Chest X-ray Research Articles

An ensemble learning model for detection of pulmonary hypertension using electrocardiogram, chest X-ray, and brain natriuretic peptide

Abstract Aims Delayed diagnosis of pulmonary hypertension (PH) is a known cause of poor patient prognosis. We aimed to develop an artificial intelligence (AI) model, using ensemble learning method to detect PH using electrocardiography (ECG), chest X-ray (CXR), and brain natriuretic peptide (BNP), facilitating accurate detection and prompting further examinations. Methods and results We developed a convolutional neural network model using ECG data to predict PH, labelled by ECG from seven institutions. Logistic regression was used for the BNP prediction model. We referenced a CXR deep learning model using ResNet18. Outputs from each of the three models were integrated into a three-layer fully connected multimodal model. Ten cardiologists participated in an interpretation test, detecting PH from patients’ ECG, CXR, and BNP data both with and without the ensemble learning model. The area under the receiver operating characteristic curves of the ECG, CXR, BNP, and ensemble learning model were 0.818 [95% confidence interval (CI), 0.808–0.828], 0.823 (95% CI, 0.780–0.866), 0.724 (95% CI, 0.668–0.780), and 0.872 (95% CI, 0.829–0.915). Cardiologists’ average accuracy rates were 65.0 ± 4.7% for test without AI model and 74.0 ± 2.7% for test with AI model, a statistically significant improvement (P < 0.01). Conclusion Our ensemble learning model improved doctors’ accuracy in detecting PH from ECG, CXR, and BNP examinations. This suggests that earlier and more accurate PH diagnosis is possible, potentially improving patient prognosis.

Role of Artificial Intelligence in Identifying Vital Biomarkers with Greater Precision in Emergency Departments During Emerging Pandemics

The COVID-19 pandemic has accelerated advances in molecular biology and virology, enabling the identification of key biomarkers to differentiate between severe and mild cases. Furthermore, the use of artificial intelligence (AI) and machine learning (ML) to analyze large datasets has been crucial for rapidly identifying relevant biomarkers for disease prognosis, including COVID-19. This approach enhances diagnostics in emergency settings, allowing for more accurate and efficient patient management. This study demonstrates how machine learning algorithms in emergency departments can rapidly identify key biomarkers for the vital prognosis in an emerging pandemic using COVID-19 as an example by analyzing clinical, epidemiological, analytical, and radiological data. All consecutively admitted patients were included, and more than 89 variables were processed using the Random Forest (RF) algorithm. The RF model achieved the highest balanced accuracy at 92.61%. The biomarkers most predictive of mortality included procalcitonin (PCT), lactate dehydrogenase (LDH), and C-reactive protein (CRP). Additionally, the system highlighted the significance of interstitial infiltrates in chest X-rays and D-dimer levels. Our results demonstrate that RF is crucial in identifying critical biomarkers in emerging diseases, accelerating data analysis, and optimizing prognosis and personalized treatment, emphasizing the importance of PCT and LDH in high-risk patients.

CXR-LLaVA: a multimodal large language model for interpreting chest X-ray images.

This study aimed to develop an open-source multimodal large language model (CXR-LLaVA) for interpreting chest X-ray images (CXRs), leveraging recent advances in large language models (LLMs) to potentially replicate the image interpretation skills of human radiologists. For training, we collected 592,580 publicly available CXRs, of which 374,881 had labels for certain radiographic abnormalities (Dataset 1) and 217,699 provided free-text radiology reports (Dataset 2). After pre-training a vision transformer with Dataset 1, we integrated it with an LLM influenced by the LLaVA network. Then, the model was fine-tuned, primarily using Dataset 2. The model's diagnostic performance for major pathological findings was evaluated, along with the acceptability of radiologic reports by human radiologists, to gauge its potential for autonomous reporting. The model demonstrated impressive performance in test sets, achieving an average F1 score of 0.81 for six major pathological findings in the MIMIC internal test set and 0.56 for six major pathological findings in the external test set. The model's F1 scores surpassed those of GPT-4-vision and Gemini-Pro-Vision in both test sets. In human radiologist evaluations of the external test set, the model achieved a 72.7% success rate in autonomous reporting, slightly below the 84.0% rate of ground truth reports. This study highlights the significant potential of multimodal LLMs for CXR interpretation, while also acknowledging the performance limitations. Despite these challenges, we believe that making our model open-source will catalyze further research, expanding its effectiveness and applicability in various clinical contexts. Question How can a multimodal large language model be adapted to interpret chest X-rays and generate radiologic reports? Findings The developed CXR-LLaVA model effectively detects major pathological findings in chest X-rays and generates radiologic reports with a higher accuracy compared to general-purpose models. Clinical relevance This study demonstrates the potential of multimodal large language models to support radiologists by autonomously generating chest X-ray reports, potentially reducing diagnostic workloads and improving radiologist efficiency.

