Evaluation of the mediastinum by EUS began in the early 1990s and has recently been expanded by the use of endobronchial ultrasound (EBUS) following the same principle applied to the bronchoscope . Thus, using different ports of access, both gastroenterologists and pulmonologists evaluate the mediastinum in a complementary manner to perform an almost complete mediastinal nodal staging in an accurate, minimally invasive, and cost-effective approach. EUS findings impact patient management and surgery can be avoided in ~ 50 % of the patients.

Direct tumor extension and nodal staging in lung cancer are paramount for deciding therapy options. Definitions of practical use for the cytopathologist assisting a nonsmall-cell lung cancer (NSCLC) staging are given in Table 4.2, which summarizes the American Joint Cancer Commission (AJCC) guidelines for lung cancer staging .

The cytopathologist needs to be aware that metastasis to subcarinal or ipsilateral mediastinal LN (N2) implies stage IIIA disease, while contralateral LN metastasis (N3) or direct mediastinal invasion (T3) implies stage IIIB disease and may prevent a patient from undergoing surgery (Fig. 4.2).

Computerized tomography (CT) scan is the standard imaging modality for lung cancer, but is poor at staging the mediastinum (sensitivity , 50–70 %) and unreliable for detecting LNs < 1 cm. Positron emission tomography (PET) seems to be superior to CT (sensitivity, 67–100 %). Magnetic resonance imaging (MRI) ­often offers no advantages when compared with CT, and has higher costs. These imaging modalities usually do not procure tissue for diagnosis.

EUS and endoscopic ultrasound fine needle aspiration (EUS-FNA) have a sensitivity of 84–94 % in the evaluation of mediastinal LNs in approachable compartments, and confirm metastatic disease in up to 25 % of patients who show no evidence of mediastinal disease on CT scan (lesions < 1 cm). Similar statistical values, although with lower sensitivity and accuracy, are reported for EUS-FNA in restaging of NSCLC after induction chemotherapy, and are superior to CT or re-mediastinoscopy, which can be technically difficult to perform . The results of EUS-FNA change the management in ~ 80 % of the patients, avoiding thoracotomy/thoracoscopy in up to 49 % and mediastinoscopy in 68 % of cases. EUS-FNA is particularly useful in decreasing the morbidity of image-guided transthoracic FNA and has the potential to prevent unnecessary surgery in patients with mediastinal disease and negative CT scan results.

EUS -FNA reduces NSCLC staging costs by up to 40 % per patient because of the decrease in surgical procedures, namely, mediastinoscopy (~ $ 8000) or exploratory thoracotomy (~ $ 26,000). It is also important that EUS-FNA has 93 % sensitivity vs. 73 % sensitivity of surgical procedures, added to the fact that EUS-FNA can be performed in an ambulatory setting and has lower morbidity (Fig. 4.3) .

In summary, EUS -FNA has a high diagnostic yield and accuracy , detects occult metastases, often prevents unnecessary surgery, is cost-effective, and has fewer complications . Thus, initial evaluation with EUS-FNA is recommended instead of mediastinoscopy with biopsy, transbronchial FNA, CT-guided FNA, and PET. According to the European Society of Thoracic Surgery guidelines, mediastinoscopy remains the gold standard for superior mediastinal LNs in primary staging of NSCLC; EUS-FNA, EBUS-FNA, and transbronchial FNA are preferred modalities for primary staging of LNs in other locations. Aspiration cytology or surgical modalities may be used for restaging; if negative cytology results, a surgical approach may be indicated, depending on the clinical setting.

We should emphasize that EUS -FNA and EBUS-FNA are complementary in sampling almost all mediastinal compartments and may replace more invasive methods for diagnosing and staging NSCLC. EUS and EUS-FNA have access to paratracheal lymph nodes and lesions located in the posterior and inferior mediastinum . EBUS and EBUS-FNA have access to pretracheal masses and those adjacent to main bronchi particularly in the right side. Sensitivity, specificity , and diagnostic accuracy approach 100 % when both modalities are used .