Mediastinal Anatomy and Mediastinoscopy

Mediastinal Anatomy

The anatomic boundaries of the mediastinum include the thoracic inlet superiorly, the diaphragm inferiorly, the sternum anteriorly, the spine posteriorly, and the pleural spaces bilaterally. It is convenient to divide the mediastinum into anatomic compartments that provide a method for classification of disease processes.

A classic description divides the mediastinum into four compartments: superior, anterior, middle, and posterior ( Fig. 40-1 ). The superior mediastinum includes all structures from the thoracic inlet superiorly to an imaginary plane that includes the lower edge of the manubrium and the lower edge of the fourth thoracic vertebra. The inferior mediastinum, which lies below this boundary, is further divided into the anterior, middle, and posterior compartments. The boundary between the anterior and middle compartments is the anterior pericardium, whereas the border between the middle and posterior compartments is the posterior aspect of the tracheal bifurcation, pulmonary vessels, and pericardium. In this four-compartment model, the upper portions of the trachea and esophagus are contained within the superior mediastinum; the lower portions are contained within the middle and posterior mediastinum.

Shields proposed a simpler three-compartment model consisting of an anterior compartment, a middle (or visceral) compartment, and a posterior compartment (paravertebral sulcus; Fig. 40-2 ). All three compartments are bounded inferiorly by the diaphragm, laterally by the pleural space, and superiorly by the thoracic inlet. The anterior compartment is contained anteriorly by the sternum and posteriorly by the great vessels and pericardium and this compartment contains the thymus, internal mammary vessels, areolar and adipose tissue, and potentially pathologic structures such as ectopic parathyroid tissue or a retrosternal goiter. The middle mediastinum is bordered posteriorly by the ventral surface of the thoracic spine; it occupies the entire thoracic inlet and contains the majority of mediastinal structures, namely, the great vessels, heart, pericardium, trachea, proximal main-stem bronchi, vagus nerves, phrenic nerves, esophagus, thoracic duct, descending aorta, and azygos venous system. The posterior compartment (paravertebral sulcus or sulci) consists of potential spaces along the thoracic vertebrae that contain the sympathetic chain, proximal portions of the intercostal neurovascular bundles, thoracic spinal ganglia, and distal azygos vein. Whereas some anatomists may argue that the paravertebral sulci are not truly a mediastinal space, they often harbor disease processes that are classically considered in the posterior mediastinum (i.e. neurogenic tumors). Detailed sagittal views of mediastinal anatomy are depicted in Figures 40-3 and 40-4 .

Potential Spaces in the Mediastinum

For ease and accuracy of communication in describing pathologic abnormalities (most commonly lymph node enlargement), several potential mediastinal spaces are described. The pretracheal space is a triangular space bounded anterolaterally by the superior vena cava and right brachiocephalic vein on the right, the aorta and pericardium on the left, and the trachea posteriorly. Continuing inferiorly from the pretracheal space is the subcarinal space. This location is bounded superiorly by the carina, laterally by the two main-stem bronchi, anteriorly by the back of the right pulmonary artery, and posteriorly by the esophagus (see Fig. 40-3 ). The pretracheal and subcarinal spaces are routinely explored in mediastinoscopy and endobronchial ultrasound. The aortopulmonary window is the space bounded superiorly by the aortic arch, medially by the trachea and esophagus, inferiorly by the pulmonary artery, and laterally by the pleura. This space contains lymph nodes, the ligamentum arteriosum, and the left recurrent laryngeal nerve (see Fig. 40-4 ). Routine cervical mediastinoscopy does not fully access this space, but anterior mediastinotomy (Chamberlain procedure), extended cervical mediastinoscopy, and thoracoscopy or thoracotomy can all provide access to the aortopulmonary window.

Mediastinal Lymph Node Anatomy

The adoption of a common thoracic regional lymph node classification by the American Joint Committee on Cancer and the Union for International Cancer Control in 1997, known as the Mountain-Dresler chart, has found widespread acceptance ( Fig. 40-5 ). This system classifies lymph nodes into 14 stations, of which stations 1 through 9 are contained within the mediastinal pleura and are considered to be mediastinal lymph nodes. The highest mediastinal station (level 1), upper right and left paratracheal nodes (level 2R, 2L), lower right and left paratracheal nodes (level 4R, 4L), and the subcarinal nodes (level 7; see Fig. 40-3 ) are the only mediastinal nodal stations accessible by standard cervical mediastinoscopy, while stations 5 (subaortic nodes) and 6 (paraaortic nodes; see Fig. 40-4 ) require an alternative approach such as extended mediastinoscopy or anterior mediastinotomy (Chamberlain procedure) or, rarely, endoscopic ultrasound (EUS).