Prediction of tuberculosis-specific mortality for older adult patients with pulmonary tuberculosis

BackgroundThis study aims to identify risk factors associated with tuberculosis-specific mortality (TSM) in older adult patients with pulmonary tuberculosis (TB) and to develop a competing risk nomogram for TSM prediction.MethodsWe conducted a retrospective cohort study and randomly selected 528 older adult pulmonary TB patients hospitalized in designated hospitals in Henan Province between January 2015 and December 2020. The cumulative incidence function (CIF) was calculated for both TSM and non-tuberculosis-specific mortality (non-TSM). A Fine and Gray proportional subdistribution hazards model and a competing risk nomogram were developed to predict TSM in older adult patients.ResultsThe 5-year cumulative incidence functions (CIFs) for TSM and non-TSM were 9.7 and 9.4%, respectively. The Fine and Gray model identified advanced age, retreatment status, chest X-rays (CXR) cavities, and hypoalbuminemia as independent risk factors for TSM. The competing risk nomogram for TSM showed good calibration and excellent discriminative ability, achieving a concordance index (c-index) of 0.844 (95% confidence interval [CI]: 0.830–0.857).ConclusionThe Fine and Gray model provided an accurate evaluation of risk factors associated with TSM. The competing risk nomogram, developed using the Fine and Gray model, provided accurate and personalized predictions of TSM.

Interpretable Deep Learning for Pneumonia Detection Using Chest X-Ray Images

Pneumonia remains a global health issue, creating the need for accurate detection methods for effective treatment. Deep learning models like ResNet50 show promise in detecting pneumonia from chest X-rays; however, their black-box nature limits the transparency, which fails to meet that needed for clinical trust. This study aims to improve model interpretability by comparing four interpretability techniques, which are Layer-wise Relevance Propagation (LRP), Adversarial Training, Class Activation Maps (CAMs), and the Spatial Attention Mechanism, and determining which fits best the model, enhancing its transparency with minimal impact on its performance. Each technique was evaluated for its impact on the accuracy, sensitivity, specificity, AUC-ROC, Mean Relevance Score (MRS), and a calculated trade-off score that balances interpretability and performance. The results indicate that LRP was the most effective in enhancing interpretability, achieving high scores across all metrics without sacrificing diagnostic accuracy. The model achieved 0.91 accuracy and 0.85 interpretability (MRS), demonstrating its potential for clinical integration. In contrast, Adversarial Training, CAMs, and the Spatial Attention Mechanism showed trade-offs between interpretability and performance, each highlighting unique image features but with some impact on specificity and accuracy.

Comparison of dark-field chest radiography and CT for the assessment of COVID-19 pneumonia

