Identifying tuberculous pleural effusion using artificial intelligence machine learning algorithms

Enhanced differential evolution algorithm for feature selection in tuberculous pleural effusion clinical characteristics analysis

Tuberculous pleural effusion poses a significant threat to human health due to its potential for severe disease and mortality. Without timely treatment, it may lead to fatal consequences. Therefore, early identification and prompt treatment are crucial for preventing problems such as chronic lung disease, respiratory failure, and death. This study proposes an enhanced differential evolution algorithm based on colony predation and dispersed foraging strategies. A series of experiments conducted on the IEEE CEC 2017 competition dataset validated the global optimization capability of the method. Additionally, a binary version of the algorithm is introduced to assess the algorithm's ability to address feature selection problems. Comprehensive comparisons of the effectiveness of the proposed algorithm with 8 similar algorithms were conducted using public datasets with feature sizes ranging from 10 to 10,000. Experimental results demonstrate that the proposed method is an effective feature selection approach. Furthermore, a predictive model for tuberculous pleural effusion is established by integrating the proposed algorithm with support vector machines. The performance of the proposed model is validated using clinical records collected from 140 tuberculous pleural effusion patients, totaling 10,780 instances. Experimental results indicate that the proposed model can identify key correlated indicators such as pleural effusion adenosine deaminase, temperature, white blood cell count, and pleural effusion color, aiding in the clinical feature analysis of tuberculous pleural effusion and providing early warning for its treatment and prediction.

Quantitative proteomics revealed protein biomarkers to distinguish malignant pleural effusion from benign pleural effusion

To identify protein biomarkers capable of early prediction regarding the distinguishing malignant pleural effusion (MPE) from benign pleural effusion (BPE) in patients with lung disease. A four-dimensional data independent acquisition (4D-DIA) proteomic was performed to determine the differentially expressed proteins in samples from 20 lung adenocarcinoma MPE and 30 BPE. The significantly differential expressed proteins were selected for Gene Ontology (GO) enrichment and Kyoto Encyclopaedia of Genes and Genomes (KEGG) pathway analysis. Protein biomarkers with high capability to discriminate MPE from BPE patients were identified by Random Forest (RF) algorithm prediction model, whose diagnostic and prognostic efficacy in primary tumors were further explored in public datasets, and were validated by ELISA experiment. 50 important proteins (30 up-regulated and 20 down-regulated) were selected out as potential markers to distinguish the MPE from BPE group. GO analysis revealed that those proteins involving the most important cell component is extracellular space. KEGG analysis identified the involvement of cellular adhesion molecules pathway. Furthermore, the Area Under Curve (AUC) of these proteins were ranged from 0.717 to 1.000，with excellent diagnostic properties to distinguish the MPE. Finally, significant survival and gene and protein expression analysis demonstrated BPIFB1, DPP4, HPRT1 and ABI3BP had high discriminating values. SIGNIFICANCE: We performed a 4D-DIA proteomics to determine the differentially expressed proteins in pleural effusion samples from MPE and BPE. Some potential protein biomarkers were identified to distinguish the MPE from BPE patients., which may provide helpful diagnostic and therapeutic insights for lung cancer. This is significant because the median survival time of patients with MPE is usually 4-12 months, thus, it is particularly important to diagnose MPE early to start treatments promptly. The most common causes of MPE are lung cancers, while pneumonia and tuberculosis are the main causes of BPE. If more diagnostic markers could be identified periodically, there would be an important significance to clinical diagnose and treatment with drugs in lung cancer patients.

