PHILADELPHIA – Endobronchial ultrasound–guided biopsy of mediastinal lymph nodes in patients with operable non–small cell lung cancer worked as effectively for staging as did the standard approach – mediastinoscopy – in the first head-to-head comparison of the two methods.
"Our results showed that EBUS-TBNA [endobronchial ultrasound–guided transbronchial needle aspiration], when performed as in this study, can replace mediastinoscopy for accurate staging of the mediastinum in potentially resectable lung cancer," Dr. Kazuhiro Yasufuku said at the annual meeting of the American Association for Thoracic Surgery.
Dr. Kazuhiro Yasufuku
Based on these results, which were obtained in 153 patients treated by any one of seven surgeons working at Toronto General Hospital, Dr. Yasufuku and his colleagues now routinely use EBUS-TBNA as their initial approach for staging patients with inoperable non–small cell lung cancer (NSCLC), who account for about 70% of all NSCLC patients they treat. As long as they can collect adequate cell specimens for cytologic analysis from the lymph node stations they routinely assess, they rely exclusively on EBUS-TBNA for staging, which allows them to avoid mediastinoscopy for most of their patients, Dr. Yasufuku said in an interview.
"We knew that EBUS-TBNA was good, but [until now] we never knew how it compared with the gold standard, mediastinoscopy," he said. The major limiting factor is lymph node size, he noted. Surgeons find it challenging to routinely obtain an adequate cell specimen from nodes smaller than 5 mm in diameter, Dr. Yasufuku said. "The smaller the node, the harder it is to put a needle into it."
The Toronto group uses rapid, onsite cytologic evaluation, which means that a cytologist attends the procedure in the combined surgical and endoscopy suite. In the study, and also in routine practice, "we can make repeated needle passes until we obtain good specimens. The surgeon can learn how to place the needle by getting immediate feedback" on the specimens, he said.
The specimens obtained allow for a tissue diagnosis, and typically provide enough material to assess cells for the presence of epidermal growth factor receptor mutations, he added.
EBUS-TBNA uses local rather than general anesthesia, is less invasive, and has fewer complications compared with mediastinoscopy, said Dr. Yasufuku, a thoracic surgeon and director of the interventional thoracic surgery program at Toronto General and the University of Toronto.
The study enrolled adults with NSCLC who required mediastinoscopy as part of their staging to determine their suitability for lung cancer resection. The study excluded patients who were not fit for definitive surgical resection, because the researchers used the status of the surgically excised lymph nodes as the basis for judging the diagnostic accuracy of both techniques.
During July 2006–August 2010, they enrolled 153 patients with an average age of 69 years. The most common NSCLC histologic subtype was adenocarcinoma (59%), followed by squamous cell carcinoma (25%). Staging by ultrasound imaging identified 57% of the patients with stage I or II disease, and 39% with stage IIIA disease. The remaining 4% had stage IIIB or stage IV disease.
All patients underwent general anesthesia. A surgeon first performed EBUS-TBNA on each patient, followed immediately by mediastinoscopy. All patients then underwent surgical lymph node resection to definitively assess their nodes if EBUS-TBNA, mediastinoscopy, or both did not show signs of metastatic disease.
The surgeons attempted biopsies at five lymph node stations in each patient: stations 2R, 2L, 4R, 4L, and 7. They successfully biopsied an average of three stations per patient using EBUSTBNA, with an inadequate specimen obtained on an average of one station per patient. Average lymph node diameter on the short axis was 7 mm, and the procedure averaged a total of 20 minutes per patient. Overall, EBUS-TBNA identified 78 biopsies as malignant. During mediastinoscopy, surgeons successfully biopsied an average of 4 nodes per patient, with inadequate specimens obtained from 10 nodes, an average of fewer than 0.1 inadequate specimen per patient. Mediastinoscopy retrieved 79 biopsies that were identified as malignant.
The surgeons reached an identical and correct diagnosis using both modalities in 136 patients (89%). Neither modality produced the correct diagnosis in four patients (3%), which meant that overall EBUS-TBNA and mediastinoscopy agreed 92% of the time. EBUS-TBNA was correct and mediastinoscopy incorrect in seven patients, and mediastinoscopy was correct and EBUS-TBNA incorrect in six patients.
These outcomes meant that EBUS-TBNA had 81% sensitivity, 91% negative predictive value, and 93% diagnostic accuracy. Mediastinoscopy led to 79% sensitivity, 90% negative predictive value, and 93% accuracy. Both methods had a specificity and positive predictive value of 100%, Dr. Yasufuku said.