BackgroundDark-field chest radiography allows the assessment of the structural integrity of the alveoli by exploiting the wave properties of x-rays.PurposeTo compare the qualitative and quantitative features of dark-field chest radiography in patients with COVID-19 pneumonia with conventional CT imaging.Materials and methodsIn this prospective study conducted from May 2020 to December 2020, patients aged at least 18 years who underwent chest CT for clinically suspected COVID-19 infection were screened for participation. Inclusion criteria were a CO-RADS score ≥4, the ability to consent to the procedure and to stand upright without help. Participants were examined with a clinical dark-field chest radiography prototype. For comparison, a healthy control cohort of 40 subjects was evaluated. Using Spearman's correlation coefficient, correlation was tested between dark-field coefficient and CT-based COVID-19 index and visual total CT score as well as between the visual total dark-field score and the visual total CT score.ResultsA total of 98 participants [mean age 58 ± 14 (standard deviation) years; 59 men] were studied. The areas of signal intensity reduction observed in dark-field images showed a strong correlation with infiltrates identified on CT scans. The dark-field coefficient had a negative correlation with both the quantitative CT-based COVID-19 index (r = −.34, p = .001) and the overall CT score used for visual grading of COVID-19 severity (r = −.44, p < .001). The total visual dark-field score for the presence of COVID-19 was positively correlated to the total CT score for visual COVID-19 severity grading (r = .85, p < .001).ConclusionCOVID-19 pneumonia-induced signal intensity losses in dark-field chest radiographs are consistent with CT-based findings, showing the technique's potential for COVID-19 assessment.

Skeletal surveys for suspected infant abuse: patient-specific radiation dose estimation using a hybrid computational phantom.

Radiographic skeletal survey plays an important role in the diagnosis of infant abuse. Some practitioners have expressed concerns about the radiation exposure from this examination. To utilize state-of-the-art hybrid computational phantoms to more accurately estimate radiation doses of skeletal surveys performed for suspected infant abuse. We searched our imaging database to identify skeletal surveys performed for suspected infant abuse (5/2020-5/2022). Initial skeletal surveys consisted of 25 standardized radiographs while follow-up skeletal surveys consisted of 16 standardized radiographs (no frontal or lateral views of the skull; or lateral views of the spine, knees, and ankles). To estimate the patient-specific organ and effective dose, we applied the National Cancer Institute dosimetry system for Radiography and Fluoroscopy (with on-the-fly 3D Monte Carlo simulation) to the male and female infant hybrid computational phantoms. The mean total effective radiation dose was 0.627 mSv (initial survey) and 0.495 mSv (follow-up survey). For both surveys, the anteroposterior chest radiograph was the largest contributor to effective dose (contributing 0.101 mSv and 0.108 mSv, respectively). In the initial skeletal survey, the lens and the eyeballs received the highest organ absorbed doses (with the skull radiographs as the major contributors); and in the follow-up skeletal survey, the breasts received the highest organ absorbed dose (with the chest radiographs as the major contributors). We employed hybrid computational phantoms to better estimate the radiation profile of skeletal surveys performed for suspected infant abuse, thus enabling us to update and optimize this life-saving imaging protocol.

Characteristics and long-term health outcomes of the first domestic COVID-19 outbreak cases in Da Nang, Vietnam: a longitudinal cohort study

BackgroundVietnam experienced the first COVID-19 domestic outbreak due to the Wuhan strain (B.1.1) in Da Nang from July 2020. COVID-19 can cause acute as well as long-term health problems. We aimed to characterise clinical features and risk factors related to severe illness of COVID-19 among Da Nang outbreak cases and to describe long-term health outcomes among survivors of this outbreak.MethodsWe conducted an ambidirectional cohort study. Study subjects were all hospitalised cases with positive real-time PCR test of SARS-CoV-2 in the three major hospitals in Da Nang from 25 July to 28 August 2020. Clinical and demographic information was retrospectively collected from medical charts. Then, the survivors were followed-up prospectively at 6 and 16 months after acute infection to assess their health status via standardized questionnaires, physical examinations, chest X-rays and pulmonary function tests.ResultsA total of 362 cases including 20 fatal cases were enrolled into the study retrospectively. The median age of the participants included in the medical chart review was 46.5 years and 60.8% were female. Overall, 7.8% of the participants required respiratory support during hospitalisation and 20 of them died. Compared to the survivors, the fatal cases were significantly older (median age of survivors 45.0 yr vs fatal cases 66.5 yr, P < 0.001) and more likely to have underlying conditions. The proportions of participants who had at least one long COVID symptom within the 7 days of each follow-up at 6 and 16 months were 72.0% (134/186) and 63.5% (47/74), respectively. We also found that females and adults reported symptoms more often in the follow-up surveys, 78.9% (90/114) [females] vs 61.1% (44/72) [males] at 6 months, P = 0.008; 68.7% (46/67) [≥ 20 years] vs 14.3% (1/7) [< 20 years] at 16 months, P = 0.004.ConclusionsIn the first domestic COVID-19 outbreak in Vietnam, the mortality rate was approximately 6% and associated with underlying medical conditions. In the follow-up surveys, a substantial proportion of participants reported long COVID related health problems, although the prevalence declined over time. Females and adults reported symptoms more often, which might be due to the pathophysiological differences according to sex and age.

