Diagnostic approach to pleural diseases: new tricks for an old trade

a. The ultrasound revolution

Thoracic ultrasound, widely available in the form of increasingly sophisticated and affordable portable and handheld units, has unquestionably changed the paradigm in bedside medical evaluation. Basic ultrasound education now complements traditional physical examination training at many medical schools, residency, and fellowship programs. Point-of-care ultrasound for diagnosis and management of thoracic diseases in particular is recommended and endorsed by most relevant scientific societies, and there are established training and certification standards^13–16. In fact, the evidence for improved outcomes and reduced complication rates with ultrasound for thoracic procedures is so overwhelming that its use is now universally considered the standard of care^17,18. Bedside ultrasound examination of the pleura can easily confirm the presence of pleural fluid when conventional chest x-ray findings are equivocal. In addition, it allows an estimation of its volume and complexity and occasionally identifies pleural nodules or masses suggestive of pleural malignancy¹⁹. While most physicians use ultrasound merely to mark the site for needle insertion before setting up for the procedure, real-time ultrasound using a sterile sleeve allows sampling of previously unreachable small loculated effusions and occasionally allows parietal pleural biopsies. Furthermore, recent evidence suggests that pre-procedural identification of the intercostal artery, the course of which can be unpredictable, is easily achieved with a high-frequency probe and could help avoid rare but potentially fatal bleeding complications^20–22. Ultrasound examination immediately after the procedure can indirectly suggest the occurrence of pneumothorax by the disappearance of the typical “sliding” sign, representing the sliding of visceral and parietal pleura over each other. Rare but potentially serious bleeding complications from injury of the intercostal artery or one of its collateral vessels may also be identified in the form of rapid re-accumulation of echogenic pleural fluid^23,24.

Interestingly, there has been relatively little research on pleural ultrasound as a predictive tool. Two small studies suggest that a highly organized pleural space with loculations and septations during pleural space infection predicts poorer outcomes, but these studies are small and unadjusted for possible confounders^25,26. An intriguing recent report suggests that ultrasound may allow the identification of unexpandable lung before thoracentesis. In some cases, such as when chronic inflammatory effusions have resulted in the creation of a thick peel around the lung, fluid drainage is not accompanied by lung re-expansion. The vacuum generated can expose patients to complications including pain, pneumothorax, and re-expansion pulmonary edema. Transmission of the heartbeat results in different lung movement and deformation changes that can be identified by using specific ultrasound modes such as the M-mode and speckle tracking²⁷. In this small study, ultrasound in fact performed better than pleural manometry, long regarded as the gold standard for the diagnosis of unexpandable lung and the utility of which is being assessed in an ongoing randomized clinical trial (ClinicalTrials.gov identifier NCT02677883).

b. Other imaging modalities

There have been comparatively less recent data on other imaging modalities. Whereas loculations (sequestrations of fluid in non-dependent areas) are well identified on computed tomography, the degree of organization within the effusion is better appreciated on ultrasound. In addition, the presence of any of the four following radiologic signs is highly predictive of malignancy: parietal pleural thickening of more than 1 cm, thickened mediastinal pleura, circumferential thickening, and the presence of pleural nodules²⁸. However, in the absence of such specific findings, computed tomography appears neither sensitive nor specific enough to obviate the need for invasive diagnostic interventions^29,30. Positron emission tomography is often used to assess the probability of malignant pleural effusion but with modest sensitivity and specificity, which in a recent meta-analysis were estimated to be 81% and 74%, respectively³¹. Recently, magnetic resonance imaging has been proposed as a potentially superior imaging modality for pleural effusions in general and malignant pleural mesothelioma in particular but remains largely underutilized compared with computed tomography^32–34. Radiomic approaches to thoracic diseases represent an early and exciting field in imaging science, and very preliminary work in quantitative imaging for the diagnosis of malignant pleural effusion—and staging and diagnosis of mesothelioma in particular—is generating considerable interest^35,36.

