MULTIFACTORIAL ANALYSIS OF THE PROGNOSIS OF NONANAPLASTIC N2 OPERATED LUNG CANCERS

The marked disparity of the published 5-year survival rates (0 to 3 %) for non-anaplastic, operated N2 lung cancers would suggest the marked heterogeneity of the patients studied. We prospectively studies 97 N2 tumours for which << curative >> resection was performed between 1983 and 1986. The overall survival rate was 28% at 3 years, 19% at 4 years and 12% at 5 years. Survival was studied on the basis of several parameters concerning the tumour: histological type and degree of differentiation, T (TNM), modality of invasion (circumscribed, infiltrating), the quality of the stroma (absent, fibrotic, cellular), the mitotic index, the necrosis rate, presence or absence of microscopic invasion of blood vessels; and concerning the nodes : site, number (solitary, multiple), size of the metastasis (massive, microscopic), macroscopic or microscopic invasion of the capsule. We know the outcome for all of our patients: 81 have died and 16 are still alive. Sixty-one have developed metastases, 5 have developed local and regional recurrences and 3 have both a recurrence and metastases. Univariate analysis demonstrated that a small number of parameters significantly influence survival: microscopic invasion of blood vessel (chi-2 = 5.24, p < 0.02), macroscopic and microscopic invasion of the lymph node capsule. In order to more clearly define the prognostic factors, we used Cox's multidimensional model. This model was applied to the three significant parameters to which we added two other variables which non-significantly tended to be related to survival: tumour necrosis and number of lymph nodes involved. This analysis showed that the prognostic information can be sumarised by a single factor: microscopic invasion of blood vessels; the other factors are contingent or redundant. This analysis also demonstrated that the best theoretical 5-year survival rate which wa can hope to obtain for this group of patients is 26%. Is strict selection leading to good numerical results justified in a disease whose intimate mechanisms of progression are unknown and in which surgery, although unreliable, remains the only chance of cure?