COMPARISON OF SURGERY AND RADIATION-THERAPY FOR NON-SMALL-CELL CARCINOMA OF THE LUNG WITH MEDIASTINAL METASTASIS

Interest in the potential role of induction chemotherapy for patients with marginally operable non-small cell carcinoma of the lung (NSCCL) led to a retrospective study of surgical resection and radiation therapy, alone or combined with each other and/or chemotherapy. All 169 patients seen at The University of Texas M. D. Anderson Cancer Center from 1980 through 1985 with evidence of NSCCL metastatic to ipsilateral mediastinal lymph nodes but without extrathoracic spread were evaluated (N2M0). All patients had histologic or cytologic confirmation of NSCCL and clinical or pathologic evidence of mediastinal involvement. Nine patients received CHM alone and were excluded. The male:female ratio was 3:1, and 50% were < 60 years old. Squamous cell carcinoma was reported in 42%, adenocarcinoma in 45%, large-cell carcinoma in 9%, and unclassified carcinoma in 4%. Radiation therapy (RT) was selected for 81 patients (+CHM in 56%), in 85% because of the extent of tumor involvement and in 15 for medical reasons. Of RT patients, 31% had a Karnofsky performance status (KPS) of less-than-or-equal-to 80, 30% had > 5% weight loss, and 9% had Stage IIIB disease. Surgical resection (SX) was used in 41 patients (+CHM in 41%), of whom 10% had KPS less-than-or-equal-to 80, 17% had > 5% weight loss, and 2% had Stage IIIB disease. SX + RT was the treatment for 38 patients (+CHM in 36%), of whom 13% had KPS less-than-or-equal-to 80, 13% had > 5% weight loss, and 13% had Stage IIIB disease. The proportions of patients with KPS less-than-or-equal-to 80 and weight loss > 5% were significantly greater (p < .01 and p < .05, respectively) in the RT group than in the other treatment groups. Actuarial survival rates at 2 and 5 years were 24% and 9%, respectively, for RT, 32% and 17% for SX, and 46% and 25% for SX + RT. Overall survival rates for all 160 patients were 30% at 2 years and t4% at 5 years. Prognostic factors that were found to be important were KPS (p = .027) and weight loss (p = .001); age, sex, histology, and Stage IIIa versus IIIB disease were not significantly related to outcome. The results of treatment with SX + RT were significantly better than with RT alone (p = .03); the difference between RT alone and SX alone was not significant (p = .39). There was no evidence that the addition of CHM to any of the treatments affected outcome. The data suggest that important differences in pretreatment characteristics among treatment groups may be of greater consequence in outcome than actual therapy. These characteristics and selection factors must be considered in reporting retrospective studies and in developing prospective, randomized trials comparing treatment modalities in patients who have NSCCL with mediastinal involvement.