Lung cancer is the leading cause of death in Taiwan. The outcomes of the disease vary depending on early detection, histologic types and staging. Because the mediastinal involvement including lymph node status is a significant prognostic factor for survival, lymph node biopsy is necessary for clinical staging of some patients. Although fluorodeoxyglucose-positron emission tomography (FDG-PET) is suggested for precise evaluation of mediastinum, tissue proof of PET positive lesions are recommended due to its limited diagnostic specificity for identifying mediastinal metastases. Cervical mediastinoscopy remains the "gold standard" for mediastinal lymph node sampling. However, it is invasive, requires general anesthesia. Another new minimally invasive method of mediastinal biopsy is real-time endobronchial ultrasound-guided transbronchial needle aspiration (EBUS-TBNA). The aim of this study is to compare the accuracy of PET and EBUS-TBNA for correct staging of the mediastinum for lung cancer patients.

Lung cancer ranks among the most commonly occurring malignancies and currently is the leading cause of cancer-related cause worldwide including Taiwan [1, 2]. Although a lot of research focus on the treatment of lung cancer, the prognosis of lung cancer remains dismal and a five year survival ate is less than 15% [3]. Unfortunately, early detection of lung cancer is still a problem. In a tertiary care hospital in Taiwan, only 27.3% of patients could received operation (stage I 15%, stage II 7.5%) [4]. Lymph node staging is also important for evaluation the possibility of operation.

Fluorodeoxyglucose-positron emission tomography (FDP-PET) is now used by oncologist to evaluate lung masses, solitary pulmonary nodules and intrathoracic lymph nodes. As the technique becomes more widespread, it is now used even as a first line imaging investigation. Although PET has a high negative predictive value, it is neither sensitive nor specific to differentiate benign from malignant mediastinal lymph nodes [5, 6]. If PET positive mediastinal lymph nodes are equal to malignant involvement, some patients might be excluded from potentially curative surgery. Several national guideline groups suggest that PET positive lymph nodes should be biopsied if it is likely that the result will alter clinical management [7, 8].

"Cervical mediastinoscopy" has been regarded as the "standard procedure" for sampling mediastinal lymph nodes. However, these techniques require general anesthesia and could not be repeated because of adhesion. Access to hilar nodal stations can be difficult for mediastinoscopy. In recent years, one minimally invasive method endobronchial ultrasound-guided transbronchial needle aspiration (EBUS-TBNA) was used for biopsy of both hilar and mediastinal lymph nodes [9]. EBUS-TBNA allows the mediastinal lymph nodes to be targeted in the areas accessible to cervical mediastinoscopy, as well as some hilar nodes (lymph node stations 2-4, 7, 10-12)[9] .

Kazuhiro Yasufuku had published the first report of real-time EBUS-TBNA in evaluating mediastinal lymphadenopathy in 2004 [10]. Currently, the main indication of EBUS-TBNA is the mediastinal nodal staging of NSCLC after recent meta-analyses established the comparable sensitivity and specificity of nodal staging by EBUS-TBNA and cervical mediastinoscopy [11]. Efficacy in evaluation of other disease processes such as sarcoidosis and lymphoma has also been established [12].

Although there were several large studies to compare the diagnostic efficacy of mediastinal malignant lymphadenopathy between FDG-PET and EBUS-TBNA, the investigators need to have our own data because of high incidence of TB lymphadenitis in Taiwan, where the diagnostic accuracy of PET may be lower than other countries.

Inclusion Criteria:

1. Age older than 18 years
2. Patient with suspected malignant mediastinal lymphadenopathy

Exclusion Criteria:

1. Age younger than 18 years
2. Bleeding diathesis(INR > 1.4, Platelet count < 10,000/mcl)

Suspected malignant mediastinal lymphadenopathy