Towards a minimally invasive staging strategy in NSCLC: analysis of pet positive mediastinal lesions by EUS-FNA

Abstract

Lung Cancer 2004;44:53–60. Annema JT, Hoekstra OS, Smit EF, Veselic M, Versteegh MIM, Rabe KF. Study Summary: This study addresses the yield of EUS-FNA for confirming mediastinal lymph node metastasis in NSCLC patients with a PET scan positive for mediastinal disease. The authors prospectively enrolled 36 suspected or confirmed NSCLC patients with positive findings on PET scan in the mediastinal area. Twenty-six (72%) patients also had >1-cm mediastinal lymph nodes on chest computed tomography (CT). Regardless of PET findings, EUS-FNA examination was performed for all lymph node stations accessible from the esophagus. The cytology specimen was obtained from lymph nodes >5-mm in diameter and was examined by an on-site cytologist. EUS-FNA confirmed mediastinal lymph node involvement in 25 (69%) patients. Subsequent surgical staging (mediastinoscopy 2 patients, thoracotomy 10 patients) showed malignancy in 3 patients and reactive lymph nodes in 7 patients. Thus, EUS-FNA correctly detected 25 of 28 (89%) of patients with N2/N3 disease. Location of lymph node metastasis was outside the reach of mediastinoscopy in 8 of these patients. Overall, the sensitivity, specificity, positive predictive value, negative predictive value, and accuracy of EUS-FNA were 93%, 100%, 100%, 80%, and 94%, respectively. There were no procedure-related complications. Comments: Several studies have used PET scan for mediastinal staging of patients with NSCLC and found it to be better than chest CT scan for this purpose. In 1 study, the sensitivity and specificity of PET scan for mediastinal staging were 91% and 86%, respectively (N Engl J Med 2000;343:254–261). An important limitation of PET scan is its high false-positive rate. For instance, in a recent study, more than one third of PET positive lymph node stations were later found to be false-positive (Ann Thorac Surg 2003;75:231–236). False-positive mediastinal lymph nodes on PET scan may lead to inappropriate denial of curative surgery. Therefore, the current recommendation is to confirm PET positive lesions with biopsy (Chest 2004;125:2300–2308). Mediastinoscopy is most commonly used for this purpose. This is the first study to combine PET with EUS-FNA for mediastinal staging. The study establishes EUS-FNA as an effective noninvasive alternative to mediastinoscopy for PET-positive patients. The study does not address the potential role of EUS-FNA in NSCLC patients with negative PET findings. An important observation was a false-positive PET in 7 of 36 (19%) of patients. This reaffirms the need for tissue diagnosis for all patients with a positive PET scan. Furthermore, 8 of 28 (29%) of PET positive lymph nodes were located in regions unsuitable for mediastinoscopy, again suggesting the complimentary value of EUS-FNA and mediastinoscopy in these patients. The safety and noninvasive nature of EUS-FNA makes it an attractive choice before subjecting the patient to a more invasive surgical staging. The future studies should address the cost-effectiveness of PET followed by EUS-FNA approach to mediastinal staging.