Differential Diagnosis of Pleural Effusion Using Machine Learning

Rationale: Differential diagnosis of pleural effusion is challenging in clinical practice. Objectives: We aimed to develop a machine learning model to classify the five common causes of pleural effusions. Methods: This retrospective study collected 49 features from clinical information, blood, and pleural fluid of adult patients who underwent diagnostic thoracentesis between October 2013 and December 2018. Pleural effusions were classified into the following five categories: transudative, malignant, parapneumonic, tuberculous, and other. The performance of five different classifiers, including multinomial logistic regression, support vector machine, random forest, extreme gradient boosting, and light gradient boosting machine (LGB), was evaluated in terms of accuracy and area under the receiver operating characteristic curve through fivefold cross-validation. Hybrid feature selection was applied to determine the most relevant features for classifying pleural effusion. Results: We analyzed 2,253 patients (training set, n = 1,459; validation set, n = 365; extra-validation set, n = 429) and found that the LGB model achieved the best performance in both validation and extra-validation sets. After feature selection, the accuracy of the LGB model with the selected 18 features was equivalent to that with all 49 features (mean ± standard deviation): 0.818 ± 0.012 and 0.777 ± 0.007 in the validation and extra-validation sets, respectively. The model's mean area under the receiver operating characteristic curve was as high as 0.930 ± 0.042 and 0.916 ± 0.044 in the validation and extra-validation sets, respectively. In our model, pleural lactate dehydrogenase, protein, and adenosine deaminase levels were the most important factors for classifying pleural effusions. Conclusions: Our LGB model showed satisfactory performance for differential diagnosis of the common causes of pleural effusions. This model could provide clinicians with valuable information regarding the major differential diagnoses of pleural diseases.

Diagnostic and comparative performance for the prediction of tuberculous pleural effusion using machine learning algorithms

OBJECTIVE

Early diagnosis and differential diagnosis of tuberculous pleural effusion (TPE) remains challenging and is critical to the patients' prognosis. The present study aimed to develop nine machine learning (ML) algorithms for early diagnosis of TPE and compare their performance.

METHODS

A total of 1435 untreated patients with pleural effusions (PEs) were retrospectively included and divided into the training set (80%) and the test set (20%). The demographic and laboratory variables were collected, preprocessed, and analyzed to select features, which were fed into nine ML algorithms to develop an optimal diagnostic model for TPE. The proposed model was validated by an independently external data. The decision curve analysis (DCA) and the SHapley Additive exPlanations (SHAP) were also applied.

RESULTS

Support vector machine (SVM) was the best model in discriminating TPE from non-TPE, with a balanced accuracy of 87.7%, precision of 85.3%, area under the curve (AUC) of 0.914, sensitivity of 94.7%, specificity of 80.7%, and F1-score of 86.0% among the nine ML algorithms. The excellent diagnostic performance was also validated by the external data (a balanced accuracy of 87.7%, precision of 85.2%, and AUC of 0.898). Neural network (NN) and K-nearest neighbor (KNN) had better net benefits in clinical usefulness. Besides, PE adenosine deaminase (ADA), PE carcinoembryonic antigen (CEA), and serum CYFRA21-1 were identified as the top three important features for diagnosing TPE.

CONCLUSIONS

This study developed and validated a SVM model for the early diagnosis of TPE, which might help clinicians provide better diagnosis and treatment for TPE patients.

Transforming cancer clinical trials: The integral role of artificial intelligence in electronic health records for efficient patient recruitment

Healthcare is one of the sectors where artificial intelligence (AI) is currently viewed as a crucial driving factor. Patient care, medical research, and clinical trial enrollment could all significantly improve due to AI's incorporation into electronic health records (EHRs). This short communication highlights how AI may improve the recruitment process regarding speed, accuracy, and overall cancer clinical trial efficiency. AI can automate this procedure by utilizing machine learning (ML) algorithms, identifying potential trial participants quickly and precisely. Many challenges could be addressed due to this integration, including data privacy and security that can be resolved through cutting-edge encryption techniques and differential privacy algorithms that ensure data anonymization. Another significant obstacle is the lack of common EHR formats and interoperability that can be addressed by creating a standardized structured layout. Automating and improving recruitment processes with AI may speed up research, increase the effectiveness of clinical trials, and open the door to more specialized cancer treatments.

