Endoscopic ultrasound in the evaluation of posterior mediastinal lesions and lung cancer

Key points

Criteria exist to differentiate benign from malignant mediastinal lymph nodes, but used alone, these criteria are not sufficiently accurate. Endoscopic ultrasonography-guided fine-needle aspiration (EUS FNA) for cytopathologic evaluation is required to make sound clinical decisions.
The overall accuracy for the diagnosis of posterior mediastinal malignancies with transesophageal EUS FNA is greater than 90%.
EUS FNA and endobronchial ultrasound-guided transbronchial needle aspiration (EBUS-TBNA) are the techniques of choice in case mediastinal nodal sampling is indicated for the diagnosis and staging of lung cancer. In centers without availability of EBUS, staging with EUS FNA is an alternative.
Unique features of EUS FNA in lung cancer diagnosis and staging are access to lower mediastinal lymph nodes, evaluation for mediastinal invasion (T4 stage), and detection of distant metastatic disease to the liver and left adrenal gland.
The diagnosis of lymphoma in the posterior mediastinum is made by cytology and flow cytometry studies on EUS FNA and core biopsy specimens.
EUS FNA can be valuable in helping to establish a diagnosis of granulomatous disease involving the mediastinum (sarcoidosis, histoplasmosis, tuberculosis).
While EUS FNB is being increasingly used to sample posterior mediastinal lymph nodes or masses, comparative performance data with EUS FNA is lacking.
Most mediastinal cysts are benign, and because the risk of infection is high, EUS FNA should not be performed. If a high suspicion of malignancy exists, the cyst should undergo one puncture and be fully drained, and antibiotics should be administered.

Introduction

Transesophageal endoscopic ultrasonography (EUS) with fine-needle aspiration (FNA) allows for minimally invasive evaluation of posterior mediastinal lesions and the diagnosis and staging of non–small cell lung cancer (NSCLC). Posterior mediastinal lesions are usually first detected with computed tomography (CT), but occasionally are incidentally detected during passage of the echoendoscope through the esophagus on the way to image gastrointestinal or pancreatic disease. Transesophageal EUS is well suited to image the posterior mediastinum but is unable to access or visualize much of the anterior mediastinum beyond cardiac structures. In lung cancer, the stage of disease determines prognosis and treatment. ^, Endosonographic evaluation and sampling of mediastinal lymph nodes is more accurate than CT or PET and less invasive than and superior to initial surgical staging, and guidelines recommend endosonography as the technique of choice when mediastinal staging is indicated. ^, ^, This chapter will discuss the EUS diagnosis of posterior mediastinal lymphadenopathy, masses, and cysts and the role of EUS in the diagnosis and staging of lung cancer.

EUS evaluation of enlarged posterior mediastinal lymph nodes

EUS appearance of posterior mediastinal lymph nodes

Appearance of benign posterior mediastinal lymph nodes

Mediastinal lymph nodes are commonly encountered during EUS for nonthoracic indications. The most common EUS feature of these benign lymph nodes is a triangular or crescent shape, with possibly an echogenic center ( Fig. 8.1 ). The echogenic center represents the hilum of the lymph node. Intranodal blood vessels also suggest benign lymph nodes. ^,

• Fig. 8.1, Benign mediastinal lymph node. Note the triangular appearance with a central hyperechoic stripe. AZ , azygos vein; LN , lymph node.

The prevalence of posterior mediastinal adenopathy varies with geographic region of the world, depending on the risk of endemic pulmonary infections. A prospective study from England and Sweden revealed that 62% of patients had posterior mediastinal lymph nodes, with a mean of 1.4 lymph nodes per patient. Nearly all of these lymph nodes had a short-axis diameter of 5 mm or less.

Appearance of malignant posterior mediastinal lymph nodes

EUS findings associated with malignant lymph nodes include round shape, short-axis diameter greater than 10 mm, hypoechoic echotexture, and well-demarcated borders ( Fig. 8.2 , ). ^, However, in mediastinal lymph nodes, if all four features are present, the chance of malignancy may be as low as 50% with only a 60% accurracy. For this reason, EUS findings alone are insufficient in evaluating for malignancy in enlarged mediastinal lymph nodes, and tissue sampling is important to obtain diagnostic cytopathologic material. This is particularly important when the EUS indication is for mediastinal staging of NSCLC as findings can change oncologic management.

• Fig. 8.2, Endoscopic ultrasonography-guided fine-needle aspiration of a malignant mediastinal lymph node.

Elastography has been reported in the evaluation of mediastinal lymph nodes and masses. However, the sensitivity and specificity (in the range of 80% to 90%) of this technique are lower than those of transesophageal or transbronchial EUS-guided FNA (>90% range). Therefore elastography needs further assessment and improvement before widespread use of this method can be recommended.

