Pleural Effusions Related to Metastatic Malignancies
INCIDENCE
Malignant disease involving the pleura is the second leading cause of exudative pleural effusions after parapneumonic effusions. Because many parapneumonic effusions are small and are not subjected to thoracentesis, malignancy is probably the leading cause of exudative effusions subjected to thoracentesis. In our series from Baltimore, 42% of 102 exudative pleural effusions were due to malignant disease (1). In an epidemiologic study from the Czech Republic, malignancy accounted for 24% of all the pleural effusions (2).
Carcinomas of the lung and breast and lymphomas account for approximately 75% of malignant pleural effusions (Table 10.1). Metastatic ovarian carcinoma is the fourth leading cause of malignant pleural effusions, whereas sarcomas, particularly melanoma, account for a small percentage of malignant pleural effusions. No other single tumor accounts for more than 1% of malignant pleural effusions. In approximately 6% of patients with malignant pleural effusions, the primary tumor is not identified (3,4).
Lung Cancer
In most series, lung cancer is the leading cause of malignant pleural effusion (3). When patients with lung cancer are first evaluated, approximately 15% have a pleural effusion (5). During the course of this disease, however, at least 50% of patients with disseminated lung cancer develop a pleural effusion. Pleural effusions occur with all the cell types of lung carcinoma but appear to be most frequent with adenocarcinoma (6,7). Patients with small-cell lung carcinoma have a lower incidence of pleural effusion (˜15%) (8). Patients with lung cancer who have anti-p53 antibodies are more likely to have pleural effusions. In one series, 9 of 10 patients (90%) with this antibody had pleural effusions, whereas 42 of 115 patients (36%) without the antibody had a pleural effusion (9).
Many patients have visceral pleural involvement but do not have pleural effusions. In one series of 1,074 patients undergoing pulmonary resection with intent to cure, visceral pleural involvement was present in 26.8% (10). The 5-year survival rate was significantly less in the patients with visceral pleural involvement (49.8%) than in those without it (76.0%) (10). Almost all patients undergoing lung cancer resection, who have a pleural lavage positive for malignant cells, have visceral pleural involvement (11).
At times, pleural effusions develop in patients who have undergone resection for adenocarcinoma. The incidence of pleural effusion is higher if there is either lymph node or pleural involvement by tumor, at the time of surgery (12). In one series, 18 of 19 patients who developed a cytology-positive pleural effusion after resection had either lymph node metastases, pleural involvement, or both (12). The median time from resection to diagnosis of malignant pleural effusion was 8 months. Most effusions that develop more than 24 months after surgery are due to another primary tumor (12).
It is important to emphasize that the presence of a pleural effusion in a patient with lung cancer almost always indicates that the patient is not curable with surgery whether or not the cytology is positive. Sugiura et al. (13) reviewed 197 patients with stage IIIB or IV non-small-cell lung cancer (NSCLC). They reported that the survival for stage IIIB without effusion, stage IIIB with effusion, and stage IV were 15.3, 7.5, and 5.5 months, respectively (13). The survival
was similar whether the pleural fluid cytology was positive or negative (13). It has been suggested that patients with lung cancer and pleural effusion be classified as M1a which would make them a stage IV (14). Another study also demonstrated that with multivariate analysis, the presence of a pleural effusion at the time of diagnosis adversely affected prognosis (6).
was similar whether the pleural fluid cytology was positive or negative (13). It has been suggested that patients with lung cancer and pleural effusion be classified as M1a which would make them a stage IV (14). Another study also demonstrated that with multivariate analysis, the presence of a pleural effusion at the time of diagnosis adversely affected prognosis (6).
TABLE 10.1 ▪ Causes of Malignant Pleural Effusions in Two Different Series | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
How should the patient with bronchogenic carcinoma and an ipsilateral cytology-negative pleural effusion be evaluated? Rodriguez-Panadero (15) performed thoracoscopy on 21 patients with lung cancer and an ipsilateral cytology-negative pleural effusion. At thoracoscopy, only five patients were believed to be potentially resectable, but when these five were subjected to thoracotomy, their tumors were found to be unresectable because of mediastinal invasion (15). In another older study from the Mayo Clinic, 5 of 73 patients with bronchogenic carcinoma and ipsilateral cytology-negative pleural effusions had long-term survival after the lung cancer was resected (16). In view of the two studies mentioned in the preceding text, it is recommended that patients with bronchogenic carcinoma and an ipsilateral cytology-negative pleural effusion undergo thoracoscopy. If the thoracoscopy is negative, a computed tomography (CT) scan of the chest should be obtained to evaluate the mediastinal lymph nodes. If the CT scan demonstrates lymph node enlargement, a mediastinoscopy should be performed. If the CT scan demonstrates no lymph node enlargement and the thoracoscopy is negative, consideration should be given to an exploratory thoracotomy if the patient has no other contraindication to curative resection.
On occasion, a significant pleural effusion will be recognized only at the time of thoracotomy. Ruffini et al. (17) reported that 52 of 1,279 patients (4%) operated upon between 1993 and 1999 had a pleural effusion with a volume more than 100 mL. The median survival for 16 patients who were inoperable was 6 months, whereas the median survival of 8 patients who had positive cytology but who underwent resection was 9 months. However, the 3- and 5-year survivals of the 21 patients with negative cytology who underwent resection were 68% and 54%, respectively (17).
When a patient with lung cancer is found to have a pleural effusion and is not symptomatic from the effusion, should the effusion be treated? The answer to this question is probably no. Tremblay et al. (18) studied 14 such patients and reported that no patient required any specific pleural intervention for the pleural effusion.
