Patterns of disease failure after trimodality therapy of nonsmall cell lung carcinoma pathologic stage IIIA (N2). Analysis of Cancer and Leukemia Group B Protocol 8935.

BACKGROUND: The impact of sequential trimodality therapy on the pattern of first site disease failure in pathologic Stage IIIA (N2) nonsmall cell lung carcinoma (NSCLC) was analyzed.

METHODS: Seventy-four eligible patients with histologically documented Stage IIIA (N2) NSCLC underwent sequential trimodality therapy on Cancer and Leukemia Group B (CALGB) Protocol 8935. Treatment consisted of 2 cycles of induction cisplatin at 100 mg/m2 intravenously (i.v.) (Days 1 and 29) and vinblastine at 5 mg/m2 i.v. weekly for 5 weeks followed by surgery. Surgery included a thoracotomy with resection of the primary tumor and hilar lymph nodes and a mediastinal lymph node dissection. Patients with resected disease then received an additional a cycles of cisplatin at 100 mg/m2 i.v. and vinblastine at 5 mg/m2 i.v. biweekly for 2 total of 4 doses followed by consolidative thoracic irradiation. Patients with completely resected disease received 54 Gray (Gy) whereas those with incompletely resected disease received 59.4 Gy at 1.8 Gy/fraction (fx) once a day. Patients with unresectable disease underwent thoracic radiation therapy (TRT) treatments only to 59.4 Gy at 1.8 Gy/fx without any additional chemotherapy. Disease recurrence was determined by clinical, radiographic, or histologic criteria. Pattern of disease failure was identified by site of involvement at first recurrence as indicated by the CALGB Respiratory Follow-Up Form.

RESULTS: Sixty-three of the 74 patients completed the induction chemotherapy as planned. Forty-six of the 63 patients underwent resection of disease whereas the remaining 17 were unresectable. Thirty-three of the 46 resected patients completed the entire adjuvant postoperative chemoradiation treatment as planned. Ten of 17 patients with unresectable disease completed postsurgical TRT. At a median follow-up interval of 27 months (range, 4-43), the 3-year overall survival and failure-free survival were 23% and 18%, respectively, for all 74 eligible patients. Overall, disease failure has occurred in 52 (70%) of the 74 eligible patients: local only: 13 (25%); distant only: 16 (31%); and both local and distant: 23 (44%), (P = not significant [NS]). Ten patients progressed during induction chemotherapy: local only: six patients; and both local and distant failure: four patients. Twenty-eight of 46 resected patients recurred: local only: 1 (4%); both local and distant failure: 11 (39%); and distant only: 16 (57%); (P < 0.001). Disease progression occurred in 14 of 17 patients with unresectable disease: local only: 6; both local and distant sites: 8. Among the 52 total patients experiencing disease relapse, isolated or combined local failure occurred commonly among patients during induction chemotherapy (n = 10, [28%]), in those with unresectable disease (n = 14, [39%]), or in those with resected disease (n = 12, [33%]), (P = NS). However, isolated or combined distant failure was more likely to occur among patients with resected disease (n = 27, [69%]) than either during induction chemotherapy (n = 4, [10%]) or in those with unresected disease (n = 8, [21%]), (P < 0.05). Among patients who relapsed, brain metastases occurred in 13 of 52 (25%) patients overall and in 12 of 28 (43%) patients with resected disease.

CONCLUSIONS: Overall, disease failure was just as likely to occur in local, distant, or combined sites on CALGB Protocol 8935 using sequential trimodality therapy in the treatment of pathologic Stage IIIA (N2) NSCLC: Isolated or combined local failure occurred commonly during sequential tri-modality therapy whereas isolated or combined distant relapse was prevalent among patients with resected disease. In addition, isolated local failure was rare among patients with resected disease. The pattern of disease failure on CALGB Protocol 8935 reflects the biology of locoregional NSCLC as much as the therapeutic impact of trimodality therapy.