Biomarkers of Distinguishing Neutrophil-Predominant Tuberculous Pleural Effusion from Parapneumonic Pleural Effusion

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.

Characteristics of pleural effusion with a high adenosine deaminase level: a case-control study

Background

Increased pleural fluid adenosine deaminase (ADA) is useful for diagnosing tuberculous pleurisy (TB), but high ADA levels are associated with other diseases. In this study, we compare various disease characteristics in patients with high-ADA pleural effusion.

Methods

We retrospectively collected data for 456 patients with pleural fluid ADA levels of ≥ 40 U/L from January 2012 to October 2021. Cases were classified as TB (n = 203), pleural infection (n = 112), malignant pleural effusion (n = 63), nontuberculous mycobacteria (n = 22), malignant lymphoma (ML) (n = 18), autoimmune diseases (n = 11), and other diseases (n = 27), and data were compared among those diseases. Predictive factors were identified by comparing data for a target disease to those for all other diseases. A diagnostic flowchart for TB was developed based on those factors.

Results

The most frequent disease was TB, though 60.0% of patients were diagnosed with other diseases. Median ADA levels in patients with TB were 83.1 U/L (interquartile range [IQR] 67.2–104.1), higher than those of patients with pleural infection (median 60.9 [IQR 45.3–108.0], p = 0.004), malignant pleural effusion (median 54.1 [IQR 44.8–66.7], p < 0.001), or autoimmune diseases (median 48.5 [IQR 45.9–58.2], p = 0.008), with no significant difference from NTM (p = 1.000) or ML (p = 1.000). Pleural fluid lactate dehydrogenase (LDH) levels of < 825 IU/L were beneficial for the diagnosis of TB. Neutrophil predominance or cell degeneration, white blood cell count of ≥ 9200/µL or C-reactive protein levels of ≥ 12 mg/dL helped in diagnosing pleural infection. Pleural fluid amylase levels of ≥ 75 U/L and a pleural fluid ADA/total protein (TP) ratio of < 14 helped in diagnosing malignant pleural effusion. High serum LDH and high serum/pleural fluid eosinophils helped in diagnosing ML and autoimmune diseases, respectively. The flowchart was comprised of the following three factors: pleural fluid LDH < 825 IU/L, pleural fluid ADA/TP of < 14, and neutrophil predominance or cell degeneration, which were decided by a decision tree. The diagnostic accuracy rate, sensitivity, and specificity for the diagnosis of TB were 80.9%, 78.8%, and 82.6%, respectively.

Conclusion

Cases involving high pleural fluid ADA levels should be investigated using several factors to distinguish TB from other diseases.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12890-022-02150-4.