Breast Carcinoma
The second leading cause of malignant pleural effusion is metastatic breast carcinoma. Fracchia et al. (21)
reviewed 601 patients with disseminated breast carcinoma and found that 48% had pleural effusions. The effusions were large enough to warrant therapeutic intervention in 48% of the patients. Goldsmith et al. (22) reviewed the autopsies of 365 patients who had died of disseminated breast carcinoma and reported that 46% had pleural effusions. Pleural effusions were more common with lymphangitic spread (63%) than without lymphangitic spread (41%) (22). In this series, the pleural effusions were on the same side as the primary breast carcinoma in 58% of these patients, on the opposite side in 26%, and on both sides in 16% (22). In a second series, the effusion was ipsilateral in 70%, contralateral in 20%, and bilateral in 10% (23). Ipsilateral effusions are less common if radiotherapy was part of the initial treatment (24). With breast carcinoma, the mean interval between the development of the primary tumor and the appearance of the pleural effusion is approximately 2 and 5 years (24,25), but this interval can be as long as 20 years (26). In patients with pleural effusions secondary to breast carcinoma, determination of the steroid receptors in the effusion is useful in planning therapy.
reviewed 601 patients with disseminated breast carcinoma and found that 48% had pleural effusions. The effusions were large enough to warrant therapeutic intervention in 48% of the patients. Goldsmith et al. (22) reviewed the autopsies of 365 patients who had died of disseminated breast carcinoma and reported that 46% had pleural effusions. Pleural effusions were more common with lymphangitic spread (63%) than without lymphangitic spread (41%) (22). In this series, the pleural effusions were on the same side as the primary breast carcinoma in 58% of these patients, on the opposite side in 26%, and on both sides in 16% (22). In a second series, the effusion was ipsilateral in 70%, contralateral in 20%, and bilateral in 10% (23). Ipsilateral effusions are less common if radiotherapy was part of the initial treatment (24). With breast carcinoma, the mean interval between the development of the primary tumor and the appearance of the pleural effusion is approximately 2 and 5 years (24,25), but this interval can be as long as 20 years (26). In patients with pleural effusions secondary to breast carcinoma, determination of the steroid receptors in the effusion is useful in planning therapy.
One paper suggested that when a patient with breast carcinoma is found to have a pleural effusion, it is preferable to treat the patient with systemic therapy plus pleurodesis rather than systemic therapy alone (27). In this article, the mean pleural progression free interval was 8.5 months in the 102 patients that underwent pleurodesis compared to 4.1 months in the 78 patients that did not undergo pleurodesis originally (27). There was no difference in overall survival.
Lymphomas
Lymphomas, including Hodgkin’s disease, are the third leading cause of malignant pleural effusions (28). The incidence of pleural effusion with Hodgkin’s disease at presentation has varied from 7% to 21% (29,30). During the course of the disease the incidence of pleural effusion is approximately 16% (31). Patients with Hodgkin’s disease who have pleural effusions almost invariably have intrathoracic lymph node involvement, frequently without microscopic pleural involvement (32). Most patients with Hodgkin’s disease and pleural effusion have the nodular sclerosis type (33). Approximately only 3% of the effusions present with Hodgkin’s disease are chylothoraces.
The reported incidence of pleural effusion at presentation in non-Hodgkin’s lymphoma has varied from 6% to 50% (29,30,34). With this neoplasm, 20% to 70% have evidence of mediastinal disease and 90% have evidence of disease elsewhere. For non-Hodgkin’s lymphoma, large-cell lymphomas more frequently have associated pleural disease than do small-cell lymphomas (33). The presence of a pleural effusion at the time of presentation does not adversely affect complete remission or survival rates with non-Hodgkin’s lymphoma (35). The cytology on the pleural fluid is positive in almost all cases (34,35). During the course of the disease, as many as 40% will have a pleural effusion (36). Approximately 20% of the effusions present with non-Hodgkin’s lymphomas are chylothoraces (34).
Vieta and Craver (31) also reported that 12% of 158 patients with lymphatic leukemia and 4% of 52 patients with myelogenous leukemia had pleural effusions. Parietal pleural involvement, however, was uncommon at autopsy with leukemia. The most common cause of pleural effusion in patients with chronic myeloid leukemia and acute lymphobastic leukemia is as a side-effect from the tyrosine-kinase inhibitor dasatinib (see Chapter 22) (37).
Angioimmunoblastic T-cell lymphoma was formerly called angioimmunoblastic lymphadenopathy (AILD) and is characterized by the acute onset of constitutional symptoms, generalized lymphadenopathy, hepatosplenomegaly, anemia, and polyclonal hypergammaglobulinemia. Pathologically, this disorder is characterized by extensive infiltration of lymph nodes with atypical lymphocytes, proliferation of arborizing small vessels, and the deposition of amorphous acidophilic material (19). Most cases of AILD contain monoclonal T-cell populations as well as clonal cytogenetic abnormalities (51). AILD is accepted as a lymphoma and a distinct clinicopathologic entity in the current World Health Organization (WHO) classification (51). A characteristic feature of angioimmunoblastic AITL, seen in more than 95% of all patients, is the presence of increased numbers of Epstein-Barr virus (EBV)-infected cells compared with both normal lymph nodes and peripheral T-cell lymphomas (51). The EBV-infected cells are mostly B cells and the EBV is not thought to have a primary role in the pathogenesis of AITL (51).
Approximately 40% of patients with AITL have pleural effusions (38). The pleural fluid is an exudate with a preponderance of mononuclear cells (39). Other findings on the chest radiograph include interstitial infiltrates and mediastinal or hilar adenopathy, each in 15% to 20% of patients (40). The diagnosis is made by biopsy examination of an enlarged lymph node.
The incidence of malignant pleural effusion in patients with multiple myeloma is approximately 1% (41). The effusions develop at an average of 12 months after the diagnosis of multiple myeloma. Most patients will have associated pleural or chest wall plasmacytomas or pulmonary parenchymal lesions on CT scan (41).