Differentiation of malignant from benign pleural effusions based on artificial intelligence

INTRODUCTION

This study aimed to construct artificial intelligence models based on thoracic CT images to perform segmentation and classification of benign pleural effusion (BPE) and malignant pleural effusion (MPE).

METHODS

A total of 918 patients with pleural effusion were initially included, with 607 randomly selected cases used as the training cohort and the other 311 as the internal testing cohort; another independent external testing cohort with 362 cases was used. We developed a pleural effusion segmentation model (M1) by combining 3D spatially weighted U-Net with 2D classical U-Net. Then, a classification model (M2) was built to identify BPE and MPE using a CT volume and its 3D pleural effusion mask as inputs.

RESULTS

The average Dice similarity coefficient, Jaccard coefficient, precision, sensitivity, Hausdorff distance 95% (HD95) and average surface distance indicators in M1 were 87.6±5.0%, 82.2±6.2%, 99.0±1.0%, 83.0±6.6%, 6.9±3.8 and 1.6±1.1, respectively, which were better than those of the 3D U-Net and 3D spatially weighted U-Net. Regarding M2, the area under the receiver operating characteristic curve, sensitivity and specificity obtained with volume concat masks as input were 0.842 (95% CI 0.801 to 0.878), 89.4% (95% CI 84.4% to 93.2%) and 65.1% (95% CI 57.3% to 72.3%) in the external testing cohort. These performance metrics were significantly improved compared with those for the other input patterns.

CONCLUSIONS

We applied a deep learning model to the segmentation of pleural effusions, and the model showed encouraging performance in the differential diagnosis of BPE and MPE.

Diagnosis of malignant pleural effusion with combinations of multiple tumor markers: A comparison study of five machine learning models

BACKGROUND

To evaluate the diagnostic value of combinations of tumor markers carcinoembryonic antigen (CEA), carbohydrate antigen (CA) 125, CA153, and CA19-9 in identifying malignant pleural effusion (MPE) from non-malignant pleural effusion (non-MPE) using machine learning, and compare the performance of popular machine learning methods.

METHODS

A total of 319 samples were collected from patients with pleural effusion in Beijing and Wuhan, China, from January 2018 to June 2020. Five machine learning methods including Logistic regression, extreme gradient boosting (XGBoost), Bayesian additive regression tree, random forest, and support vector machine were applied to evaluate the diagnostic performance. Sensitivity, specificity, Youden's index, and the area under the receiver operating characteristic curve (AUC) were used to evaluate the performance of different diagnostic models.

RESULTS

For diagnostic models with a single tumor marker, the model using CEA, constructed by XGBoost, performed best (AUC = 0.895, sensitivity = 0.80), and the model with CA153, also by XGBoost, showed the largest specificity 0.98. Among all combinations of tumor markers, the combination of CEA and CA153 achieved the best performance (AUC = 0.921, sensitivity = 0.85) in identifying MPE under the diagnostic model constructed by XGBoost.

CONCLUSIONS

Diagnostic models for MPE with a combination of multiple tumor markers outperformed the models with a single tumor marker, particularly in sensitivity. Using machine learning methods, especially XGBoost, could comprehensively improve the diagnostic accuracy of MPE.

DNA Methylation Analysis of the SHOX2 and RASSF1A Panel Using Cell-Free DNA in the Diagnosis of Malignant Pleural Effusion

Objectives

The differential diagnosis of pleural effusion (PE) is a common but major challenge in clinical practice. This study aimed to establish a strategy based on a PE-cell-free DNA (cfDNA) methylation detection system for the differential diagnosis of malignant pleural effusion (MPE) and benign pleural effusion (BPE).