Technique for EUS and EUS FNA of posterior mediastinal lesions

Performing EUS of the posterior mediastinum

Cross-sectional imaging studies including CT, magnetic resonance imaging (MRI), and positron emission tomography (PET) exams of the chest and abdomen should be reviewed to identify the location of the lesion of interest and its relative location to surrounding structures. This will determine whether a lesion is accessible by transesophageal EUS FNA and assist in locating the lesion with the linear array echoendoscope. If the EUS is performed to evaluate enlarged mediastinal or periesophageal lymph nodes, it is often advisable to perform an esophagogastroduodenoscopy (EGD) with a forward-viewing scope first to assess for any esophageal or gastric pathology that may explain the lymphadenopathy. Identification of a luminal mass amenable to biopsy may preclude the need for EUS FNA. The linear echoendoscope is passed into the esophagus and stomach, and then ultrasound imaging is performed as the scope is withdrawn with sweeping motions in a clockwise and counterclockwise fashion to obtain a 360-degree view with the linear echoendoscope. The posterior mediastinum and the celiac axis lymph node regions are evaluated for lymphadenopathy, while the left adrenal gland and liver are examined for masses or ascites that can provide clues as to a primary lung cancer, lymphoma, or other advanced malignancy. In the diagnosis and staging of NSCLC, primary lung tumors can be evaluated when located adjacent to the esophagus and mediastinal lymph nodes stations are examined closely and systematically. The left adrenal gland is a predilection site of distant metastases.

During the EUS evaluation of enlarged mediastinal lymph nodes, there are usually several identified. These may be discrete, separate lymph nodes, which is often the case in malignancy, or a conglomeration of matted lymph nodes which can be seen in granulomatous disease. The location of each lesion is documented in terms of the distance in centimeters of the transducer tip from the incisors and the anatomic site (e.g., subcarinal, left paraesophageal, right paratracheal, aortopulmonary window) to assist in future identification and comparison on future examinations. For each lesion, the short-axis and long-axis dimensions are measured, and the degree of demarcation is described (well demarcated or poorly demarcated). The shape is described in terms of round, oval, triangular, draping, or matted. The echogenicity is described in terms of hypoechoic, hyperechoic, heterogeneous, or anechoic.

Performing EUS FNA of posterior mediastinal lymph nodes and lesions

Table 8.1 shows the types of pathologic lesions that can be diagnosed with transesophageal EUS FNA cytology. Transesophageal EUS FNA is generally performed using a linear array echoendoscope and usually a 22- or 25-gauge aspiration or core biopsy needle. Given the esophageal location of mediastinal FNA, the scope is generally straight, which allows for easier movement and maneuverability of the needle compared to transduodenal FNA where the scope is more angulated. If there is more than one possible lesion to sample for biopsy, the lesion that is most likely to be malignant (i.e., rounder, larger, more hypoechoic, more demarcated; Fig. 8.3 ) is chosen as the target. The presence of intervening blood vessels and distance from the esophagus should also be taken into account, as this may make FNA of certain lymph nodes easier and safer. The Doppler imaging function should be utilized to examine the FNA needle trajectory to ensure there are no significant vessels that may complicate FNA. Needle passage through an adjacent blood vessel can usually be avoided by displacing the esophagus with the scope tip to create a different needle path. While avoiding intervening vessels is the preferred method, some mediastinal lesions can only be sampled through a transvascular approach. Several databases and case reports have shown that transaortic or transpulmonary arterial EUS FNA of mediastinal lesions can be safely performed with over 70% yield. Once the lesion has been brought into view, the needle is passed through the esophageal wall and into the lymph node under constant ultrasound visualization. The internal stylet (optional) is then removed, intermittent suction may be applied, and the needle is moved back and forth within the lesion to sample the edges as well as the center of the lesion ( ).

TABLE 8.1

Posterior Mediastinal Lesions That Can Be Diagnosed With Endoscopic Ultrasonography-Guided Fine-Needle Aspiration

Malignant	Benign
Lung cancer	Reactive
Primary or metastatic	Granulomatous disease
NSCLC	Histoplasmosis
Small cell	Sarcoid
Mesothelioma	Tuberculosis
Lymphoma	Duplication cyst
Metastatic from nonlung primary	Leiomyoma
GIST	Mediastinitis/abscess
Spindle cell neoplasm	Pleural effusion

GIST , gastrointestinal stromal tumor; NSCLC , non–small cell lung cancer.

• Fig. 8.3, Appearance of a malignant mediastinal lymph node—note the hypoechoic, round, well-defined appearance and size greater than 10 mm.

