Can chest ultrasonography differentiate between benign and malignant effusions?

Background: The differential diagnosis of exudative pleural effusion is a major challenge for chest physicians particularly in a country with limited financial resources. Objective: The aim of this study was to evaluate the role of the sonographic features in the prediction of exudative malignant pleural effusion. Design: This was a prospective cross-sectional study. Setting: This study was carried out between May 2013 and June 2014 in the Chest Department of Assiut University Hospital. Patients: The patients enrolled included 25 patients with malignant pleural effusion and 25 patients with other different benign causes. Main outcome measures: The sonographic appearances of pleural effusions were defined in terms of five patterns: anechoic, complex septated, complex nonseptated, pleural thickening, and pleural nodules. Results: Among the 25 malignant exudative pleural effusions, a complex nonseptated pattern is a useful diagnostic predictor, with sensitivity, specificity, positive predictive value, and negative predictive value of 60, 68, 65, and 63%, respectively. Pleural nodules were only found in malignant effusion (100% specificity). If we define the complex septated sonographic pattern as a predictor for benign effusion, we can achieve sensitivity, specificity, positive predictive value, and negative predictive value of 52, 88, 81, and 65%, respectively. Conclusion: Pleural nodules and a complex nonseptated pattern in the sonographic appearance are useful predictors of malignant pleural effusions, whereas a complex septated pattern is a useful predictor in nonmalignant effusion.


Introduction
Pleural eff usion is a highly common clinical presentation in malignant and benign diseases. Th e diff erential diagnosis is broad and includes heart failure, parapneumonic eff usion, empyema, pulmonary emboli, infl ammatory disease, and malignancies. Th e diff erentiation between malignant and nonmalignant pleural eff usions has often been made with cytologic examinations of pleural eff usions, histologic examinations of pleural biopsies, helpful biomarkers, and even pleural biopsy [1][2][3].
With the advances in imaging technology and computerized functions, the chest sonographic examination has been used widely in the diagnosis and management of lung cancer, uncommon pulmonary consolidations, mediastinal tumors, and pleural diseases. Chest sonography is a very useful imaging tool for assessing the nature of pleural eff usions [4].

Aim of the work
Th e aim of this study was to evaluate the role of the sonographic features in the prediction of malignant pleural eff usion.

Patients and methods
We prospectively recruited patients with the diagnosis of malignant and nonmalignant exudative pleural eff usion from May 2013 to June 2014. Th e study was approved by the ethical committee of Assiut University. A total of 25 patients with malignant pleural eff usion and 25 patients with nonmalignant exudative pleural eff usion were enrolled. Th ere were 32 men and 18 women whose ages ranged from 23 to 75 years (mean 49 years).
All patients included in the study were subjected to the following: (1) Full assessment of medical history and clinical examination. (2) Complete blood count, blood urea and serum creatinine, liver, and renal function tests. (3) Radiological examination: chest radiography and multislice computer tomographic scan of the chest whenever needed. (4) Sputum cytology for the detection of infl ammatory or malignant cells. (5) Diagnostic thoracocentesis; about (300-500 ml) of pleural fl uid was aspirated for the following: (a) Chemical examination including protein level and lactate dehydrogenase level.

Statistical analysis
Sonographic appearances of each group were categorized and analyzed for comparison between each of them. Statistical analysis of the data was carried out using the SPSS 16 software package under (SPSS 16, IPM, Chicago, USA) the Windows7 operating system. Categorical data parameters were presented as frequency and percent. Quantitative data were expressed as mean and SD. Comparison was performed using the Z-test a nd χ 2 for categorical data and a paired t-test for quantitative data. Probability level (P-value) was assumed to be signifi cant if it was equal to or less than 0.05 and highly signifi cant if the P-value was equal to or less than 0.001.  (Tables 1 and 2).

Sonographic patterns between malignant and nonmalignant pleural effusions
As shown in Table 3, 25 malignant pleural eff usions had the following sonographic appearances: an anechoic pattern in 28% (7/25), a complex nonseptated pattern in 60% (15/25), and a complex septated pattern in 12% (b) Total and diff erential cell count and adenosine deaminase if tuberculosis was suspected.
(c) Bacteriological examination. (d) Cytological examination for malignant cells. (6) Pathological examination of the specimens obtained by bronchoscopy, closed pleural biopsy, thoracoscopy, or lymph node biopsy. (7) Other relevant investigations to make a fi nal diagnosis.

Procedure of chest ultrasound
A real-time ultrasound scanner, high-resolution linear probe was used to provide a description of chest wall, pleura, and lung. Pictures were saved in a JPG format. Th e size of eff usion was assessed by posteroanterior chest radiograph. Pleural eff usions were defi ned as minimal if they occupied less than 1/3 of the hemithorax, moderate if they occupied more than 1/3 and less than 2/3 of the hemithorax, and massive if they occupied more than 2/3 of the hemithorax.

