneumoconioses, dust-related diffuse fibrosis (DDF) and chronic obstructive pulmonary disease (COPD), are occupational lung diseases attributed to respirable coal mine dust. Following the re-identification of CMDLD in Queensland in 2015, we undertook a case series to understand their radiological presentation. METHODS: Chest radiographs and high-resolution computed tomography (HRCT) were retrospectively reviewed for 79 male individuals diagnosed by a respiratory physician with a CMDLD since 2015. Radiological findings were characterised as per the International Labour Office Classification System (ILO system) and the International Classification of HRCT for Occupational and Environmental Respiratory Diseases (ICOERD). RESULTS: Subjects with pneumoconiosis (n = 56) demonstrated widespread opacities with bilateral upper zone predominance. The majority of the lung was impacted, with 72% and 79% of zones demonstrating opacities on chest radiograph and HRCT, respectively. Most pneumoconiosis subjects (71%) demonstrated ILO category 1 disease, while 29% had advanced disease (ILO grades ≥ 2/1). A high proportion (81%) of pneumoconiosis subjects demonstrated at least one radiological feature associated with exposure to respirable crystalline silica (RCS). DDF subjects (n = 5) had radiologically severe disease (mean ILO 2/1) with lower zone-predominant irregular opacities. Widespread emphysema, with no zone dominance, was the key radiological feature in those with COPD (n = 18). CONCLUSION: Radiological findings of particular interest included the high burden of opacities observed and the presence of RCS-associated features in the majority of subjects. Radiologists are at the front line in occupational lung disease screening/diagnosis and must be aware of the imaging spectrum.

Radiological appearance of coal mine dust lung diseases in Australian workers.

Coal Mine Dust Lung Disease (CMDLD) encompasses a spectrum of lung diseases caused by prolonged exposure to coal mine dust. This review presents high-resolution computed tomography (HRCT) images from men diagnosed with a CMDLD since the resurgence of these diseases in Queensland in 2015.