PATHOPHYSIOLOGIC FEATURES
There are several different mechanisms that can be responsible for the development of a pleural effusion in patients with malignancy (Table 10.2). Although it is frequently written that lymphatic obstruction is the primary pathophysiologic abnormality responsible for the pleural effusion with malignancy, this appears not to be true (42). The basis for the contention that lymphatic obstruction is responsible is the observation that at postmortem studies, the presence of pleural effusions is correlated with metastases to the lymph nodes (43). However, the normal rate of pleural fluid formation is thought to be only 15 mL/day. Therefore, if there was complete blockage of the lymphatics, the rate of pleural fluid accumulation should only be 15 mL/day. Certainly, the rate of pleural fluid accumulation frequently exceeds this in patients with pleural malignancy (44). Moreover, if the fluid accumulation was solely due to lymphatic obstruction, one would expect the fluid to be a transudate, but it is almost always an exudate.
We believe that the most likely explanation for the pleural effusion with metastatic disease to the pleura is increased permeability of the pleura (42). Indeed, in the series of Leckie and Tothill (45), a patient with bronchogenic carcinoma had the second highest amount of protein entering the pleural space of the 40 patients studied. The mechanism by which pleural metastases increase the permeability of the pleura is not definitely known. However, we postulate that it is due to the production of vascular endothelial growth factor (VEGF) by the tumor (42). Indeed, the median level of VEGF in pleural effusions secondary to malignancy is much higher than that in patients with effusions secondary to inflammatory disease (46,47). The pleural fluid VEGF levels are also higher in hemorrhagic malignant effusions than in nonhemorrhagic malignant pleural effusions (46). VEGF is one of the most potent agents known for increasing vascular permeability (48). Yano et al. have developed an animal model of a malignant pleural effusion by injecting human adenocarcinoma cells into the pleural space of nude mice (49). The formation of pleural fluid in this model is markedly reduced if the animals are given an inhibitor of the VEGF receptor (49) or if the cells are transfected with an antisense VEGF-165 gene (50) (see discussion on malignant pleural effusions in Chapter 4).
TABLE 10.2 ▪ Mechanisms by Which Malignant Disease Leads to Pleural Effusions | ||||||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| ||||||||||||||||||||||||
It is likely that lymphatic blockade and the resulting decreased clearance of fluid from the pleural space contributes to the accumulation of pleural fluid although, for the reasons outlined earlier, this is not the predominant mechanism in most cases. Leckie and Tothill (45) reported that the mean amount of protein leaving the pleural space in patients with malignant pleural effusions was less than that leaving the pleural space in patients with tuberculosis, pulmonary embolism, or congestive heart failure. This decreased lymphatic drainage can occur through two separate mechanisms. First, because the fluid leaves the pleural space through stomas in the lymphatic vessels in the parietal pleura (51), metastases to the parietal pleura that obstruct these stomas can decrease fluid clearance. Second, the lymphatic vessels of the parietal pleura drain mainly through the mediastinal lymph nodes. Therefore, neoplastic involvement of the mediastinal lymph nodes can decrease the lymphatic clearance of the pleural space.
Malignant tumors can also produce pleural effusions by obstructing the thoracic duct, in which case the resulting pleural effusion is a chylothorax. In fact, most chylothoraces that are not traumatic in origin are secondary to neoplastic involvement of the thoracic duct. Lymphomas are responsible for 75% of chylothoraces secondary to malignant disease (see Chapter 26).
Another mechanism by which malignant tumors produce pleural effusion is through bronchial obstruction. When a neoplasm obstructs the mainstem bronchus or a lobar bronchus, the lung distal to the obstruction becomes atelectatic. Therefore, the remaining lung must overexpand or the ipsilateral hemithorax must contract to compensate for the loss of volume of the atelectatic lung. These events result in a more negative pleural pressure, and it is easy to see from Figure 2.1 that such a negative pleural pressure causes pleural fluid to accumulate. My associates and I studied a patient with obstruction of the bronchus intermedius in whom the pleural pressure dropped from -12 to -48 cm H2O as 200 mL pleural fluid was removed (52).
Pericardial involvement is frequent with metastatic malignant diseases. When a pericardial effusion is caused by such involvement and hydrostatic pressures become elevated in the systemic and pulmonary circulation, transudative pleural effusions may result. It is also likely that some of the malignant pericardial fluid is cleared through the pleural space, which can lead to an exudative effusion (see Chapter 19).
Not all pleural effusions in patients with malignant disease are related to intrathoracic involvement by the neoplasm. Pulmonary infection distal to a partially or totally occluded bronchus may produce a parapneumonic effusion (see Chapter 12). The incidence of pulmonary embolization is higher in patients with malignant disease, and emboli frequently cause exudative pleural effusions (see Chapter 17). Patients with intrathoracic neoplasms frequently receive radiotherapy for their tumors, and this treatment can also result in pleural effusions (see Chapter 23), as can some types of chemotherapy (see Chapter 22). Many patients with malignant disease are malnourished and have hypoproteinemia, and this disorder can, on rare occasions, lead to the formation of transudative pleural effusions (see Chapter 9).
AUTOPSY STUDIES
The most detailed autopsy series on pleural involvement in malignant disease are those of Meyer (43) and Rodriguez-Panadero et al. (53). It appears that pleural metastases with bronchogenic carcinoma are usually due to pulmonary arterial emboli to the ipsilateral pleura. Virtually, all patients with metastatic pleural involvement from lung carcinoma have involvement of the visceral pleura (43,53). In the series of Rodriguez-Panadero et al., pulmonary vascular invasion by the tumor was found in 19 of the 24 cases (53). Parietal pleural metastases result from direct extension from the visceral pleura (43,53).