Application of Artificial Intelligence in Radiological Image Analysis for Pulmonary Disease Diagnosis: A Review of Current Methods and Challenges

Introduction and purposeArtificial intelligence (AI), particularly machine learning (ML) and deep learning (DL), is revolutionizing radiology by improving diagnostic accuracy and efficiency. This paper examines AI applications, especially convolutional neural networks (CNNs), in diagnosing pulmonary diseases, such as pneumonia, tuberculosis, and lung cancer. The goal is to explore the impact of these technologies and assess challenges in their integration into clinical practice. Material and methodsThis review is based on articles from the PubMed database, published between 2015 and 2024, using keywords such as artificial intelligence in radiology, AI in medicine, AI in chest X-ray, and AI in chest-CT. ResultsAI, driven by ML and DL, has significantly enhanced medical imaging analysis, automating tasks that require expert interpretation. CNNs excel in processing raw image data and identifying hierarchical features, surpassing traditional methods in diagnosing lung diseases from radiographs and CT scans. AI systems demonstrate exceptional accuracy in detecting pneumonia, tuberculosis, and lung cancer, providing rapid, consistent results, particularly valuable in resource-limited settings. However, challenges persist, including the need for diverse training datasets, model interpretability, and integration into existing workflows. ConclusionsAI, especially CNN-based DL models, is reshaping radiology by advancing diagnostic capabilities. While it often outperforms traditional methods, AI is best used to complement human expertise. Overcoming challenges in data quality, system integration, and training is essential for broader clinical adoption. Continued research will enhance AI’s reliability and utility, ultimately improving patient outcomes.

Predictors of developing severe COVID-19 among hospitalized patients: a retrospective study

BackgroundBased on the comparative statistical tests, patients with severe COVID-19 were statistically significantly associated with older age and had higher respiratory rate, longer hospital stay, and higher prevalence of diabetes than non-severe cases. This work aimed to explore risk factors for developing severe COVID-19 to enhance the quality of care provided to patients and to prevent complications.MethodsA retrospective study was conducted in Saudi Arabia’s eastern province, including all COVID-19 patients aged 18 years or older who were hospitalized at Prince Saud Bin Jalawi Hospital in July 2020. Comparative tests as well as both univariate and multivariate logistic regression analyses were performed to identify risk factors for developing severe COVID-19 and poor outcomes.ResultsBased on the comparative statistical tests patients with severe COVID-19 were statistically significantly associated with older age and had higher respiratory rate, longer hospital stay, and higher prevalence of diabetes than non-severe cases. They also exhibited statistically significant association with high levels of potassium, urea, creatinine, lactate dehydrogenase (LDH), D-dimer, and aspartate aminotransferase (AST). The univariate analysis shows that having diabetes, having high severe acute respiratory infection chest X-ray scores, old age, prolong hospitalization, high potassium and lactate dehydrogenase, as well as using insulin, heparin, corticosteroids, favipiravir or azithromycin were all statistically significant associated with severe COVID-19. However, after adjustments in the multivariate analysis, the sole predictor was serum LDH (p = 0.002; OR 1.005; 95% CI 1.002–1.009). In addition, severe COVID-19 patients had higher odds of being prescribed azithromycin than non-severe patients (p = 0.001; OR 13.725; 95% CI 3.620–52.043). Regarding the outcomes, the median hospital stay duration was statistically significantly associated with death, intensive care unit admission (ICU), and mechanical ventilation. On the other hand, using insulin, azithromycin, beta-agonists, corticosteroids, or favipiravir were statistically significantly associated with reduced mortality, ICU admission, and need of mechanical ventilation.ConclusionThis study sheds light on numerous parameters that may be utilized to construct a prediction model for evaluating the risk of severe COVID-19. However, no protective factors were included in this prediction model.

