Management of Malignant Pleural Effusion: Where Are We Now?

Abstract

Pleural malignancies are among the most common causes of pleural disease and form the basis of our daily pleural practice. There has been significant research and increase in both diagnostic and therapeutic management of malignant pleural diseases in the last decade. Good-quality data have led to a paradigm shift in the management options of pleural malignancies, and indwelling pleural catheter is now recommended and widely used as first-line intervention. Several trials compared different treatment modalities for pleural malignancies and continue to emphasize the need to reduce hospital length of stay and unnecessary pleural intervention, and the importance of patient choice in clinical decision making. This practical review aims to summarize the current knowledge for the management of pleural malignancies, and the understanding of the steps that we still have to climb to optimize management and reduce morbidity.

The Diagnosis of Malignant Pleural Effusion Using Tumor-Marker Combinations: A Cost-Effectiveness Analysis Based on a Stacking Model

“…Utilizing the strong association between tumor markers and malignant tumors, the analysis of tumor-marker concentrations has been recognized as a fast and less invasive method to diagnose MPE. Clinically, tumor markers such as carcinoembryonic antigen (CEA), carbohydrate antigen (CA)19-9, CA125, and CA15-3 have been frequently employed, and show satisfying diagnostic accuracy [7][8][9][10][11]. Therefore, using cut-off values of tumor-marker concentrations to diagnose MPE has been popular in past years [6,7,10,11].…”

“…Clinically, tumor markers such as carcinoembryonic antigen (CEA), carbohydrate antigen (CA)19-9, CA125, and CA15-3 have been frequently employed, and show satisfying diagnostic accuracy [7][8][9][10][11]. Therefore, using cut-off values of tumor-marker concentrations to diagnose MPE has been popular in past years [6,7,10,11]. To handle the complex associations between tumor markers, machine learning methods, such as logistic regression, support vector machine (SVM), random forest, etc., have been widely used in this field and show good performance [12][13][14].…”

“…Moreover, many studies have found minimal differences in AUC between single markers and multi-marker combinations. Notably, the diagnostic accuracy of multi-marker combinations does not always rise with the inclusion of more tumor markers [10,11,14]. Therefore, the cost-effectiveness issue, especially the necessity of including many tumor markers in diagnosing MPE, needs to be evaluated to give comprehensive recommendations on tumor-marker selection.…”