Methods

A total of 104 patients with PE were enrolled in this study, among which 50 patients had MPE, 9 malignant tumor patients had PE of indefinite causes, and the other 45 patients were classified as benign controls. The methylation status of short stature homeobox 2 (SHOX2) and RAS association domain family 1, isoform A (RASSF1A) was detected using PE-cfDNA specimens by real-time fluorescence quantitative PCR. Total methylation (TM) was defined as the combination of the methylation levels of SHOX2 and RASSF1A. The electrochemiluminescence immunoassay was applied to evaluate the levels of multiple serum tumor markers.

Results

The PE-cfDNA methylation status of either SHOX2 or RASSF1A was much higher in MPE samples than in benign controls. The combination of SHOX2 and RASSF1A methylation in PE yielded a diagnostic sensitivity of 96% and a specificity of 100%, respectively. When compared with the corresponding serum tumor marker detection results, TM showed the highest diagnostic efficiency (AUC = 0.985). Furthermore, the combination of the SHOX2 and RASSF1A methylation panels using PE-cfDNA could apparently improve the differential diagnostic efficacy of BPE and MPE and could help compensate for the deficiency of cytology.

Conclusions

Our results indicated that SHOX2 and RASSF1A methylation panel detection could accurately classify BPE and MPE diseases and showed better diagnostic performance than traditional serum parameters. The SHOX2 and RASSF1A methylation detection of PE-cfDNA could be a potentially effective complementary tool for cytology in the process of differential diagnosis. In summary, PE-cfDNA could be used as a promising non-invasive analyte for the auxiliary diagnosis of MPE.

Diagnostic value of soluble biomarkers for parapneumonic pleural effusion

Parapneumonic pleural effusion (PPE) is a common complication in patients with pneumonia. Timely and accurate diagnosis of PPE is of great value for its management. Measurement of biomarkers in circulating and pleural fluid have the advantages of easy accessibility, short turn-around time, objectiveness and low cost and thus have utility for PPE diagnosis and stratification. To date, many biomarkers have been reported to be of value for the management of PPE. Here, we review the values of pleural fluid and circulating biomarkers for the diagnosis and stratification PPE. The biomarkers discussed are C-reactive protein, procalcitonin, presepsin, soluble triggering receptor expressed on myeloid cells 1, lipopolysaccharide-binding protein, inflammatory markers, serum amyloid A, soluble urokinase plasminogen activator receptor, matrix metalloproteinases, pentraxin-3 and cell-free DNA. We found that none of the available biomarkers has adequate performance for diagnosing and stratifying PPE. Therefore, further work is needed to identify and validate novel biomarkers, and their combinations, for the management of PPE.

Development and validation of a machine learning model for differential diagnosis of malignant pleural effusion using routine laboratory data

BACKGROUND

The differential diagnosis of malignant pleural effusion (MPE) and benign pleural effusion (BPE) presents a clinical challenge. In recent years, the use of artificial intelligence (AI) machine learning models for disease diagnosis has increased.

OBJECTIVE

This study aimed to develop and validate a diagnostic model for early differentiation between MPE and BPE based on routine laboratory data.

DESIGN

This was a retrospective observational cohort study.

METHODS

A total of 2352 newly diagnosed patients with pleural effusion (PE), between January 2008 and March 2021, were eventually enrolled. Among them, 1435, 466, and 451 participants were randomly assigned to the training, validation, and testing cohorts in a ratio of 3:1:1. Clinical parameters, including age, sex, and laboratory parameters of PE patients, were abstracted for analysis. Based on 81 candidate laboratory variables, five machine learning models, namely extreme gradient boosting (XGBoost) model, logistic regression (LR) model, random forest (RF) model, support vector machine (SVM) model, and multilayer perceptron (MLP) model were developed. Their respective diagnostic performances for MPE were evaluated by receiver operating characteristic (ROC) curves.