In patients with nonbronchogenic carcinoma, the visceral pleura is also almost always involved. Involvement of the parietal pleura again appears to result from direct extension from the visceral pleura (43,53). The origin of these metastases is controversial. Meyer attributed them to tertiary spread from secondary hepatic tumors (43). In his series of 23 patients with pleural metastases, 19 (83%) had hepatic metastases. In the series by Rodriguez-Panadero et al., however, hepatic metastases could be demonstrated only in 71% of the patients and they attributed the visceral pleural metastases to blood-borne metastases from the primary (53). The latter explanation appears more plausible to me. The presence of pleural metastases does indicate systemic dissemination of the disease and renders the patient incurable with surgery alone.
Not all patients with pleural metastases have pleural effusions. In Meyer’s series, only 60% of patients with pleural metastases had pleural effusion (43). In Rodriguez-Panadero’s study, only 30 of 55 patients (55%) with metastatic disease to the pleura had pleural effusion (53). Meyer found that the presence of a pleural effusion was more closely related to neoplastic invasion of the mediastinal lymph nodes than to the extent of pleural involvement by nodular metastases (43).
CLINICAL MANIFESTATIONS
The most common symptom reported by patients with malignant pleural effusions is dyspnea, which occurs in more than 50% (7). Symptoms attributable to the tumor itself are also frequent. In one series, weight loss occurred in 32%, malaise in 21%, and anorexia in 14% of patients (7). When patients with malignant pleural effusions are compared to those with benign pleural effusions, patients with malignant pleural effusions are more likely to have dull chest pain (34% vs. 11%), whereas patients with benign disease are more likely to have pleuritic chest pain (51% vs. 24%) (54). Temperature elevations are significantly more common in patients with benign disease (73%) than in patients with malignant disease (37%) (54).
Chest Radiographs
The size of a malignant pleural effusion varies from a few milliliters to several liters, with the fluid occupying the entire hemithorax and shifting the mediastinum
to the contralateral side. Malignant disease is the most common cause of a massive pleural effusion and accounted for 31 of 46 (67%) of effusions occupying an entire hemithorax, in one series (55). In a more recent series, malignancy was responsible for 55% of 163 effusions that occupied more than two thirds of the hemithorax (56).
to the contralateral side. Malignant disease is the most common cause of a massive pleural effusion and accounted for 31 of 46 (67%) of effusions occupying an entire hemithorax, in one series (55). In a more recent series, malignancy was responsible for 55% of 163 effusions that occupied more than two thirds of the hemithorax (56).
Almost all patients with pleural effusions secondary to bronchogenic carcinoma have radiographically demonstrable pulmonary abnormalities besides the effusion. At times, a therapeutic thoracentesis must be performed before the pulmonary abnormality is evident. Although almost all patients with pleural effusions secondary to lymphoma have mediastinal lymph node involvement at autopsy, this involvement is not always evident in chest radiographs (31). In a series of 22 patients with chylothorax due to lymphoma, only 5 patients (23%) had hilar or mediastinal adenopathy demonstrable on routine chest radiographs (57). In another series, however, 71% of 21 patients with pleural effusions secondary to lymphoma had visible mediastinal lymph node involvement on chest radiographs (58). In a third series of 19 patients with non-Hodgkin’s lymphoma, only 4 patients had mediastinal lymphadenopathy (59). The chest radiographs of patients with pleural effusions due to malignant tumors other than lung carcinoma or lymphoma often reveal only a pleural effusion. In a series of 105 patients with pleural effusion due to breast carcinoma (26), only 9% had radiographically evident pulmonary metastases.
In patients with undiagnosed pleural effusions, the chest CT scan is useful in indicating whether the effusion has a benign or malignant etiology. Yilmaz et al. (60) reviewed the CT scans in 146 patients with pleural effusions including 59 that were malignant and 87 that were benign. They reported that the following four findings were suggestive of malignancy: (a) pleural nodularity, (b) pleural rind, (c) mediastinal pleural involvement, and (d) pleural thickening greater than 1 cm (60). In a second study, pleural surfaces assessed by CT scan were abnormal in 27 of 32 patients with malignancy but in none of the 8 patients with benign disease (61). It should be noted that both the series mentioned in the preceding text had a large percentage of mesotheliomas, which are more likely to have abnormalities of the pleural surfaces. Concurrent abnormalities are frequently present in patients with documented malignant effusions. In one study of 86 patients, the following incidences of concurrent abnormalities were reported: pericardial effusion, 3%; pericardial thickening, 14%; mediastinal adenopathy, 43%; chest wall involvement, 12%; lymphangitic carcinoma, 7%; and suspicious lung masses, nodules, or infiltrates, 53%.
Pleural Fluid
The pleural fluid from a malignant pleural effusion is almost always an exudate (62,63). Ashchi et al. (64) reviewed the medical records of 171 patients with malignant pleural effusion and found that in 8 of the cases, the fluid was transudative. There were alternative explanations for the transudative effusions in seven of the eight cases (64). In another study, 97 of 98 patients with malignant pleural effusions had exudates (63). The ratio of the pleural fluid to the serum protein level is less than 0.5 in approximately 20% of malignant pleural effusions (7,62). However, in these 20%, the ratio of the pleural fluid to the serum lactate dehydrogenase (LDH) or the absolute pleural fluid LDH almost always meet exudative criteria (62). Most pleural effusions that meet exudative criteria by the LDH level but not by the protein level are malignant pleural effusions (62).
