Résumé
Introduction: Optimal preoperative treatment of stage IIB (Pancoast)/III non-small cell lung cancer (NSCLC) remains undetermined and a subject of controversy. The goal of our study is to confirm feasibility and pathological response rates after induction chemoradiation (CRT) in our community-based treatment center. Patients and methods: Patients were selected according to functional and resectability criteria. Induction treatment comprised 3D conformal 4500 cGy radiotherapy delivered to the primary tumor and pathologic hilar and/or mediastinal lymph nodes on CT scan with an extra-margin of 1–1.5 cm. Concurrent chemotherapy regimen was cisplatinum 20 mg/m2 d1–d5 and etoposide 50 mg/m2 d1–d5, d1–5 d29–33. Within 3–4 weeks after CRT completion, operability was re-assessed accordingly. Surgery was performed 4–6 weeks after CRT completion in patients (pts) deemed resectable. Inoperable pts were referred for a 20–25 Gy boost ±1 extra-cycle of cisplatinum + etoposide. Results: From 1996 to 2005, 107 pts were initially selected for treatment and received induction chemoradiation (stage IIB-Pancoast 18, IIIA 58 and IIIB 31, squamous cell carcinoma 48%, adenocarcinoma 44%, large-cell undifferentiated carcinoma 14%). After preoperative evaluation, 72 pts (67%) had a thoracotomy (pneumonectomy 21, lobectomy 45, bilobectomy 5) and all but one (unresectable tumor) had a macroscopic complete resection. During the 3-month postoperative time, five patients (6.9%) died, four after pneumonectomy (right 3, left 1). The analysis of tumoral samples showed a pathological complete response rate or microscopic residual foci of 39.5%. Median follow-up time was 22.3 months (survivors: 36.8 months), 2-year and 3-year overall survival rates were 55% and 40%, respectively (median = 26.7 months) for all the intention-to-treat population (n = 107), 62% and 51% (median = 36.5 months) for 71 resected pts, 41% and 16% for 36 non-resected pts (median = 19.1 months). On multivariate analysis, surgical resection and tumoral necrosis >50% (or pathological complete response) were the most pertinent predictive factors of the risk of death (hazard ratio = 0.50 and 0.48, p = 0.006 and 0.038, respectively). Conclusion: Surgery was feasible after induction chemoradiation, particularly lobectomy in PS 0–1, stage IIB (Pancoast)/III NSCLC pts but pneumonectomy carries a high risk of postoperative death (particularly, right pneumonectomy). Pathological response to induction chemoradiation was complete in 39.5% of patients and was a significant predictive factor of overall survival.