RESULTS

Among the five models, the XGBoost model exhibited the best diagnostic performance for MPE (area under the curve (AUC): 0.903, 0.918, and 0.886 in the training, validation, and testing cohorts, respectively). Additionally, the XGBoost model outperformed carcinoembryonic antigen (CEA) levels in pleural fluid (PF), serum, and the PF/serum ratio (AUC: 0.726, 0.699, and 0.692 in the training cohort; 0.763, 0.695, and 0.731 in the validation cohort; and 0.722, 0.729, and 0.693 in the testing cohort, respectively). Furthermore, compared with CEA, the XGBoost model demonstrated greater diagnostic power and sensitivity in diagnosing lung cancer-induced MPE.

CONCLUSION

The development of a machine learning model utilizing routine laboratory biomarkers significantly enhances the diagnostic capability for distinguishing between MPE and BPE. The XGBoost model emerges as a valuable tool for the diagnosis of MPE.

Pleural fluid biochemical analysis: the past, present and future

Identifying the cause of pleural effusion is challenging for pulmonologists. Imaging, biopsy, microbiology and biochemical analyses are routinely used for diagnosing pleural effusion. Among these diagnostic tools, biochemical analyses are promising because they have the advantages of low cost, minimal invasiveness, observer independence and short turn-around time. Here, we reviewed the past, present and future of pleural fluid biochemical analysis. We reviewed the history of Light’s criteria and its modifications and the current status of biomarkers for heart failure, malignant pleural effusion, tuberculosis pleural effusion and parapneumonic pleural effusion. In addition, we anticipate the future of pleural fluid biochemical analysis, including the utility of machine learning, molecular diagnosis and high-throughput technologies. Clinical Chemistry and Laboratory Medicine (CCLM) should address the topic of pleural fluid biochemical analysis in the future to promote specific knowledge in the laboratory professional community.

A scoring model for diagnosis of tuberculous pleural effusion

Background

Due to the low efficiency of a single clinical feature or laboratory variable in the diagnosis of tuberculous pleural effusion (TBPE), the diagnosis of TBPE is still challenging. This study aimed to build a scoring diagnostic model based on laboratory variables and clinical features to differentiate TBPE from non-tuberculous pleural effusion (non-TBPE).

Methods

A retrospective study of 125 patients (63 with TBPE; 62 with non-TBPE) was undertaken. Univariate analysis was used to select the laboratory and clinical variables relevant to the model composition. The statistically different variables were selected to undergo binary logistic regression. Variables B coefficients were used to define a numerical score to calculate a scoring model. A receiver operating characteristic (ROC) curve was used to calculate the best cut-off value and evaluate the performance of the model. Finally, we add a validation cohort to verify the model.

Results

Six variables were selected in the scoring model: Age ≤ 46 years old (4.96 points), Male (2.44 points), No cancer (3.19 points), Positive T-cell Spot (T-SPOT) results (4.69 points), Adenosine Deaminase (ADA) ≥ 24.5U/L (2.48 point), C-reactive Protein (CRP) ≥ 52.8 mg/L (1.84 points). With a cut-off value of a total score of 11.038 points, the scoring model’s sensitivity, specificity, and accuracy were 93.7%, 96.8%, and 99.2%, respectively. And the validation cohort confirms the model with the sensitivity, specificity, and accuracy of 92.9%, 93.3%, and 93.1%, respectively.

Conclusion

The scoring model can be used in differentiating TBPE from non-TBPE.

Diversity in Machine Learning: A Systematic Review of Text-Based Diagnostic Applications

OBJECTIVE

 As the storage of clinical data has transitioned into electronic formats, medical informatics has become increasingly relevant in providing diagnostic aid. The purpose of this review is to evaluate machine learning models that use text data for diagnosis and to assess the diversity of the included study populations.

METHODS

 We conducted a systematic literature review on three public databases. Two authors reviewed every abstract for inclusion. Articles were included if they used or developed machine learning algorithms to aid in diagnosis. Articles focusing on imaging informatics were excluded.