The needle is then pulled out of the scope, the stylet is slowly reintroduced into the needle, and the aspirated material is slowly expressed either onto a microscope slide and/or into medium for cell block pathology or flow cytometry. The availability of immediate cytologic evaluation may increase the diagnostic yield and allow for fewer passes. ^, If immediate cytologic evaluation raises the possibility of lymphoma or lymphoma is suspected based on clinical history, additional passes should be obtained for flow cytometry, and fine-needle core biopsies should be considered. Several core needles are now available that are capable of obtaining larger samples of tissue for histopathology; these can provide tissue architecture that is usually necessary to differentiate the various subtypes of lymphoma. If immediate cytologic evaluation suggests infection, additional passes may be made for microbiologic studies. A final diagnosis is provided only after the cytopathologist has evaluated all processed specimen slides and cell block material. In general, diagnostic material from posterior mediastinal lesions can be obtained within two to five EUS FNA cytology passes, and one to two core biopsy passes.

Endobronchial ultrasound

EBUS TBNA has become increasingly available, especially as performed by interventional pulmonologists and thoracic surgeons. ^, EBUS provides unique access to lymph nodes and masses adjacent to the trachea, as well as the subcarinal and perihilar areas. The combination of transesophageal and transbronchial EUS provides nearly complete mediastinal evaluation and avoids risks of mediastinoscopy or other more invasive surgeries to obtain a tissue diagnosis. The role of EBUS and EUS in the diagnosis and staging of NSCLC will be further discussed in the subsequent section.

Accuracy of EUS FNA for diagnosing posterior mediastinal lesions

The overall accuracy rate for diagnosing posterior mediastinal malignancy with transesophageal EUS FNA cytology is approximately 93%. A meta-analysis of 76 studies ( n = 9310 patients) found a pooled sensitivity of 88% and specificity of 96%. Table 8.2 shows a summary of the accuracy rates for EUS FNA for diagnosing malignancy in posterior mediastinal lesions. Several studies showed that the diagnostic accuracy of malignant posterior mediastinal lymph nodes increases with the use of EUS FNA cytology over simple EUS appearance alone.

TABLE 8.2

Summary of Studies Evaluating the Operating Characteristics of Endoscopic Ultrasonography-Guided Fine-Needle Aspiration for Diagnosing Malignant Posterior Mediastinal Lesions

Authors (Year)	n	Sensitivity (%)	Specificity (%)	Accuracy (%)	PPV (%)	NPV (%)
Giovannini and coworkers ¹³⁷ (1995)	24	81	100	83	—	—
Silvestri and coworkers ¹³⁸ (1996)	27	89	100	—	—	—
Gress and coworkers ¹³⁹ (1997)	52	95	81	96	—	—
Hünerbein and coworkers ¹⁴⁰ (1998)	23	89	83	87	—	—
Serna and coworkers ¹⁴¹ (1998)	21	86	100	—	—	—
Vazquez-Sequeiros and coworkers ¹⁴² (2001)	82	96	100	98	94	100
Fritscher-Ravens and coworkers (2000)	153	92	100	95	—	—
Wallace and coworkers (2001)	121	87	100	—	—	—
Devereaux and coworkers (2002)	49	—	—	94	—	—
Larsen and coworkers ¹⁴³ (2002)	79	92	100	94	100	80
Hernandez and coworkers ¹⁴⁴ (2004)	59	—	—	84	—	—
Savides and coworkers (2004)	59	96	100	98	100	97
Eloubeidi and coworkers ¹⁴⁵ (2005)	104	93	100	97	100	97
Overall	91	97	100	97	99	94

NPV , negative predictive value; PPV , positive predictive value.

Risks of EUS FNA of posterior mediastinal lesions

EUS FNA of posterior mediastinal lesions is safe, with few complications reported in the thousands of patients described in retrospective and prospective trials. A pooled review of prospective studies revealed a 0.43% risk of complications with mediastinal EUS FNA, mostly pain, bleeding, and perforation. However, several cases of mediastinitis have also been reported after transesophageal EUS FNA. A case of mediastinitis with osteomyelitis has also been reported in the setting of FNA. Although most of these cases have involved mediastinal cysts, some patients with solid lesions (nodes or masses) have also developed post-EUS FNA mediastinitis.

