A Prospective Observational Study on the Role of Transthoracic Ultrasound in Differentiating Tuberculous From Malignant Pleural Effusion

Tuberculous (TBE) and malignant pleural effusion (MPE) is the commonest cause of exudative pleural effusion in Malaysia. Early differentiation between these two diagnoses is essential as TBE only requires drainage if symptomatic, whereas MPE would require tissue biopsy for diagnosis confirmation and molecular profiling. However, both TBE and MPE present almost similarly with lymphocytic exudates. Pleural fluid (Pf) indices such as adenosine deaminase (ADA) may allow differentiation between these two entities in appropriate clinical circumstances. However, Pf ADA may not readily be accessible in resource-limited regions and the optimal cut-off varies depending on the local prevalence of tuberculosis. As a result, TBE diagnosis in our region is still heavily dependent on the analysis of pre-existing clinical demographic data and Pf parameters, where ultimately requiring pleural biopsy for a confident clinical diagnosis.

Point-of-care predictors for TBE, such as ultrasound imaging appearance, may be helpful, but have rarely been described. Previous studies have demonstrated that a complex septated ultrasound pattern in lymphocytic pleural effusion is a potentially useful diagnostic predictor to differentiate TBE from MPE with a positive predictive value of 94% and likelihood ratio of 12.2 TBE diagnostic algorithms frequently include only clinical indices with Pf parameters such as Pf differential cell count, ADA, protein and lactate dehydrogenase (LDH). Incorporation of point-of-care ultrasound finding into the clinical diagnostic algorithm has not been extensively explored.

TBE is the result of a delayed type IV hypersensitivity reaction to mycobacterial protein; fibrin formation in the pleural cavity is largely driven by pro-inflammatory cytokines as well as a reduction of fibrinolytic activity due to pleural inflammation. In contrast, MPE is believed to be driven by a high degree of anaerobic metabolism leading to lactic acid production, rather than an inflammatory response. These different pathogenic mechanisms in TBE and MPE resulted in different pleural fluid parameters such as lower Pf glucose and pH with higher Pf LDH level in MPE when compared to TBE. We believed that this principle may be extrapolated to discriminative ultrasound findings between TBE and MPE.

The result from our retrospective pilot study found that the presence of echographic septation had an adjusted odds ratio of 9.28 in prediction of TBE diagnosis from MPE. Along with other clinical parameters (male gender, serum leucocyte counts 9 x 109/L or, pleural fluid protein 50g/L or more), these parameters collectively report a diagnostic accuracy of 79.61% (95% CI 74.13-84.38) for TBE. In a previous study conducted in region with low tuberculosis burden, pleural thickening of >1cm, pleural nodularity and diaphragmatic thickening of >7mm on transthoracic ultrasound were highly suggestive of MPE. However, not uncommonly, we observed similar pattern of pleural and diaphragmatic thickening in TBE patients in our region as well.

As TBE is a hypersensitivity process with significant inflammatory response, we hypothesize that echographic septation, in addition to pleural thickening and other sonographic findings, may be a good indicator, as part of a clinical prediction model to discriminate TBE from MPE in our region.