RESULTS

 From 2,260 identified papers, we included 78. Of the machine learning models used, neural networks were relied upon most frequently (44.9%). Studies had a median population of 661.5 patients, and diseases and disorders of 10 different body systems were studied. Of the 35.9% (N = 28) of papers that included race data, 57.1% (N = 16) of study populations were majority White, 14.3% were majority Asian, and 7.1% were majority Black. In 75% (N = 21) of papers, White was the largest racial group represented. Of the papers included, 43.6% (N = 34) included the sex ratio of the patient population.

DISCUSSION

 With the power to build robust algorithms supported by massive quantities of clinical data, machine learning is shaping the future of diagnostics. Limitations of the underlying data create potential biases, especially if patient demographics are unknown or not included in the training.

CONCLUSION

 As the movement toward clinical reliance on machine learning accelerates, both recording demographic information and using diverse training sets should be emphasized. Extrapolating algorithms to demographics beyond the original study population leaves large gaps for potential biases.

Development and Validation of a Scoring System for Early Diagnosis of Malignant Pleural Effusion Based on a Nomogram

Background

The diagnostic value of clinical and laboratory features to differentiate between malignant pleural effusion (MPE) and benign pleural effusion (BPE) has not yet been established.

Objectives

The present study aimed to develop and validate the diagnostic accuracy of a scoring system based on a nomogram to distinguish MPE from BPE.

Methods

A total of 1,239 eligible patients with PE were recruited in this study and randomly divided into a training set and an internal validation set at a ratio of 7:3. Logistic regression analysis was performed in the training set, and a nomogram was developed using selected predictors. The diagnostic accuracy of an innovative scoring system based on the nomogram was established and validated in the training, internal validation, and external validation sets (n = 217). The discriminatory power and the calibration and clinical values of the prediction model were evaluated.

Results

Seven variables [effusion carcinoembryonic antigen (CEA), effusion adenosine deaminase (ADA), erythrocyte sedimentation rate (ESR), PE/serum CEA ratio (CEA ratio), effusion carbohydrate antigen 19-9 (CA19-9), effusion cytokeratin 19 fragment (CYFRA 21-1), and serum lactate dehydrogenase (LDH)/effusion ADA ratio (cancer ratio, CR)] were validated and used to develop a nomogram. The prediction model showed both good discrimination and calibration capabilities for all sets. A scoring system was established based on the nomogram scores to distinguish MPE from BPE. The scoring system showed favorable diagnostic performance in the training set [area under the curve (AUC) = 0.955, 95% confidence interval (CI) = 0.942-0.968], the internal validation set (AUC = 0.952, 95% CI = 0.932-0.973), and the external validation set (AUC = 0.973, 95% CI = 0.956-0.990). In addition, the scoring system achieved satisfactory discriminative abilities at separating lung cancer-associated MPE from tuberculous pleurisy effusion (TPE) in the combined training and validation sets.

Conclusions

The present study developed and validated a scoring system based on seven parameters. The scoring system exhibited a reliable diagnostic performance in distinguishing MPE from BPE and might guide clinical decision-making.

Diagnostic accuracy of adenosine deaminase for pleural tuberculosis in a low prevalence setting: A machine learning approach within a 7-year prospective multi-center study

OBJECTIVE

To analyze the performance of adenosine deaminase in pleural fluid combined with other parameters routinely measured in clinical practice and assisted by machine learning algorithms for the diagnosis of pleural tuberculosis in a low prevalence setting, and secondly, to identify effusions that are non-tuberculous and most likely malignant.

PATIENTS AND METHODS

We prospectively analyzed 230 consecutive patients diagnosed with lymphocytic exudative pleural effusion from March 2013 to June 2020. Diagnosis according to the composite reference standard was achieved in all cases. Pre-test probability of pleural tuberculosis was 3.8% throughout the study period. Parameters included were: levels of adenosine deaminase, pH, glucose, proteins, and lactate dehydrogenase, red and white cell counts and lymphocyte percentage in pleural fluid, as well as age. We tested six different machine learning-based classifiers to categorize the patients. Two different classifications were performed: a) tuberculous/non-tuberculous and b) tuberculous/malignant/other.