There are two reported cases of esophageal wall seeding with tumor after EUS FNA of malignant posterior mediastinal lesions—one with multiple passes of a 19-gauge needle into a malignant node from primary gastric cancer, and the other with a 22-gauge needle into a mediastinal carcinoma of unclear origin. ^, A case of esophagomediastinal fistula formation after EUS FNA of a posterior mediastinal lymph node resulting from tuberculosis has also been reported. Two cases of esophageal wall rupture in the setting of FNA requiring thoracotomy have been reported. In rare situations where difficult-to-access mediastinal lesions are sampled using transvascular EUS FNA, a single asymptomatic aortic pseudoaneurysm has been reported, with other minor complications including incidental hematomas and self-limited bleeding. Despite rare reports of severe complications, EUS FNA and EBUS FNA of mediastinal lesions appear to be very safe. Two large, retrospective studies with 16,181 and 16,750 EUS and EBUS procedures with mediastinal FNA demonstrated a mortality of 0% to 0.04% and an adverse event rate of 0.22%. ^,

EUS FNA compared with other modalities for evaluation and biopsy of posterior mediastinal lymph nodes or masses

The noninvasive imaging modalities commonly used to evaluate enlarged mediastinal lymph nodes are CT and PET scans. These modalities have mostly been compared with EUS FNA in the setting of suspected lung cancer. Both EUS alone and EUS FNA have been shown to be more accurate than CT alone (using short-axis lymph node diameter >10 mm) for diagnosing malignant posterior mediastinal lymph nodes.

PET scanning detects increased uptake of the glucose analog 18F-2-deoxy- d -glucose. Increased uptake can occur both in malignancy and in inflammatory conditions. A meta-analysis comparing CT with PET scan for evaluation of mediastinal adenopathy in patients with lung cancer revealed that when the CT scan showed enlarged lymph nodes, the sensitivity of PET was 100%, but the specificity was only 78%, in contrast to no CT findings of lymph node enlargement (sensitivity of 82% and specificity of 93%). The low specificity of PET scan implies that 22% of patients with PET-positive enlarged mediastinal lymph nodes actually do not have malignancy (false-positive PET scan). Therefore these PET-positive lymph nodes should undergo tissue biopsy if it is critical to be certain about the diagnosis of malignancy in the nodes. Several studies confirmed the poor specificity of PET compared with transesophageal EUS FNA. ^, One large study found that the EUS FNA positive predictive value of malignancy was 100% compared with 40% for PET. Another study comparing EUS FNA to PET CT demonstrated superior performance of EUS FNA in the evaluation of mediastinal and upper abdominal lymphadenopathy. The sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and overall accuracy of EUS FNA were 91.3%, 100%, 100%, 92.5%, and 95.8%, respectively. The same values for PET CT were 75%, 25%, 50%, 50%, and 50%, respectively. The combination of PET and EUS FNA improves the specificity and overall accuracy as compared with PET alone. ^,

The other modalities for obtaining tissue samples from posterior mediastinal lesions are percutaneous CT-guided transthoracic FNA, bronchoscopy with transbronchial biopsy, EBUS with transbronchial FNA, and mediastinoscopy with biopsy. Percutaneous transthoracic FNA is generally not used for biopsy of posterior mediastinal lesions because of the risk of pneumothorax or puncture of a major vessel. The diagnostic yield of transbronchial FNA without EBUS is lower than that of EUS FNA, whereas EBUS has a similar diagnostic yield in the sampling of adenopathy locations visualized by both transesophageal EUS and EBUS. Mediastinoscopy is associated with greater difficulty (and potentially increased risk) in accessing the lymph nodes in stations that are the most easily visualized and biopsied with transesophageal EUS FNA (subcarina, posterior aortopulmonic window, and periesophageal stations). Therefore the less invasive EUS FNA and EBUS FNA have replaced mediastinoscopy at most referral centers.

Differential diagnosis of enlarged posterior mediastinal lymph nodes

Enlarged mediastinal lymph nodes are usually defined by CT findings of lymph nodes 10 mm diameter or larger. In the setting of a peripheral lung mass and mediastinal lymph nodes, the main concern is primary lung cancer with metastatic disease. The finding of numerous posterior mediastinal and hilar lymph nodes raises the question of whether the diagnosis is benign (sarcoid, histoplasmosis, tuberculosis, reactive) or malignant (especially lymphoma). Often the clinical history suggests the origin.

Malignant posterior mediastinal lymph nodes

The rate of diagnosis of malignancy with EUS FNA of posterior mediastinal nodes in patients without a known diagnosis of cancer varies depending on prior pulmonary evaluation and local referral patterns; however, it is approximately 50%, and most cancers are of pulmonary origin. ^, Table 8.2 shows the reported operating characteristics of EUS FNA for diagnosing malignancy in posterior mediastinal adenopathy. The overall sensitivity, specificity, and accuracy are greater than 90%.

You're Reading a Preview

Become a Clinical Tree membership for Full access and enjoy Unlimited articles

Become membership

If you are a member. Log in here