a. Pleural fluid analysis

The “diagnostic separation of transudates and exudates” using a combination of pleural fluid protein and lactate dehydrogenase levels was published in 1972 (in a study that has been cited more than 1,500 times) and to this day remains the single most important step in determining the etiology of a pleural effusion¹¹. Briefly, at the risk of oversimplifying complex processes, transudative effusions typically result from local alterations of the Starling rules that govern capillary microcirculation in the parietal pleura with increased hydrostatic (for example, heart failure) or decreased oncotic (for example, liver failure) pressures, whereas exudative effusions suggest a disease process involving the parietal pleura itself, whether from malignancy, infection, or inflammatory diseases. However, these processes are complex and overlaps are frequent in terms of both pathophysiology and pleural fluid analysis. For example, an estimated 25% of effusions due to congestive heart failure may classify as exudates (“pseudo-exudates”), particularly if the patient is on diuretics³⁷, and conversely a small percentage of documented malignant pleural effusions may classify as transudative effusions. While a high index of suspicion should lead to pleural biopsies in the latter scenario, recent data suggest that the pleural fluid–to–serum albumin gradient could help clarify the etiology of heart failure–related pseudo-exudative pleural effusions³⁸. Pleural and serum brain natriuretic peptide and amino-terminal pro-brain natriuretic peptide levels are also useful to establish heart failure as the cause of the effusion³⁹. Pleural space infections, potentially life-threatening conditions, are usually characterized by typical pleural fluid biochemistry and hence are more straightforward from a diagnostic standpoint, and an interesting recent observation is that the yield of pleural fluid cultures can be increased by 20% by simply inoculating blood culture bottles with pleural fluid at the bedside rather than sending the fluid directly to the laboratory⁴⁰.

b. Cytology and biomarkers

The diagnostic utility of pleural fluid analysis for malignant pleural effusions has been the object of several recent studies. The sensitivity of cytology for malignancy is estimated around 60%, and there is a 15% increase with a second procedure⁴¹, although some reports suggest much lower estimates⁴². Others focusing on modern immunohistochemistry techniques seem much more optimistic, particularly for mesothelioma, a primary pleural malignancy which traditionally requires pleural biopsies for definitive diagnosis^43,44. In addition, pleural fluid has been shown to be a biospecimen suitable for the majority of molecular analyses required for targeted therapy⁴⁵. High-throughput techniques and modern molecular techniques promise to identify biomarkers expressed by cancer cells or their environment that could ultimately transform our diagnostic approach. Novel proposed techniques include single-cell mechanophenotyping which evaluates the deformability of pleural cells⁴⁶, circulating tumor cells and cell-free tumor DNA^47,48, and metabolic-based assays to identify non-leukocyte metabolically active tumor cells⁴⁹. These recent developments, though exciting, remain preliminary and are still closer to the bench than they are to the clinic. The search for optimal biomarkers has been hampered by a lack of standardized methodology and failure to externally validate promising results, as in the case of fibulin-3, a biomarker once anticipated to transform our approach to malignant mesothelioma^50–53. Large ongoing research projects are attempting to identify reliable alternative candidates.

c. Clinical prediction models

An unprecedented number of well-designed randomized controlled studies published in the last decade by a growing international pleural research network have clarified and sometimes transformed patient management for malignant pleural effusions and pleural space infections in particular. However, one major obstacle faced by researchers and clinicians has been the lack of clinical prediction models allowing appropriate patient selection and stratification. The RAPID (renal, age, purulence, infection source, and dietary factors) score was derived and validated from two large datasets and proposes to risk-stratify patients with pleural space infections, and ultimately its purpose is to individualize management, which currently is subject to local preferences and expertise rather than patient characteristics⁵⁴. A large multicenter observational prospective study to validate this score is ongoing. Similarly, the LENT (pleural fluid lactate dehydrogenase, Eastern Cooperative Oncology Group performance score, neutrophil-to-lymphocyte ratio, and tumor type) score is a clinical prediction model that provides estimates of survival for patients with malignant pleural effusions, which could prove very useful in subject selection in clinical trials and ultimately individualized medical or surgical management of pleural effusions⁵⁵.