The presence of grossly bloody pleural fluid (red blood cell [RBC] count >100,000/mm3) suggests malignant pleural disease. In our series of 22 such effusions, 12 (55%) were due to malignant disease (1). Approximately 30% to 50% of malignant pleural effusions, however, have red blood cell counts less than 10,000/mm3 and do not appear bloody (1,65). The pleural fluid white blood cell (WBC) count with malignant pleural effusion is variable, with the usual count between 1,000 and 10,000/mm3 (1). The predominant cells in the pleural fluid differential white cell count of these effusions are lymphocytes in approximately 45%, other mononuclear cells in approximately 40%, and polymorphonuclear leukocytes in approximately 15% (1). In the past, it was stated that pleural fluid eosinophilia (>10%) made pleural malignancy unlikely. However, in a review of 392 cases of eosinophilic pleural effusions that were not associated with pleural air and/or blood, malignancy was the final diagnosis in 17% (66). A more recent paper from a single institution reported that malignancy was responsible for 34.8% of 135 eosinophilic pleural effusions (67).
The pleural fluid glucose level is reduced to below 60 mg/dL in approximately 15% to 20% of malignant pleural effusions (68,69,70). A low pleural fluid glucose level in association with a malignant pleural effusion indicates that the patient has a high tumor burden in the pleural space. Rodriguez-Panadero
and Lopez-Mejias (70) performed thoracoscopy on 77 patients with a malignant pleural effusion and found that the extent of the tumor was significantly greater in those with a low pleural fluid glucose level. Cytology and pleural biopsy are more likely to be positive in patients with low-glucose pleural effusions (70). Because of the large tumor burden, patients with a low pleural fluid glucose level have a worse prognosis (71). It appears that the low-glucose levels with malignant pleural effusion are due to impaired glucose transfer from blood to pleural fluid (72). Increased glucose utilization by the pleural tumor also probably plays a role in producing the low pleural fluid glucose.
and Lopez-Mejias (70) performed thoracoscopy on 77 patients with a malignant pleural effusion and found that the extent of the tumor was significantly greater in those with a low pleural fluid glucose level. Cytology and pleural biopsy are more likely to be positive in patients with low-glucose pleural effusions (70). Because of the large tumor burden, patients with a low pleural fluid glucose level have a worse prognosis (71). It appears that the low-glucose levels with malignant pleural effusion are due to impaired glucose transfer from blood to pleural fluid (72). Increased glucose utilization by the pleural tumor also probably plays a role in producing the low pleural fluid glucose.
Approximately one third of patients with malignant pleural effusions have a pleural fluid pH below 7.3 (70,73,74). Patients with a low pleural fluid pH level also tend to have a low pleural glucose level (70,74). As one might anticipate, they have a greater tumor burden, are more likely to have positive pleural fluid cytology and pleural biopsy, and have a shorter survival than individuals with malignant pleural effusions and a pH level above 7.3 (70,74). The pathogenesis of the low pH level with malignant pleural effusions appears to be due to the combination of acid production by the pleural fluid or the pleura and a block to the movement of carbon dioxide out of the pleural space (72).
Approximately 10% of patients with malignant pleural effusions have an elevated pleural fluid amylase level (69). Usually, the primary tumor is not in the pancreas in these patients (69,75). Analysis of the amylase isoenzymes has demonstrated that the amylase in malignant effusions is the salivary isoenzyme rather than the pancreatic isoenzyme (76), and therefore amylase isoenzyme analysis can be used to differentiate pancreatic effusions from malignant effusions.
DIAGNOSIS
The diagnosis of a malignant pleural effusion is established by demonstrating malignant cells in the pleural fluid or in the pleura itself. In most cases, this is done by cytologic examination of the pleural fluid or biopsy of the pleura.
Cytologic Examination
The easiest way to establish the diagnosis of pleural malignancy is with pleural fluid cytology. The characteristics of malignant cells in pleural fluid are described in Chapter 7. The percentage of cases in which cytologic study of the pleural fluid establishes the diagnosis of a malignant pleural effusion ranges from 40% to 87% (77,78,79,80,81). The reasons for this variability in the diagnostic yield with cytologic study are discussed in Chapter 7. When three separate pleural fluid specimens from a patient with malignant pleural disease are submitted to an experienced cytologist, one should expect a positive diagnosis in approximately 80% of patients. The incidence of positive results depends on the primary tumor. Most cases of metastatic adenocarcinoma can be diagnosed by pleural fluid cytology. Positive results are uncommon with squamous cell carcinoma (SCC) because the pleural effusions are usually due to bronchial obstruction or lymphatic blockade (1,3,82). With lymphoma, the cytologic test is positive in approximately 25% of patients with Hodgkin’s disease and in 50% to 60% of patients with non-Hodgkin’s lymphoma. By cytologic examination, the neoplasm can usually be classified into a histologic type such as adenocarcinoma, but the primary site of the tumor cannot usually be identified (82).
Immunohistochemical Tests
The use of monoclonal antibodies directed against various antigens is useful in distinguishing malignant from benign pleural effusions (83). Metastatic adenocarcinomas tend to stain positive with carcinoembryonic antigen (CEA), MOC-31, B72.3, Ber-EP4, BG-8, and TTF-1 (84), whereas malignant mesothelial cells and benign mesothelial cells stain positive with calretinin, keratin5/6, podoplanin, and WT1 (83) (see Chapter 7). TTF-1 has high specificity for lung carcinoma (83). When attempts are made to differentiate metastatic carcinoma from mesothelioma, stains should be made with two of the antibodies that stain positive for adenocarcinomas and two that stain positive for mesothelioma (84). In the situation where the cytology of the pleural fluid is positive, but the primary is unknown, immunohistochemical tests can help identify the site of the primary (see Chapter 7). The technology with these monoclonal antibodies continues to improve, and it is recommended that all laboratories that deal with significant numbers of pleural fluids develop the capability to perform these immunohistochemical tests.
