closed pleural biopsy had pooled sensitivity and specifcities of 77% and 99%, respectively, compared with 93% and 100% for medical thoracoscopy [9]. Medical thoracoscopy allows for directed biopsies of abnormalities identifed by direct inspection of the pleura, at least partially explaining the increased sensitivity of the procedure. Both procedures are quite safe with pooled complication rates of 8% for medical thoracoscopy­ and 5% for closed pleural biopsy. Although the complication rate appears signifcantly higher for medical thoracoscopy, the most common complication documented is subcutaneous emphysema which for the most part is of little clinical signifcance [10].

Diagnostic Indications for Medical Thoracoscopy

Lung Cancer

As stated above, medical thoracoscopy has an important clinical role in the diagnostic algorithm for unclear exudates and in some cases transudative effusions. Staging of cancers is one of the most common indications for medical thoracoscopy with a pooled sensitivity of 92% (range 88–95%). A major advantage of medical thoracoscopy is the ability to visually assess the pleura. Sanchez-Armengol frst described a simple, semi-­quantitative method to assess disease involvement of the pleura at time of thoracoscopy in 1993. Essentially it assigns a score of 0–3 depending on whether there is no, isolated, diffuse or massive tumor involvement of the costo-­ parietal, visceral, and diaphragmatic pleural surfaces [11]. This method has subsequently been modifed to add a grade for adhesions (0–4) and whether or not the effusion is bloody. The adhesions are graded as follows: 0 none, 1 localized and non-obstructive, 2 obstructing 1/3 of thoracic cavity from visualization, 3 obstructing 2/3 pleural surface from view, and 4 fused pleural space unable to perform procedure. Extent of pleural carcinomatosis (EPC) and grade of adhe-

sion are both associated with prognosis on multivariate analysis (p = 0.007 and p = 0.019, respectively). Of note, these fndings predict a median survival difference for non-small cell lung cancer of 3 months for EPC and 6 months for grade of adhesions [12].

Mesothelioma

Mesothelioma can often portend a more complex diagnostic dilemma as disease extent is directly correlated with prognosis which in turn impacts on treatment. The extent of tumor invasion, cellular subtype, and molecular profle may impact prognosis and treatment. Literature is sparse comparing the diagnostic techniques for malignant mesothelioma. Pleural uid cytology alone for the diagnosis and prognostication of malignant mesothelioma is limited with reported low sensitivity. Reactive mesothelial cells may have a similar appearance in cytology specimens. Newer techniques including an increasing array of biomarkers will most assuredly increase the utility of pleural uid analysis for mesothelioma in the future but at present these remain experimental due to small sample sizes and variability in methods of analysis [13]. Computed tomography or ultrasound-guided biopsy do have the advantage in that they may be safely performed in patients with overall poor performance status including those with small or no effusion or those unable to tolerate sedation due to comorbid illness. Recently published guidelines by the American Society for Clinical Oncology suggest that thoracoscopy with pleural biopsy be performed as it may assist with staging, elucidate pathologic cell type, confrm or establish the diagnosis, as well as provide material for additional testing such as molecular profling [14]. Medical thoracoscopy provides an advantage over other surgical options such as video-assisted thoracoscopic surgery or robotic surgery in that it is generally performed using a single, small access port that may be readily resected or radiated to avoid tumor spread along the procedural tract.

Other Tumors

Medical thoracoscopy may be benefcial in other tumor cell types such as hematologic malignancy, and breast cancer. The literature regarding the utility of the procedure in the management of pleural involvement from hematologic malignancies is limited. Effusions in patients with hematologic­ malignancy may be either malignant or benign. Pleural involvement does portend a poorer prognosis for patients with hematologic malignancies [15].

Diagnosis of hematologic malignancies via pleural uid analysis may be limited by lack of cellularity and ambiguous morphology. Flow cytometry and immunohistochemistry may be performed on tissue samples when pleural uid studies are non-diagnostic. In these cases, medical thoracoscopy with directed biopsy of the pleura may assist with diagnosis, prognostication, and treatment.

With respect to breast cancer, many patients with advanced-stage breast cancer will develop a pleural effusion over the course of their illness. The effusion may be secondary to drug toxicity, radiation injury, volume overload, or metastatic disease. The ability to determine the cause of the effusion may markedly alter therapy. Pleural fluid may be used for tumor markers including estrogen and progesterone receptors, HER2/neu, and even check point inhibitors such as PD-L1 (programmed deathligand 1). This does require preparation of a cell block. Biopsies performed via thoracoscopy are large, generally have a high tumor tissue burden when taken from areas of obvious malignancy, and may reliably provide tissue for the detection of hormone receptors [16]. Our unpublished experience is that thoracoscopic biopsies from grossly abnormal areas of the pleura also provides ample tissue for molecular analysis including next-genera- tion sequencing. In the era of personalized medicine, the pleura represents a rich source of tissue that may be obtained at the same time

as performing palliative procedures such as pleurodesis or tunneled pleural catheter placement.

Tuberculosis

Worldwide tuberculosis (TB) represents one of the predominant causes of pleural effusion; however, in developed countries the prevalence is much lower. Although pleural uid is often sent for acid fast bacilli (AFB), it is believed that uid accumulation is primarily the result of an in ammatory response to the pathogen, as evidenced by granulomatous in ammation of the pleura, as opposed to a heavy burden of organisms. As such it is not surprising that the diagnostic yield of pleural uid AFB alone is low. When combined with an adenosine deaminase (ADA) >40 μ/L and LDH>200 μ/L, the sensitivity of pleural uid analysis increases to 74% with a specifcity of 80%. False positives may occur in patients with empyema or lymphoma related-effu- sions which may also have elevated ADA and LDH levels [17]. Medical thoracoscopy provides the opportunity to inspect the pleural space. Kong and colleagues described the classic visual fndings seen at thoracoscopy for pleural TB. These include necrosis, diffuse miliary nodules, pleural nodularity, and thickened pleura with or without loculation [18]. The classic appearance alone carries a diagnostic accuracy of 93%. When biopsies are performed, the sensitivity of thoracoscopy nears 100% with granulomatous in ammation as the defning feature. Closed pleural biopsy in the setting of a lymphocytic effusion with an ADA>40 μ/L and LDH>200 μ/L approaches the diagnostic yield of medical thoracoscopy when both histopathology and culture results are considered; however, for the subset of patients with culture positivity alone it may take longer to obtain a diagnosis considering the high sensitivity and specifcity of the simple appearance of the pleura in pleural TB. Closed pleural biopsy cultures are diagnostic in approximately 60% of cases, but these results are often delayed several weeks [18, 19].