Predicting Intensive Care Unit Admissions in COVID-19 Patients: An AI-Powered Machine Learning Model

Intensive Care Units (ICUs) have been in great demand worldwide since the COVID-19 pandemic, necessitating organized allocation. The spike in critical care patients has overloaded ICUs, which along with prolonged hospitalizations, has increased workload for medical personnel and lead to a significant shortage of resources. The study aimed to improve resource management by quickly and accurately identifying patients who need ICU admission. We designed an intelligent decision support system that employs machine learning (ML) to anticipate COVID-19 ICU admissions in Kuwait. Our algorithm examines several clinical and demographic characteristics to identify high-risk individuals early in illness diagnosis. We used 4399 patients to identify ICU admission with predictors such as shortness of breath, high D-dimer values, and abnormal chest X-rays. Any data imbalance was addressed by employing cross-validation along with the Synthetic Minority Oversampling Technique (SMOTE), the feature selection was refined using backward elimination, and the model interpretability was improved using Shapley Additive Explanations (SHAP). We employed various ML classifiers, including support vector machines (SVM). The SVM model surpasses all other models in terms of precision (0.99) and area under curve (AUC, 0.91). This study investigated the healthcare process during a pandemic, facilitating ML-based decision-making solutions to confront healthcare problems.

Coexistent anemia modulates systemic inflammation and exacerbates disease severity and adverse treatment outcomes in tuberculosis

IntroductionAnemia has been shown to be an independent predictor of disease progression and death in tuberculosis (TB) patients, significantly impacting TB in several ways. This dual burden poses significant challenges for TB control efforts. However, the mechanism by which anemia influences disease severity, bacterial burden, and TB treatment outcomes remains poorly understood.MethodsIn this study, we aimed to compare bacterial burdens, disease severity, and TB treatment outcomes in TB patients with or without anemia. Participants were recruited from Chennai, South India, as part of the prospective Effect of Diabetes on Tuberculosis Severity (EDOTS) study conducted from February 2014 to August 2018. Anemia was defined as hemoglobin (Hb) levels &amp;lt;13 g/dL and &amp;lt;12 g/dL for males and females, respectively. We employed chest X-rays to assess bilateral lung and cavitary diseases and sputum smear grades to measure bacterial loads in TB subjects. Treatment outcomes were defined as favorable or unfavorable. Cytokine profile was measured using multiplex ELISA.ResultsThe study comprised of 483 culture-confirmed TB individuals, with 288 positives for anemia {Median Hb was 11.0 [interquartile range (IQR)], 10.3–12.3} and 195 negatives [Median Hb was 14.3 (IQR), 13.5–15.2]. The study revealed that TB patients with anemia had significantly higher bacterial loads [adjusted prevalence ratio (aPR), 4.01; 95% CI, 2.22–6.63; p &amp;lt; 0.001], cavitary lung lesions [aPR, 3.36; 95% CI, 1.95–5.68; p &amp;lt; 0.001] and unfavorable treatment outcomes [aPR, 1.61; 95% CI, 1.31–2.19; p = 0.046] compared to those without anemia. Our data also show that TB is associated with significantly lower levels of type-1 cytokines (IFNγ and IL-2) but significantly higher levels of pro-inflammatory cytokines (IL-6, IFNα, and IFNβ) and pro-fibrotic factors (VEGF, EGF, FGF-2, and PDGF-AB/BB) in anemic individuals compared to those without anemia.ConclusionsThese findings highlight a clear association between anemia and increased TB severity, elevated bacterial loads, and poor treatment outcomes. Our data also suggest that anemia might be associated with the modulation of cytokine responses, which could impart a detrimental effect on TB pathogenesis.