Tumor Markers in Pleural Fluid
In the last decade, there have been many articles purporting to show the usefulness of tumor markers
in establishing the diagnosis of malignant pleural effusions. Tumor markers evaluated have included CEA; carbohydrate antigens (CA) 15-3, 19-9, 549 and 72-4 (85); neuron-specific enolase (85); SCC antigen; cytokeratin 19 fragments (CYFRA21-1) (85,86); and sialyl state-specific mouse embryonic antigen (SSEA-1) (87). Although the mean levels of tumor markers in the pleural fluid are significantly higher in malignant effusions than in benign effusions, there is almost always some overlap (88). For tumor markers to be useful in the diagnosis of malignant effusions, they have to be 100% specific. One does not want to misdiagnose a benign pleural effusion as a malignant effusion and inform the patient that they only have several months to live. In view of the reasons mentioned in the preceding text, the use of tumor markers for the diagnosis of pleural malignancy is not recommended (89). One possible use of pleural fluid tumor markers is to use high levels to select patients for more invasive studies (88,89). However, other criteria are probably better at predicting that a patient has malignancy. Ferrer et al. (65) found that the following four characteristics were predictive of malignancy in 93 patients undergoing thoracoscopy: a symptomatic period of more than 1 month, the absence of fever, the presence of serosanguineous pleural fluid, and a chest CT scan suggestive of pleural malignancy. All 30 patients who had all four criteria had malignancy, whereas all 20 patients with one or no criteria had a nonmalignant disease (65).
in establishing the diagnosis of malignant pleural effusions. Tumor markers evaluated have included CEA; carbohydrate antigens (CA) 15-3, 19-9, 549 and 72-4 (85); neuron-specific enolase (85); SCC antigen; cytokeratin 19 fragments (CYFRA21-1) (85,86); and sialyl state-specific mouse embryonic antigen (SSEA-1) (87). Although the mean levels of tumor markers in the pleural fluid are significantly higher in malignant effusions than in benign effusions, there is almost always some overlap (88). For tumor markers to be useful in the diagnosis of malignant effusions, they have to be 100% specific. One does not want to misdiagnose a benign pleural effusion as a malignant effusion and inform the patient that they only have several months to live. In view of the reasons mentioned in the preceding text, the use of tumor markers for the diagnosis of pleural malignancy is not recommended (89). One possible use of pleural fluid tumor markers is to use high levels to select patients for more invasive studies (88,89). However, other criteria are probably better at predicting that a patient has malignancy. Ferrer et al. (65) found that the following four characteristics were predictive of malignancy in 93 patients undergoing thoracoscopy: a symptomatic period of more than 1 month, the absence of fever, the presence of serosanguineous pleural fluid, and a chest CT scan suggestive of pleural malignancy. All 30 patients who had all four criteria had malignancy, whereas all 20 patients with one or no criteria had a nonmalignant disease (65).
Other Noninvasive Tests
Other noninvasive tests for the diagnosis of malignant pleural effusions are discussed in Chapter 7. These include measurement of oncogenes, hyaluronic acid, lectin binding, chromosomal analysis, fluorescence in situ hybridization (FISH), flow cytometry and proteomics. In general, the only test that is recommended is flow cytometry when a pleural lymphoma is suspected. Demonstration of the clonality of the lymphocytes in the pleural fluid establishes the diagnosis of lymphoma (90).
Pleural Biopsy
Needle biopsy of the pleura can establish the diagnosis of a malignant pleural effusion. The percentage of positive pleural biopsies in patients with malignant pleural disease ranges from 39% to 75% (80,91,92). In general, pleural fluid cytology is superior to pleural biopsy in establishing the diagnosis of pleural malignancy. Pleural biopsy has a lower diagnostic yield than pleural fluid cytologic examination because, in approximately 50% of patients with malignant pleural disease, the costal parietal pleura is not involved (93). In one large series of 281 patients with malignant pleural disease, the cytology was positive for malignant cells in 162 patients (58%), whereas the pleural biopsy was positive in 123 (44%) (80). The diagnosis of malignancy was established by pleural biopsy alone in only 20 of the 281 patients (7.1%) (80). Another way to look at this series is to consider the 118 patients with malignancy but negative cytology. Pleural biopsy established the diagnosis of malignancy in only 20 of these 118 patients (17%).
What is the role of blind needle biopsy of the pleura in establishing the diagnosis of malignant pleural effusion in the twenty-first century? Because thoracoscopy is very effective at establishing this diagnosis and because the needle biopsy is diagnostic in less than 20% of patients with malignancy and negative cytology, I rarely perform needle biopsy of the pleura. Certainly, if the cytology is negative and thoracoscopy is unavailable or an outpatient procedure is desired, consideration can be given to performing needle biopsy of the pleura.
An alternative means by which specimens of the pleural biopsy can be obtained is with CT-guided cutting-needle biopsy. The CT-guided biopsy is made where the pleura is the most thickened. In one study, 50 patients with unilateral pleural effusion and a clinical suspicion of malignancy were randomized to pleural biopsy through an Abram’s needle biopsy or a CT-guided cutting-needle biopsy (94). The Abram’s needle biopsy was not necessarily aimed where the pleural thickening was greatest. The diagnosis of malignancy was made in 13 of 15 patients (87%) with the CT-guided biopsy but in only 8 of 17 patients (47%) with the Abram’s needle (94). Although these results look very promising, it should be noted that 20 of the 27 patients with malignancy in this study had mesothelioma, which is notoriously difficult to diagnose with Abram’s needle biopsy of the pleura.