RESULTS

Out of a total of 230 patients with pleural effusion included in the study, 124 were diagnosed with malignant effusion and 44 with pleural tuberculosis, while 62 were given other diagnoses. In the tuberculous/non-tuberculous classification, and taking into account the validation predictions, the support vector machine yielded the best result: an AUC of 0.98, accuracy of 97%, sensitivity of 91%, and specificity of 98%, whilst in the tuberculous/malignant/other classification, this type of classifier yielded an overall accuracy of 80%. With this three-class classifier, the same sensitivity and specificity was achieved in the tuberculous/other classification, but it also allowed the correct classification of 90% of malignant cases.

CONCLUSION

The level of adenosine deaminase in pleural fluid together with cell count, other routine biochemical parameters and age, combined with a machine-learning approach, is suitable for the diagnosis of pleural tuberculosis in a low prevalence scenario. Secondly, non-tuberculous effusions that are suspected to be malignant may also be identified with adequate accuracy.

Uncertainty-guided graph attention network for parapneumonic effusion diagnosis

Parapneumonic effusion (PPE) is a common condition that causes death in patients hospitalized with pneumonia. Rapid distinction of complicated PPE (CPPE) from uncomplicated PPE (UPPE) in Computed Tomography (CT) scans is of great importance for the management and medical treatment of PPE. However, UPPE and CPPE display similar appearances in CT scans, and it is challenging to distinguish CPPE from UPPE via a single 2D CT image, whether attempted by a human expert, or by any of the existing disease classification approaches. 3D convolutional neural networks (CNNs) can utilize the entire 3D volume for classification: however, they typically suffer from the intrinsic defect of over-fitting. Therefore, it is important to develop a method that not only overcomes the heavy memory and computational requirements of 3D CNNs, but also leverages the 3D information. In this paper, we propose an uncertainty-guided graph attention network (UG-GAT) that can automatically extract and integrate information from all CT slices in a 3D volume for classification into UPPE, CPPE, and normal control cases. Specifically, we frame the distinction of different cases as a graph classification problem. Each individual is represented as a directed graph with a topological structure, where vertices represent the image features of slices, and edges encode the spatial relationship between them. To estimate the contribution of each slice, we first extract the slice representations with uncertainty, using a Bayesian CNN: we then make use of the uncertainty information to weight each slice during the graph prediction phase in order to enable more reliable decision-making. We construct a dataset consisting of 302 chest CT volumetric data from different subjects (99 UPPE, 99 CPPE and 104 normal control cases) in this study, and to the best of our knowledge, this is the first attempt to classify UPPE, CPPE and normal cases using a deep learning method. Extensive experiments show that our approach is lightweight in demands, and outperforms accepted state-of-the-art methods by a large margin. Code is available at https://github.com/iMED-Lab/UG-GAT.

Pleural biomarkers in diagnostics of malignant pleural effusion: a narrative review