Typical and Local Diagnostic Reference Levels for Chest and Abdomen Radiography Examinations in Dubai Health Sector

Chest and abdomen radiographs are the most common radiograph examinations conducted in the Dubai Health sector, with both involving exposure to several radiosensitive organs. Diagnostic reference levels (DRLs) are accepted as an effective safety, optimization, and auditing tool in clinical practice. The present work aims to establish a comprehensive projection and weight-based structured DRL system that allows one to confidently highlight healthcare centers in need of urgent action. The data of a total of 5474 adult males and non-pregnant females who underwent chest and abdomen radiography examinations in five different healthcare centers were collected and retrospectively analyzed. The typical DRL (TDRL) for each healthcare center was established and defined per projection (chest: posterior–anterior (PA), anterior–posterior (AP) and lateral (LAT); abdomen: erect and supine) for a weight band (60–80 kg) and for the whole data (no weight band). Local DRL (LDRL) values were established per project for the selected radiograph for the whole data (no weight band) and the 60–80 kg population. Chest radiography data from 1755 (60–80 kg) images were used to build this comprehensive DRL system (PA: 1471, AP: 252, and LAT: 32). Similarly, 611 (60–80 kg) abdomen radiographs were used to establish a DRL system (erect: 286 and supine: 325). The LDRL values defined per chest and abdomen projection for the weight band group (60–80 kg) were as follows: chest—0.51 PA, 2.46 AP, and 2.13 LAT dGy·cm2; abdomen—8.08 for erect and 5.95 for supine dGy·cm2. The LDRL defined per abdomen projection for the 60–80 kg weight band highlighted at least one healthcare center in need of optimization. Such a system is efficient, easy to use, and very effective clinically.

ZooCNN: A Zero-Order Optimized Convolutional Neural Network for Pneumonia Classification Using Chest Radiographs

Pneumonia, a leading cause of mortality in children under five, is usually diagnosed through chest X-ray (CXR) images due to its efficiency and cost-effectiveness. However, the shortage of radiologists in the Least Developed Countries (LDCs) emphasizes the need for automated pneumonia diagnostic systems. This article presents a Deep Learning model, Zero-Order Optimized Convolutional Neural Network (ZooCNN), a Zero-Order Optimization (Zoo)-based CNN model for classifying CXR images into three classes, Normal Lungs (NL), Bacterial Pneumonia (BP), and Viral Pneumonia (VP); this model utilizes the Adaptive Synthetic Sampling (ADASYN) approach to ensure class balance in the Kaggle CXR Images (Pneumonia) dataset. Conventional CNN models, though promising, face challenges such as overfitting and have high computational costs. The use of ZooPlatform (ZooPT), a hyperparameter finetuning strategy, on a baseline CNN model finetunes the hyperparameters and provides a modified architecture, ZooCNN, with a 72% reduction in weights. The model was trained, tested, and validated on the Kaggle CXR Images (Pneumonia) dataset. The ZooCNN achieved an accuracy of 97.27%, a sensitivity of 97.00%, a specificity of 98.60%, and an F1 score of 97.03%. The results were compared with contemporary models to highlight the efficacy of the ZooCNN in pneumonia classification (PC), offering a potential tool to aid physicians in clinical settings.

Single-cell transcriptome analysis reveals cellular reprogramming and changes of immune cell subsets following tetramethylpyrazine treatment in LPS-induced acute lung injury

