Endoscopic Ultrasound in Non-Small Cell Lung Cancer and Negative Mediastinum on Computed Tomography

Lung cancer is our nation's leading cancer killer of men and women (1-2). An unpublished report from the Society of Cardiothoracic Surgeons of Great Britain and Ireland (1985) revealed that approximately one in five thoracotomies performed for lung cancer were "open and close." Certainly, thoracotomy for unresectable non-small cell lung cancer (NSCLC) is not only costly but would be expected to impact negatively on the patient's quality of life.

The accuracy of computed tomography (CT) staging of lung cancer varies widely (3-10). Because of the lack of contrast within the mediastinum and signal averaging, CT detects abnormal lymph nodes that are greater than 1 cm. However, lymph nodes that are undetected by CT may harbor metastases, and benign lymph nodes are often greater than 1 cm (5-19). Based on CT findings, approximately 23% of NSCLC patients are overstaged, and 19% are understaged (20). Although many studies have shown the benefit of positron emission tomography (PET) in staging NSCLC (21-32), Fritscher-Ravens and colleagues demonstrated an advantage of endoscopic ultrasound (EUS)-guided fine-needle aspiration (EUS-FNA) compared with CT and PET in detecting small mediastinal metastases (33, 34). EUS-FNA combined with polymerase chain reaction analysis of mediastinal lymph nodes is also useful in detecting cancer-specific mRNA (35). In addition, new technologies such as endobronchial ultrasonography have shown some promise in staging centrally located early squamous cell lung cancers (36).

Differentiating extent of mediastinal nodal disease remains a challenge in staging NSCLC. Mediastinoscopy remains the standard of care in staging the mediastinum, however, at our institution; the use of mediastinoscopy for staging lung cancer has decreased by 20%. Certainly, EUS is complimentary to mediastinoscopy and is not used in place of mediastinoscopy for lung cancer staging. However, EUS is useful in identifying patients with positive posterior mediastinal lymph nodes, particularly N2 and N3 disease, and positive lymph nodes as small as 3 mm in size can be detected (33, 37-60). Endoscopic ultrasound has also been demonstrated to influence subsequent rates of thoracic surgery (59). Patients with N2 (positive ipsilateral mediastinal lymph nodes stations 1 through 9) disease may benefit from surgery, whereas patients with N3 (contralateral mediastinal lymph nodes) disease do not (61, 62). In addition, identifying patients with N2 disease is significant as they may benefit from neoadjuvant chemotherapy (63, 64).

The role of EUS in staging NSCLC without lymphadenopathy on CT, however, remains unclear. Patients with enlarged mediastinal lymph nodes on CT often are further evaluated with mediaslinoscopy, PET, or endoscopic ultrasound, whereas patients without evidence of mediastinal lymphadenopathy on CT do not undergo further interrogation of the mediastinum. Certainly, occult cancer may exist in both types of patients. The aim of this study was to evaluate the clinical impact of EUS in NSCLC patients without mediastinal lymphadenopathy on CT. The hypothesis for this investigation was that patients without mediastinal lymphadenopathy on CT might harbor occult metastases that may preclude unnecessary surgery if EUS identified occult malignancy in contralateral (N3) mediastinal lymph nodes or at distant sites. Findings related to this research have been presented in the form of an abstract (65).

METHODS

The institutional review board at Indiana University approved this protocol. Patients signed informed written consent. The funding agency played no role in the study or writing of this manuscript. NSCLC patients without mediastinal lymphadenopathy on CT were enrolled if they were operative candidates. Three patients with ipsilateral hilar lymphadenopathy (N1) on CT scan were included, as these lymph nodes are hilar and not mediastinal. All patients underwent a chest CT scan with intravenous contrast. However, not all patients underwent PET. Mediastinal lymph nodes on CT with a short axis diameter greater than 1 cm were considered abnormal, rendering a patient ineligible for enrollment.

Definitions

By EUS criteria, generally accepted echo characteristics for suspicious lymph nodes include more than 5 mm in short axis dimension, round, hypoechoic, and well demarcated. Lymph nodes were defined as positive by EUS if all four ultrasound features were present. EUS-FNA was performed if two or more ultrasound features were present. A change in patient management was defined as avoidance of unnecessary surgery or other deviation from the management plan that would have occurred as a result of the CT findings alone.

EUS Procedure

The EUS examination was performed on an outpatient basis using conscious sedation. A radial echoendoscope (GF-UM20 or GF-UM130; Olympus America Inc., Melville, NY) was introduced into the patient's stomach, where continuous sonographic imaging of the posterior mediastinum, left lobe of the liver, left adrenal gland, retroperitoneum, and celiac region were obtained. EUS-FNA was performed using a linear array echoendoscope (FG-32 UA or FG-36 UX; Pentax Precision Instruments Corp., Orangeburg, NY). FNA was performed on sonographically suspicious lymph nodes and lesions under direct real-time ultrasound imaging using a 22-gauge needle (Wilson-Cook, Winston-Salem, NC) that was passed through the echoendoscope through the esophageal wall or stomach. A cytotechnologist who was present during the procedure prepared the aspirate.

Patients were precluded from surgery if EUS-FNA histologically confirmed contralateral lymph node metastases (N3) or distant metastases. Thoracotomy and complete lymph node dissections were performed in accordance with American Thoracic Society guidelines and the American Joint Committee on Cancer. The surgeon was blinded to the results of the EUS examination to minimize bias unless findings were present that precluded the patient from surgery. Cytology and clinical follow-up were both used as reference standards in patients that did not undergo surgery.

Data Analysis

Using surgical pathology and EUS-FNA cytology results as the reference standards, the sensitivity of EUS imaging alone was calculated as the proportion of malignant mediastinal lymph nodes correctly identified by EUS. The specificity of EUS was calculated as the proportion of benign mediastinal lymph nodes correctly identified by EUS. Exact 95% confidence intervals for sensitivity and specificity of EUS were reported. Fisher's exact test was used to test the null hypothesis that the proportion of patients with mediastinal lymph node metastases correctly identified by EUS did not differ by location (upper or lower lobe). Malignant mediastinal lymph nodes were also stratified by location of the primary tumor.

RESULTS

There were 76 consecutive patients enrolled during a 3-year period (September 1997 to August 2000). Our cohort consisted of 48 men and 28 women between the ages of 33 and 89 (mean 61 years). Four patients were lost to follow-up or later refused to participate, leaving 72 patients in our study. Sixty-two (86%) patients underwent surgery, and 10 (14%) did not. Of the 10 patients that did not undergo surgery, 1 refused and 9 were precluded based on EUS-FNA findings: five patients were staged IIIb (N3 disease), three were staged IV, and one had a synchronous esophageal cancer (Table 1). EUS-FNA results altered management in 18 of 72 (25%) NSCLC patients (Table 2). There were no EUS- or CT-related complications.