Observation, Thoracoscopy or an Open Thoracotomy
In many patients with exudative pleural effusions, no diagnosis is apparent after the initial diagnostic thoracentesis including pleural fluid cytology and a pleural fluid marker for tuberculosis. In such instances, a CT scan of the chest is recommended. If pulmonary
emboli are demonstrated with the CT scan, the etiology of the effusion is established. If parenchymal abnormalities are present, a bronchoscopy should be performed. If the CT scan suggests mesothelioma, then a CT-guided needle biopsy or thoracoscopy should be performed to establish this diagnosis.
emboli are demonstrated with the CT scan, the etiology of the effusion is established. If parenchymal abnormalities are present, a bronchoscopy should be performed. If the CT scan suggests mesothelioma, then a CT-guided needle biopsy or thoracoscopy should be performed to establish this diagnosis.
If the CT scan shows nothing other than the pleural effusion, the approach to the patient should be governed by the clinical picture. If there is nothing in the patient’s history to suggest carcinoma and if the patient’s symptoms are improving, then it is probably best to observe the patient for several weeks because only a small percentage of these patients have malignant pleural disease (95). Alternatively, if the symptoms of the patient are worsening or if there is something in the clinical picture that suggests malignancy, the patient should undergo thoracoscopy. Ferrer et al. (65) reported that the following four characteristics were predictive of malignancy in 93 patients undergoing thoracentesis including 50 with malignancy: (a) duration of symptoms more than 1 month, (b) absence of fever, (c) blood-tinged pleural fluid, and (d) chest radiograph suggestive of malignancy (pulmonary or pleural masses, pulmonary atelectasis, or lymphadenopathy). Twenty-eight patients had all four criteria and all had malignancy. Twenty-one patients had at most one criterion and none had malignancy (65). Thoracoscopy will establish the diagnosis of malignancy in approximately 90% of patients with malignancy (96,97). If facilities for thoracoscopy are not available, an alternative approach is to perform a thoracotomy with open biopsy of the pleura or to perform a needle biopsy of the pleura. If thoracoscopy or thoracotomy is performed, a procedure such as a pleural abrasion should be performed to prevent recurrence of the pleural effusion.
Mesothelioma
The possibility of a malignant mesothelioma (see Chapter 11) should be considered whenever a patient’s pleural fluid cytologic study or pleural biopsy suggests metastatic adenocarcinoma, because the epithelial form of malignant mesothelioma is frequently misinterpreted as adenocarcinoma on cytologic examination or pleural biopsy (98). If no primary tumor is evident, a CT scan of the thorax should be obtained. If the CT scan suggests mesothelioma, one should consider thoracoscopy or exploratory thoracotomy for staging and possible radical pleuropneumonectomy (see Chapter 11).
As discussed early in this chapter and in Chapter 7, immunohistochemical tests are useful in making this differentiation. In addition, histochemical tests using periodic acid-Schiff stain (PAS) or Alcian blue and electron microscopy are very useful in differentiating metastatic adenocarcinoma from mesothelioma (see Chapter 11).
Lipid Analysis
The possibility of a chylothorax should be considered in every patient with malignant disease and a pleural effusion. If a chylothorax is present, the mediastinal lymph nodes are probably involved, and the treatment of choice is radiation to the mediastinum or chemotherapy. The supernatant of the pleural fluid from patients with malignant pleural effusions should be examined. If the supernatant is turbid, a chylothorax should be suspected, and the triglyceride level in the pleural fluid should be determined. If the pleural fluid triglyceride level exceeds 110 mg/dL, the patient probably has a chylothorax; if the level is below 50 mg/dL, the patient does not have a chylothorax (99). If the level is between 50 and 110 mg/dL, lipoprotein electrophoresis should be performed (99) (see Chapter 26).
Other Diagnostic Tests
Numerous articles have recommended various diagnostic tests such as flow cytometry, chromosomal analysis of pleural fluid cells, or LDH isoenzymes in the diagnosis of malignant pleural effusions. These various tests are discussed in Chapter 7. In general, they are not recommended. Flow cytometry with immunophenotyping is useful in making the diagnosis of lymphoma (59).
Unknown Primary
Most patients who are diagnosed with a malignant pleural effusion are already known to have a malignancy. However, if the patient presents with a malignant pleural effusion, where is its likely origin? One study (100) reviewed 42 consecutive patients referred to the Royal Marsden Hospital with a malignant pleural effusion and no known primary. The patients included 27 men and 15 women. Despite CT scans of chest and abdomen and mammography and pelvic ultrasound in 10 patients, a primary was determined in only 15 (10 men and 5 women), and the primary was in the lung in all cases (100). The median survival in this group was 12 months from diagnosis (100).
In view of the findings mentioned in the preceding text, it is recommended that patients with a malignant pleural effusion and an unknown primary tumor have a CT scan of the chest, abdomen, and pelvis. If pulmonary parenchymal abnormalities are discovered, then a bronchoscopy is indicated with special attention to the area of abnormality. If there are no parenchymal abnormalities, then bronchoscopy will probably be nondiagnostic (100). Masses in the abdomen should be evaluated. If the patient has symptoms referable to a specific organ, that organ should be evaluated. If the patient is a woman, mammography and a careful pelvic examination should be performed. If the foregoing sequence of tests does not identify the site of the primary tumor, it is recommended that further tests not be undertaken (100).
PROGNOSIS
The prognosis of patients with malignant pleural effusions is not good. In a one report, the median survival of 417 patients with malignant pleural effusions was only 4 months (71). This 4-month median survival is on the optimistic side because all 417 patients were judged to be fit enough to undergo pleurodesis (71). The most important factor influencing the life expectancy in patients with malignant pleural effusion is the source of the tumor. In the study mentioned in the preceding text, the median survivals were 3 months for 146 patients with lung cancer, 2.3 months for 18 patients with gastrointestinal primaries, 5 months for 60 patients with breast carcinoma and 51 patients with unknown primary, and 6 months for 29 patients with mesothelioma (71). In a more recent report (101) of 284 patients with malignant pleural effusion, the overall median survival was 5.4 months following diagnosis. Again survival varied significantly depending on the primary tuber being 17.4 months for mesothelioma, 13.2 months for breast cancer, 7 months for lymphoma and 2.6 months for lung cancer (101). A second factor that is very important in determining the prognosis of patients is their Karnofsky Performance Scale (KPS) score. Burrows et al. (102) reported that the median survival of patients with a KPS score less than 30 was 34 days, whereas the median survival of patients with a KPS score greater than 70 was 395 days.
