Survival after recurrence of breast cancer.

A series of 456 patients with cancer of the breast that was first diagnosed between 1961 and 1970 were studied to determine the effect of certain prognostic factors on survival after recurrence. The prognostic factors studied included those recorded at the time of initial diagnosis and at the time of recurrence. Life-table and log-rank methods were used to analyse the influence of individual prognostic factors, and the simultaneous effects of multiple factors were assessed with Cox's proportional hazards model. The analysis indicated that, in addition to the strong effect of the length of the free interval and the nature and number of sites of recurrence, variables assessed at the time of initial diagnosis -- stage and histologic subtypes of the tumour -- also influenced survival after recurrence. This supports the contention that the presenting characteristics of breast cancer reflect the underlying biologic aggressiveness of the tumour as much as the length of time the tumour has been present.