Chest sonographic examination, and sonographic appearances
All patients underwent chest sonographic examination with the Aloka ultrasound machine in the Chest Department of Assiut University Hospital (Figs 1 and 2).

Discussion
Clinically, exudative eff usion is a challenge and a daily problem for the diff erential diagnosis between its diff erent etiologies, especially in an area with     sonographic septations in pleural eff usions and focused on whether septations in exudative pleural eff usions could actually be useful in predicting the infl ammatory nature of nonmalignant eff usion (empyema, tuberculosis, parapneumonic), particularly tuberculosis, as because of a higher infl ammatory reaction.
Th e complex septated sonographic patterns of pleural eff usions were not specifi c for tuberculosis. As reported, tuberculous pleural eff usions and empyema commonly had complex septated sonographic appearances [5,10,11]. However, the diff erentiation between empyema and tuberculosis can be made easily on the basis of clinical criteria. Th erefore, we studied diff erent varieties of nonmalignant pleural eff usion.
In our results, the complex septated sonographic pattern as a predictor for nonmalignant pleural eff usions had sensitivity, specifi city, positive predictive value, and negative predictive values of 52, 88, 81, and 65%, respectively. Th us, when complex septated sonographic appearances are found in exudative plural eff usions, the possibility of a nonmalignant cause of pleural eff usions is high. In tuberculosis, a sputum acid-fast stain may also be useful. If the diagnosis is not conclusive, pleural biopsy should be performed or antituberculous chemotherapy should be attempted, and a follow-up clinical course closely 2 to 4 weeks later [12].
Conversely, a complex nonseptated sonographic pattern was found in most of our cases with malignant eff usions. Sensitivity, specifi city, negative, and positive predictive values were 60, 68, 65, and 63%, respectively. Th erefore, repeated thoracentesis for eff usion cytologic examination, and sometimes transbronchial biopsy or sonographically guided transthoracic biopsy increase the diagnostic yield [13].
On reviewing the published literature, Philip-Joët et al. [14] and Hua et al. [15] reported that plasminogen activator inhibitor and von Willebrand factor levels were signifi cantly higher in patients with empyema or tuberculosis than in those with cancer or cardiac failure, and they also found increased levels of tissue-type plasminogen activators in some malignant pleural eff usions. Th is means that fi brinolytic activity is higher in malignant pleural eff usions than nonmalignant pleural eff usions, and confi rms our fi ndings that fi brins were less common in malignant pleural eff usions. In this study, only 12% of patients with malignant pleural eff usions had a complex septated sonographic appearance compared with 52% of patients with nonmalignant exudative eff usion. Certainly, one of the limitations of this study was the lack of study of the fi brinolytic activities in our enrolled patients. Another typical fi nding on limited fi nancial resources. Yang et al. [5] defi ned and classifi ed the sonographic patterns of pleural eff usions as anechoic, complex nonseptated, complex septated, and homogeneous. Transudates were often anechoic, whereas an anechoic pattern of eff usions could be either a transudate or an exudate. Pleural eff usions appearing with complex septated, complex nonseptated, or homogeneously echogenic patterns were almost always exudates [5]. Tu et al. [6] reported that complex nonseptated and relatively hyperechoic, complex septated, and homogeneously echogenic patterns of pleural eff usions might predict the possibility of empyema. Chian et al. [7] also published fi ndings in which the echogenic swirling pattern of pleural eff usions could predict malignant pleural eff usions in patients with malignancies.
From our study, we have found that complex septated sonographic patterns of exudative eff usion can aid the early diagnosis of nonmalignant eff usion. Certainly, our results require a larger sample size and multicenter studies for confi rmation of these fi ndings.
Fibrinous bands appear in various causes of pleuritis and divide the eff usions into a network of septa [8]. Chung et al. [9] found that repeated thoracentesis may cause pleural infl ammation and may lead to fi brin formation in malignant eff usions. Th us, we also analyzed the  ultrasound has been associated with malignancy is pleural nodules. In a study involving 54 patients with pleural eff usion [16], the presence of nodular thickening of the diaphragm was associated with malignancy in all cases [15 of 15].
Ultrasound has a sensitivity of 73% and a specifi city of 100% in distinguishing malignant pleural eff usions from other causes on the basis of pleural thickening, pleural nodularity, diaphragmatic thickening, and an echogenic swirling pattern visible in the pleural fl uid [3,4].
We found pleural nodules in 10 cases; all of them were malignant. Pleural nodules could predict malignant eff usion with 100% specifi city and positive predictive values. Pleural thickening was another signifi cant sonographic fi nding in 22 cases of malignant eff usion (P = 0.005) with 88% sensitivity and 80% specifi city.

Conclusion
Th e sonographic appearances of septations are often present in nonmalignant pleural eff usions, whereas absence of septations or the presence of pleural nodules and thickening are good predictors in malignant exudative pleural eff usions. Transthoracic ultrasound is useful in diff erentiating malignant from nonmalignant pleural exudates and may become an important step in the diagnostic pathw ay.