Other factors associated with a poor prognosis are a pleural fluid pH level below 7.20, a pleural fluid glucose level below 60 mg/dL, or a pleural fluid LDH level more than twice the upper limit of normal for serum (71,101,102). In addition, the greater the number of pleural adhesions at thoracoscopy (103) and the higher the VEGF level in pleural effusions due to lung cancer (104), the poorer the prognosis. All of the poor prognostic factors mentioned in the preceding text probably reflect a greater tumor burden in the pleural space (71,102).
TREATMENT
The presence of a malignant pleural effusion indicates that the malignancy cannot be cured by surgery. Management of the effusion centers on palliation of symptoms because no available treatments prolong survival (Fig. 10.1).
The initial step is to identify the location of the primary lesion. Frequently, the location of the primary is already known when the pleural effusion is first identified. If the primary is unknown, then the procedures outlined in the preceding paragraph on unknown primary should be followed.
The main reason to identify the primary tumor is to decide whether systemic chemotherapy is indicated. Systemic chemotherapy is effective at least in some patients with small-cell carcinoma, breast carcinoma non-small-cell lung cancer and lymphoma. If the patient has a chylothorax, radiation should be administered to the mediastinum.
If the patient has a tumor that is not responsive to chemotherapy or fails to respond to chemotherapy, a procedure should be considered to remove the pleural fluid. The two primary modes of treatment to control the accumulation of pleural fluid are the insertion of an indwelling pleural catheter or the creation of a pleurodesis. Only patients who are dyspneic and whose dyspnea improves after a therapeutic thoracentesis should be considered candidates for fluid removal. If the patient is not symptomatic from the pleural effusion, no treatment is recommended. Most patients who have small effusions that do not produce symptoms never become symptomatic (105). The dyspnea of patients who do not improve after a therapeutic thoracentesis should be treated with opiates and/or oxygen. Before a pleurodesis is attempted, the position of the mediastinum on the chest radiograph should be noted. If the mediastinum is shifted toward the side of the effusion, a bronchoscopy should be done before pleurodesis is attempted because it is likely that the patient has an obstructed bronchus. The presence of an obstructed bronchus is a contraindication to pleurodesis.
If the patient is dyspneic and if the dyspnea is relieved by a therapeutic thoracentesis, the procedure
of choice for outpatients who can receive home health care or who have strong family support for controlling the pleural fluid is the insertion of an indwelling catheter with periodic drainage of the pleural fluid through vacuum bottles. If patients do not meet these criteria, pleurodesis with a sclerosing agent through a chest tube is the procedure of choice. If the lung does not expand after the chest tube is inserted, pleurodesis should not be attempted (106). These patients can be treated with either an indwelling pleural catheter or a pleuroperitoneal shunt.
of choice for outpatients who can receive home health care or who have strong family support for controlling the pleural fluid is the insertion of an indwelling catheter with periodic drainage of the pleural fluid through vacuum bottles. If patients do not meet these criteria, pleurodesis with a sclerosing agent through a chest tube is the procedure of choice. If the lung does not expand after the chest tube is inserted, pleurodesis should not be attempted (106). These patients can be treated with either an indwelling pleural catheter or a pleuroperitoneal shunt.
 FIGURE 10.1 ▪ Algorithm for managing patients with malignant pleural effusions. |
Systemic Chemotherapy
The presence of a malignant pleural effusion usually indicates disseminated tumor, at least for nonbronchogenic carcinoma (43). Therefore, the only hope for cure or prolonged palliation is with systemic
chemotherapy. Fentiman et al. (26) reported that pleural effusions were controlled in 7 of 22 patients (32%) with metastatic breast carcinoma who were given systemic chemotherapy, whereas Jones et al. (107) reported positive responses in 6 of 8 patients (75%) given systemic chemotherapy. There are no recent reports to my knowledge concerning the outcomes of malignant pleural effusions with breast cancer treated with systemic chemotherapy. Livingston et al. (108) reported that 19 of 53 patients (36%) with small-cell lung carcinoma had complete disappearance of their pleural effusions with chemotherapy. The median survival of patients with small-cell lung carcinoma with limited disease and a pleural effusion is 13.9 months, compared with a median survival of 18.3 months if no pleural effusion is present (109). In a second study of 62 patients with small-cell lung carcinoma, the pleural effusion disappeared after first-line chemotherapy in 34 patients (55%) (8). In addition, pleural effusions in lymphomas frequently respond to chemotherapy (35).
chemotherapy. Fentiman et al. (26) reported that pleural effusions were controlled in 7 of 22 patients (32%) with metastatic breast carcinoma who were given systemic chemotherapy, whereas Jones et al. (107) reported positive responses in 6 of 8 patients (75%) given systemic chemotherapy. There are no recent reports to my knowledge concerning the outcomes of malignant pleural effusions with breast cancer treated with systemic chemotherapy. Livingston et al. (108) reported that 19 of 53 patients (36%) with small-cell lung carcinoma had complete disappearance of their pleural effusions with chemotherapy. The median survival of patients with small-cell lung carcinoma with limited disease and a pleural effusion is 13.9 months, compared with a median survival of 18.3 months if no pleural effusion is present (109). In a second study of 62 patients with small-cell lung carcinoma, the pleural effusion disappeared after first-line chemotherapy in 34 patients (55%) (8). In addition, pleural effusions in lymphomas frequently respond to chemotherapy (35).
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
