A Randomized Trial of Radical Surgery versus Thoracic Radiotherapy After Response to Induction Chemotherapy in Patients with Histo-/Cytologically Proven Irresectable Stage IIIA-N2 NSCLC (EORTC 09841)

Reviewer: S. Jack Wei, MD
University of Pennsylvania School of Medicine
Last Modified: May 17, 2005

Presenter: J. van Meerbeeck
Presenter's Affiliation: EORTC Lung Cancer Group
Type of Session: Scientific


  • Stage IIIA non-small cell lung cancer (NSCLC) with stage N2 disease is a heterogeneous group of patients representing 15% of all cases of NSCLC at presentation.
  • Variable treatment approaches have been attempted with 5-year overall survival (OS) as shown:
    • Surgery: 5%
    • Chemotherapy + Radical Radiotherapy (RT): 15%
    • Chemotherapy + Resection: 18%
    • Concurrent Chemoradiotherapy (CTRT): 20%
    • Concurrent CTRT + Resection: 27%
  • This trial was conducted to determine if surgical resection after response to induction chemotherapy improves outcomes in patients with IIIA-N2 NSCLC

Materials and Methods

  • Patients with stage IIIA-N2 NSCLC were treated with 3 cycles of platinum-based induction chemotherapy (cisplatin (>=80 mg/m2) or carboplatin (AUC>=5) combined with at least one other drug) and then assessed for response.
  • Patients with stable or progressive disease were treated off study, while patients with any type of response were randomized to:
    1. Thoracic RT (60 Gy in 2 Gy/day)
    2. Surgical Resection with intent of radical resection of tumor and dissection of hilar and mediastinal lymph nodes within 6 weeks of randomization
  • Post-operative RT was recommended in cases of incomplete resection (either positive surgical margin or highest sampled mediastinal lymph node positive) with dose of 56 Gy (2 Gy/day) 4-10 weeks after resection.
  • Eligible patients had cytologically or histologically proven N2 disease with clinical stage IIIA disease, were considered unresectable by a local surgeon, had measurable disease, and were fit for combined modality therapy.
  • Primary endpoint was OS.
  • Secondary endpoints included progression-free survival (PFS) and toxicity.


  • 579 patients registered with 332 (57%) randomized.  The majority of non-randomized patients were non-responders to induction chemotherapy.
  • 94% of patients in RT arm received RT, and 92% in surgery arm received surgery.
  • Treatment arms were well-balanced for gender, age, histology, stage, and response to induction chemotherapy.
  • For the induction chemotherapy, overall response rate (OR) was 62% with 4% complete response (CR), and 58% partial response (PR).
  • In the RT arm, 80% received >=60 Gy
    • 8% late 3/4 pulmonary toxicity.
  • In the surgery arm, 47% underwent pneumonectomy, 38% underwent lobectomy, and 21% underwent thoracotomy only.
    • 50% received complete resection
    • 4% operative mortality
    • 56% have ypN2 disease at time of resection, 42% have ypN0-1 disease.
  • Median follow-up: RT 73.1 mo, surgery 67.2 mo.
  • Median survival for RT vs. surgery: 17.5 mo vs. 16.4 mo (p=NS)
  • 5-year OS: 14.0% vs. 15.7% (p=NS)
  • Median PFS: 11.3 mo vs. 9.0 mo (p=NS)
  • 2-year PFS: 24.2% vs. 26.5% (p=NS)
  • Patients undergoing pneumonectomy and those with pathologic N2 disease after induction chemotherapy had worse outcome.

Author's Conclusions

  • Surgical resection after response to chemotherapy does not improve outcomes in stage IIIA-N2 patients compared to RT.
  • Surgery is feasible and has acceptable rate of post-operative mortality in this setting.
  • Induction chemotherapy results in mediastinal downstaging in 40% of patients with resected disease, a minority of which have CR.
  • Mediastinal downstaging at surgery predicts improved outcome if pneumonectomy is avoided.
  • Chemoradiotherapy is to remain the standard treatment for future EORTC studies in stage IIIA-N2 NSCLC

Clinical/Scientific Implications

Currently, the largest amount of controversy in NSCLC revolves around the treatment of stage IIIA patients.  The current study addresses the question as to whether radiotherapy should remain an integral part of the management of these patients.  The substitution of radiotherapy for surgery following induction chemotherapy alone failed to improve outcomes in patients with N2 disease and appeared to be somewhat worse than patients receiving RT, although non-significantly so.  It is important to note that there was a 4% post-operative mortality rate in the patients undergoing surgery, while there was an 8% rate of grade 3/4 late lung toxicity in the RT arm.  Given the similar disease control between the two arms, it would be interesting to examine the quality of life in these patients.  Given an almost 50% rate of pneumonectomy in patients in the surgery arm, it is likely that there is a high rate of pulmonary symptoms in this group; however, this data was not reported. 
The results of this study are consistent with the results of RTOG 93-09 which was first reported at ASCO 2004 and updated at this meeting.  In that study, patients underwent induction chemoradiotherapy (CTRT) and were randomized to completion of definitive CTRT versus surgical resection in stage IIIA NSCLC patients.  Although there was a PFS benefit with surgery, there was no OS benefit.  Interestingly, an exploratory analysis in that study showed that patients who required a pneumonectomy had a near-significantly worse outcome than patients receiving definitive CTRT, while patients who were able to receive a lobectomy did significantly better than patients undergoing CTRT.  It appears from the current study that induction chemotherapy alone is not enough to produce tumor regression sufficient to avoid pneumonectomy.  Because of this, CTRT remains the standard treatment with the approach of induction chemotherapy.  However, in patients who can tolerate concurrent induction CTRT, given the results of RTOG 93-09, it appears that surgical resection should be the treatment of choice as long as the extent of surgery can be limited to a lobectomy.