Background Acute lung injury (ALI) is a disordered pulmonary disease characterized by acute respiratory insufficiency with tachypnea, cyanosis refractory to oxygen and diffuse alveolar infiltrates. Despite increased research into ALI, current clinical treatments lack effectiveness. Tetramethylpyrazine (TMP) has shown potential in ALI treatment, and understanding its effects on the pulmonary microenvironment and its underlying mechanisms is imperative. Methods We established a mouse model of lipopolysaccharide (LPS)-induced ALI and performed single cell RNA sequencing (scRNA-seq). Bioinformatic analyses of the immune, epithelial and endothelial cells were then performed to explore the dynamic changes of the lung tissue microenvironment. We also analyzed the effects of TMP on the cell subtypes, differential gene expression and potential regulation of transcriptional factors involved. Immunohistochemistry and enzyme-linked immunosorbent assay were performed to identify the effects of TMP on immune inflammatory response. Results We found that TMP efficiently protected against LPS-induced acute lung injury. Results of scRNA-seq showed that the cells were divided into seven major cell clusters, including immune cells, fibroblasts, endothelial cells and epithelial cells. Neither dexamethasone (Dex) nor TMP treatment showed any significant protective effects in these clusters. However, TMP treatment in the LPS-induced ALI model significantly increased follicular helper T cells and reduced CD8+ naive T cells, Vcan-positive monocytes and Siva-positive NK cells. In addition, TMP treatment increased the number of basal epithelial cells and lymphatic endothelial cells (LECs), indicating its protective effects on these cell types. Scenic analysis suggested that TMP likely mitigates LPS-induced injury in epithelial and endothelial cells by promoting FOSL1 in basal epithelial cells and JunB in LECs. Conclusions Our findings suggest that TMP appears to alleviate LPS-induced lung injury by regulating the immune response, promoting epithelial cell survival and boosting the antioxidant potential of endothelial cells. This study highlights the potential therapeutic use of TMP in the management of ALI.

Using the Regression Slope of Training Loss to Optimize Chest X-ray Generation in Deep Convolutional Generative Adversarial Networks.

Diffusion models, variational autoencoders, and generative adversarial networks (GANs) are three common types of generative artificial intelligence models for image generation. Among these, GANs are the most frequently used for medical image generation and are often employed for data augmentation in various studies. However, due to the adversarial nature of GANs, where the generator and discriminator compete against each other, the training process can sometimes end with the model unable to generate meaningful images or even producing noise. This phenomenon is rarely discussed in the literature, and no studies have proposed solutions to address this issue. Such outcomes can introduce significant bias when GANs are used for data augmentation in medical image training. Moreover, GANs often require substantial computational power and storage, adding to the challenges. In this study, we used deep convolutional GANs for chest X-raygeneration, and three typical training outcomes were found. Two scenarios generated meaningful medical images and one failed to produce usable images. By analyzing the loss history during training, we observed that the regression line of the overall losses tends to diverge slowly. After excluding outlier losses, we found that the slope of the regression line within the stable loss segment indicates the optimal point to terminate training, ensuring the generation of meaningful medical images.

Co-creation of a gender responsive TB intervention in Nigeria: a researcher-led collaborative study

BackgroundIn Nigeria, men constitute over half of the people notified with tuberculosis (TB), experience longer delays before reaching care, and are estimated to account for two thirds of people who miss out on care. The higher TB risk and burden in men has implications for the whole population and reaching them earlier with TB services will reduce onward transmission in households, communities, and workplaces. The absence of a comprehensive guidance and the lack of substantial empirical evidence on TB care approaches that are responsive to the needs of men in Nigeria exacerbates this problem. Therefore, this research aimed to co-create a gender-responsive intervention for men in peri-urban communities in Nigeria.MethodsOur study utilised a researcher-led collaborative approach to engage local TB stakeholders including communities adversely affected by the disease to co-create a gender-responsive TB intervention. Between March and November 2022, we engaged 13 local TB stakeholders in a three-phase participatory intervention design process. This engagement involved two iterative cycles of Delphi research online, and an in-person workshop. In the first and second phases, participants described the potential impact of 15 listed interventions and prioritised combinations of nine interventions deemed to be effective in overcoming identified gendered barriers. Responses were analysed using a combination of qualitative framework approach, content analysis, and summary descriptive statistics assisted by NVivo software. Stakeholder consensus on a preferred intervention package was reached during the participatory workshop.ResultsOverall, participants prioritised approaches that sought to actively find and systematically screen men for TB including awareness creation as a crucial component. The stakeholders placed significant considerations on the synergy between interventions and their programmatic sustainability when making their final choices. Consequently, a complex intervention package comprising three components was developed. These included targeted awareness creation among men in communities; TB screening in male-dominated socio-cultural congregate settings; and the use of digital chest X-ray screening. Anticipated early outputs of this intervention included improved TB knowledge, increased care-seeking, reduced TB-related costs and TB stigma, and accelerated early diagnosis among men in Nigeria.ConclusionLeveraging the insights and experiences of local stakeholders through iterative engagements yielded consensus on a viable gender-responsive TB intervention.

