A randomized trial comparing endosonography followed by surgical staging versus surgical mediastinal staging alone in non-small cell lung cancer: The ASTER study

Reporter: J. Nick Lukens, MD
The Abramson Cancer Center of the University of Pennsylvania
Last Modified: June 5, 2010

Presenter: K.G. Tournoy, MD
Ghent University Hospital, Ghent, Belgium


  • Prognosis of patients with non-small cell lung cancer (NSCLC) depends heavily on nodal staging (Rusch, Journal of Thoracic Oncology 2007, 2:603-12)
  • Appropriate treatment of NSCLC depends on accurate mediastinal lymph node staging. Non-operative strategies may be preferred for N2/3 disease, while N1 disease may be treated with up-front, definitive surgery.
  • Invasive mediastinal staging in patients with potentially resectable NSCLC is a standard of care. This staging technique, while associated with potential risk to the patient, allows accurate, pathologic assessment of the mediastinal lymph nodes. This may allow avoidance of unnecessary thoracotomy in patients who have mediastinal lymph node disease.
    • Staging of the mediastinum by PET/CT is not sufficiently accurate to deny patients potentially curative surgery, and needs to be confirmed by invasive means.
  • Techniques that may be employed as part of invasive mediastinal staging include:
    • Endoscopic techniques (trans-esophageal FNA and endobronchial ultrasound with transbronchial needle aspiration (EBUS-TBNA)).
    • Surgical staging (mediastinoscopy, mediastinotomy or video-assisted thoracoscopic surgery)
  • Endosonography remains a newer, less-seasoned technique; for this reason, surgical staging of the mediastinum is commonly performed after negative endosonography for confirmation of negative sonographic results.
  • In the modern era, at least one study suggests that the diagnostic yield of endoscopic staging may surpass that of mediastinoscopy (Ernst, J Thorac Oncol, 2008)
  • The purpose of the study presented here was to test the hypothesis that endosonography followed by surgical staging (for negative findings) has a higher sensitivity than surgical staging alone in mediastinal lymph node staging.


  • Design: prospective randomized multicenter trial.
  • Eligibility: Consecutive patients with NSCLC in whom mediastinal lymph node invasion (either N2 or N3) was suspected based on the available thoracic imaging (PET/CT).
    • Enlarged mediastinal nodes
    • Enlarged hilar nodes with a central tumor
    • FDG-avid small (<1cm) mediastinal nodes
  • Tumors were required to be otherwise resectable, and the patients to be operative candidates. Patients with evidence of metastatic disease on PET/CT scan were excluded from this study.
  • Patients were randomly assigned between:
    • Endosonographic staging (ES): Combined trans-esophageal and endobronchial ultrasound (EUS-FNA and EBUS-TBNA) followed by surgical staging (if negative findings) versus
    • Surgical staging (SS): mediastinoscopy, mediastinotomy or video-assisted thoracoscopic surgery alone.
  • Thoracotomy with systematic mediastinal lymph node dissection was performed in the absence of mediastinal metastases after surgical staging in both arms.
  • Primary endpoint: detection of nodal metastasis (N2/3 disease).
  • Secondary endpoints: the rates of futile thoracotomy and development of complications.


  • 241 patients were randomized as planned (of 357 assessed):
    • 123 to ES+SS and 118 to SS alone.
    • The groups were well balanced in terms of age, gender, and T stage and N stage by PET/CT.
  • Nodal metastases were found in:
    •  62 patients (50%) by ES + SS (56 by EUS/EBUS + 6 by subsequent SS) vs.
    • 41 patients (35%) by SS alone (p = 0.019).
    • The sensitivity for mediastinal metastases for ES+SS was 94% (95% CI, 85- 98) versus 80% (95 CI, 68-89) for SS alone (p = 0.042).
    • The negative predictive value of ES+SS was 93% vs. 86% for SS alone (p = NS)
  • Thoracotomy was futile in:
    •  8 patients (7%) staged by ES+SS vs.
    • 21 patients (18%) of those staged by upfront SS (p = 0.009).
  • The rate of complications during staging was similar in both arms (6 vs. 7 patients, p = 0.8), however 12/13 patients who developed complications developed them in the setting of SS
    • 1 patient in the ES arm had a pneumothorax.
  • There was a 47% reduced rate of need for surgical staging in patients undergoing upfront ES.

Author’s Conclusions

  • The authors recommend initiating mediastinal staging with endosonographic techniques in potentially resectable NSCLC because of:
    1. Improved detection of nodal metastasis (increased sensitivity)
    2. Reduced rate of futile thoracotomy
    3. Lack of increase in complication rates compared to surgical staging alone.

Clinical Implications

  • For patients with potentially resectable NSCLC but radiographic evidence of mediastinal lymph node invasion by PET/CT, this study provides compelling evidence that invasive staging of the mediastinum should begin with endoscopic evaluation, followed by surgical staging for negative findings, prior to thoracotomy.
    • This approach is more sensitive for detection of N2/3 disease, without an increased risk of complications.
    • It may decrease the need for more invasive surgical staging in some patients.
  • While prior studies have compared the diagnostic yield of endoscopic staging to surgical staging (Ernst, J Thorac Oncol, 2008), this is the first study to directly compare these two approaches of ES followed by SS for negative findings versus SS alone.
  • Strengths: this was a randomized, appropriately powered, and clinically relevant study.
  • Unanswered questions:
    • If the EBUS is positive and mediastinoscopy is not performed, does this adversely affect subsequent treatment?
      • Sampling of multiple lymph node stations by mediastinoscopy guides involved-field external beam radiation planning, so will this adversely affect radiation planning?
      • Does EBUS alone provide enough tissue for molecular studies to guide targeted therapeutics?
    • Is there a need for trans-esophageal EUS in addition to EBUS?
    • Is there a subset of patients who do not require surgical staging following negative endosonographic evaluation?
    • What are the performance characteristics of PET/CT compared to invasive staging in this study?
  • We await the outcome of confirmatory studies to validate the findings of this compelling study.