Although cytology and pleural biopsy of pleural effusion (PE) are the gold standards for diagnosing malignant pleural effusion (MPE), these tools’ diagnostic accuracy is plagued by some limitations such as low sensitivity, considerable inter-observer variation and invasiveness. The assessment of PE biomarkers may hence be seen as an objective and non-invasive diagnostic alternative in MPE diagnostics. In this review, we summarize the characteristics and diagnostic accuracy of available PE biomarkers, including carcinoembryonic antigen (CEA), neuron-specific enolase (NSE), carbohydrate antigens 125 (CA125), carbohydrate antigen 19-9 (CA19-9), carbohydrate antigen 15-3 (CA15-3), a fragment of cytokeratin 19 (CYFRA 21-1), chitinase-like proteins (CLPs), vascular endothelial growth factor (VEGF) and its soluble receptor, endostatin, calprotectin, cancer ratio, homocysteine, apolipoprotein E (Apo-E), B7 family members, matrix metalloproteinase (MMPs) and tissue-specific inhibitors of metalloproteinases (TIMPs), reactive oxygen species modulator 1 (Romo1), tumor-associated macrophages (TAMs) and monocytes, epigenetic markers (e.g., cell-free microRNA and mRNA). We summarized the evidence from systematic review and meta-analysis for traditional tumor markers’ diagnostic accuracy. According to the currently available evidence, we conclude that the traditional tumor markers have high specificity (around 0.90) but low sensitivity (around 0.50). The diagnostic accuracy of novel tumor markers needs to be validated by further studies. None of these tumor biomarkers would have sufficient diagnostic accuracy to confirm or exclude MPE when used alone. A multi-biomarker strategy, also encompassing the use of artificial intelligence algorithms, may be a valuable perspective for improving the diagnostic accuracy of MPE.

Driverless artificial intelligence framework for the identification of malignant pleural effusion

Our study aimed to explore the applicability of deep learning and machine learning techniques to distinguish MPE from BPE. We initially used a retrospective cohort with 726 PE patients to train and test the predictive performances of the driverless artificial intelligence (AI), and then stacked with a deep learning and five machine learning models, namely gradient boosting machine (GBM), extreme gradient boosting (XGBoost), extremely randomized trees (XRT), distributed random forest (DRF), and generalized linear models (GLM). Furthermore, a prospective cohort with 172 PE patients was applied to detect the external validity of the predictive models. The area under the curve (AUC) in the training, test and validation set were deep learning (0.995, 0.848, 0.917), GBM (0.981, 0.910, 0.951), XGBoost (0.933, 0.916, 0.935), XRT (0.927, 0.909, 0.963), DRF (0.906, 0.809, 0.969), and GLM (0.898, 0.866, 0.892), respectively. Although the Deep Learning model had the highest AUC in the training set (AUC = 0.995), GBM demonstrated stable and high predictive efficiency in three data sets. The final AI model by stacked ensemble yielded optimal diagnostic performance with AUC of 0.991, 0.912 and 0.953 in the training, test and validation sets, respectively. Using the driverless AI framework based on the routinely collected clinical data could significantly improve diagnostic performance in distinguishing MPE from BPE.

Development and validation of a novel scoring system developed from a nomogram to identify malignant pleural effusion

BACKGROUND

This study aimed to establish and validate a novel scoring system based on a nomogram for the differential diagnosis of malignant pleural effusion (MPE) and benign pleural effusion (BPE).

METHODS

Patients with PE and confirmed aetiology who underwent diagnostic thoracentesis were included in this study. One retrospective set (N = 1261) was used to develop and internally validate the predictive model. The clinical, radiological and laboratory features were collected and subjected to logistic regression analyses. The primary predictive model was displayed as a nomogram and then modified into a novel scoring system, which was externally validated in an independent set (N = 172).

FINDINGS

The novel scoring system was composed of fever (3 points), erythrocyte sedimentation rate (4 points), effusion adenosine deaminase (7 points), serum carcinoembryonic antigen (CEA) (4 points), effusion CEA (10 points) and effusion/serum CEA (8 points). With a cutoff value of 15 points, the area under the curve, specificity and sensitivity for identifying MPE were 0.913, 89.10%, and 82.63%, respectively, in the training set, 0.922, 93.48%, 81.51%, respectively, in the internal validation set and 0.912, 87.61%, 81.36%, respectively, in the external validation set. Moreover, this scoring system was exclusively applied to distinguish lung cancer with PE from tuberculous pleurisy and showed a favourable diagnostic performance in the training and validation sets.

INTERPRETATION

This novel scoring system was developed from a retrospective study and externally validated in an independent set based on six easily accessible clinical variables, and it exhibited good diagnostic performance for identifying MPE.

FUNDING

NFSC grants (no. 81572942, no. 81800094).