Indications for Mediastinal Lymph Node Assessment

The most common indication for surgical assessment of mediastinal lymph nodes is non–small cell lung cancer (NSCLC). Other indications include mediastinal lymphadenopathy of unknown etiology, mediastinal masses, primary tracheal tumors, and occasional esophageal tumors. Rare indications for mediastinoscopy include drainage of bronchogenic cysts, abscess drainage, identification of ectopic parathyroid tissue, and tissue sampling for causes of superior vena cava syndrome.

Efficacy and Utility of Mediastinoscopy

Ideally, levels 2R, 2L, 4R, 4L, and 7 should be sampled (with at least one specimen from each level) during mediastinoscopy, though this ideal is seldom achieved. It is estimated that approximately half of false-negative mediastinoscopy results were due to pathologically positive mediastinal nodes not accessible by the mediastinoscope. A retrospective series comparing conventional mediastinoscopy to video-assisted mediastinoscopy has demonstrated a significantly lower complication rate (3.6 versus 1.6; P = 0.03), higher number of sampled lymph nodes, and lower number of missed positive lymph nodes after definitive lung resection with the use of video mediastinoscopy at the time of staging. Another retrospective series also showed a higher average number of lymph node samples (7 versus 5; P < 0.001) and number of lymph node stations sampled (3.6 versus 2.6; P < 0.01) with video mediastinoscopy, but in contrast to the previous study, there was a higher complication rate with video assistance (3.8% with versus 0.8% without; P = 0.04), which was attributed to a more aggressive dissection. The most frequent nodal station to be implicated in a false-negative result at time of eventual pulmonary resection was level 7 (approximately 70% of surprise N2 cases found at thoracotomy). However, almost 90% of these false-negative cases at level 7 were in the conventional mediastinoscopy group, only one false-negative level 7 case resulted after video mediastinoscopy.

The use of modern imaging techniques, including high-resolution computed tomography (CT) and positron emission tomography (PET), has led to an appropriately selective strategy for invasive mediastinal lymph node assessment, especially for patients with stage I NSCLC. Some series have shown that the sensitivity and negative predictive value of positron emission tomography–computed tomography (PET/CT) approaches levels achieved with cervical mediastinoscopy ( Table 40-1 ). However, reviews of large series have reminded surgeons to interpret PET results with caution. In one such study in which preoperative PET (alone) was followed by mediastinoscopy, the sensitivity of PET was 64.4%, specificity was 77.1%, positive predictive value was 44.6%, negative predictive value was 88.3%, and accuracy was 74.3%. In a similarly sized series that evaluated combined PET-CT in patients who subsequently underwent mediastinoscopy, these values improved with a sensitivity of 70%, specificity of 94%, positive predictive value of 64%, and negative predictive value of 95%. However, both studies demonstrate a clinically important number of false-positive results with PET or PET/CT that, if accepted without confirmation, would result in unnecessary induction therapy or, in some practices, the possibility of the patient not being offered surgical management at all. These results suggest the continued need for pathologic confirmation of mediastinal lymph node staging before definitive management.

The European Society of Cardiothoracic Surgery has provided recommendations for mediastinal lymph node assessment in preoperative staging of NSCLC ( Fig. 40-6 ). Similarly, the National Comprehensive Cancer Network guidelines currently recommend preoperative PET/CT scanning and pathologic mediastinal lymph node evaluation (either mediastinoscopy or endobronchial ultrasound-transbronchial needle aspiration [EBUS-TBNA]), EUS, or CT-guided biopsy for clinical stage IA-IIIB NSCLC disease. However, a cost-effectiveness analysis found that patients with clinical stage I lung cancer staged by PET/CT benefit little from mediastinoscopy. In this series of stage I NSCLC patients who underwent mediastinoscopy, only 3% demonstrated N2 disease, with an additional 3.6% found to have occult N2 disease after their definitive surgery. With this rate of N2 disease, mediastinoscopy added 0.008 years of life expectancy, costing more than $250,000 per life-year gained. Many practitioners do not routinely perform mediastinoscopy for patients with clinical stage IA or IB disease after CT and PET screening. Exceptions may include patients with central tumors, bronchoalveolar cell carcinoma, or in situations in which the primary tumor has low fluorodeoxyglucose uptake and there are enlarged, but PET-negative, mediastinal nodes. Any patient with suspected N1 disease after CT or PET scan should undergo mediastinal evaluation.