A multimodal similarity-aware and knowledge-driven pre-training approach for reliable pneumoconiosis diagnosis

Background Pneumoconiosis staging is challenging due to the low clarity of X-ray images and the small, diffuse nature of the lesions. Additionally, the scarcity of annotated data makes it difficult to develop accurate staging models. Although clinical text reports provide valuable contextual information, existing works primarily focus on designing multimodal image-text contrastive learning tasks, neglecting the high similarity of pneumoconiosis imaging representations. This results in inadequate extraction of fine-grained multimodal information and underutilization of domain knowledge, limiting their application in medical tasks. Objective The study aims to address the limitations of current multimodal methods by proposing a new approach that improves the precision of pneumoconiosis diagnosis and staging through enhanced fine-grained learning and better utilization of domain knowledge. Methods The proposed Multimodal Similarity-aware and Knowledge-driven Pre-Training (MSK-PT) approach involves two stages. In the first stage, we deeply analyze the similar features of pneumoconiosis images and use a similarity-aware modality alignment strategy to explore the fine-grained representations and associated disturbances of pneumoconiosis lesions between images and texts, guiding the model to match more appropriate feature representations. In the second stage, we utilize data-associated features and pre-stored domain knowledge features as priors and constraints to guide the downstream model in the visual domain without annotations. To address potential erroneous labels generated by model predictions, we further introduce an uncertainty threshold strategy to mitigate the negative impact of imperfect prediction labels and enhance model interpretability. Results We collected and created the pneumoconiosis chest X-ray (PneumoCXR) dataset to evaluate our proposed MSK-PT method. The experimental results show that our method achieved a classification accuracy of 81.73%, outperforming the state-of-the-art algorithms by 2.53%. Conclusions MSK-PT showed diagnostic performance that matches or exceeds the average radiologist's level, even with limited labeled data, highlighting the method's effectiveness and robustness.

Chest Radiography Scores for Predicting the Severity of Respiratory Status in Newborns Weighing More Than 1,500 g

Chest Radiography Scores for Predicting the Severity of Respiratory Status in Newborns Weighing More Than 1,500 g

Development of an image quality evaluation system for bedside chest X-ray images using scatter correction processing.

In plain radiography, scattered X-ray correction processing (Virtual Grid: VG) is used to estimate and correct scattered rays in images. We developed an objective evaluation system for bedside chest X-ray images using VG and investigated its usefulness. First, we trained the blind/referenceless image spatial quality evaluator (BRISQUE) on 200 images obtained by portable chest radiography. We then evaluated optimal chest phantom VG images as well as those that deviated from the VG setting conditions using BRISQUE. Furthermore, we conducted a subjective evaluation using the mean opinion score (MOS) and established an objective evaluation system for VG images. Finally, the degree of agreement between the MOS subjectively evaluated by 14 radiological technologists and that determined by the objective evaluation system for 100 clinical images obtained by portable chest radiography was calculated using Cohen's kappa coefficient. The correlation coefficient between the BRISQUE score and MOS for chest phantom images was - 0.96 (p < 0.05). The two scores showed a very high linear correlation, indicating the potential of the BRISQUE score as an alternative to MOS. The Cohen's kappa coefficient for the objective evaluation system using the optimal conversion table was 0.42. Conversely, there was a very high detection rate of 82.86% for poor-quality images. An objective evaluation system for bedside chest X-ray images using VG that uses no-reference image quality evaluation helps provide proper image quality. Furthermore, such a system can be constructed with a small amount of training data, which increases the possibility of introducing it to a variety of facilities.