New NTP centers meet the need to know.

The role of immunological mechanisms in controlling the growth of malignant tumours is controversial (Underwood, 1974) and, in the case of the common human malignancies, the evidence is slight. It is well recognised, however, that in breast cancer in particular there can be a long disease-free interval terminating in late recurrence, suggesting that important tumour-host interactions influence tumour growth kinetics. The frequent presence of an infiltrate of mononuclear cells in primary breast carcinomas has been taken as evidence of a host defence mechanism and in several studies the intensity of this putative response has been positively correlated with good prognosis (Hamlin, 1968; Bloom et al., 1970). Some 75% of the leucocytes in primary breast carcinomas are T-lymphocytes (Whitwell et al., 1984); their function in this site is unknown but may be either a nonspecific reaction to tumour necrosis or a more complex response to malignant cells. The peripheral blood lymphocyte count has also been reported to have a positive association with 5 year survival and disease-free survival rates (Papatestas et al., 1976). Immune competence, measured by a summary score derived from in vitro lymphocyte function tests and in vivo cutaneous reactivity, has been related to prognosis though in this study the peripheral lymphocyte count was not found to have prognostic value (Adler et al., 1980). However, a subsequent large investigation of potential prognostic factors in operable breast cancer indicated that a pre-operative blood lymphocyte count of 1.5 x 109 1 1 or less is predictive of early recurrence (Ownby et al., 1983). We have tested this hypothesis in a retrospective study of a consecutive series of women undergoing mastectomy and have sought to define further the relationships between peripheral blood lymphocyte count, nodal involvement, tumour size and disease-free interval. The study group consisted of 308 women with confirmed carcinoma of the breast who underwent mastectomy at Selly Oak Hospital between January 1978 and December 1982. All were judged on clinical grounds to have local disease only. Peripheral lymphocyte count was recorded pre-operatively. Node status was ascertained pathologically by routine axillary node sampling in 248 patients; the node status was considered to be unknown in the remainder for statistical analysis. Tumour size was measured from the surgically resected specimen in 285 cases and was unrecorded in the remaining 23 patients. Simple mastectomy was performed in all patients and clinical follow-up was undertaken according to a standard protocol. Patients were reviewed routinely every 3 months for 18 months after operation and then every 6 months indefinitely. Further investigations were undertaken if history or examination suggested recurrent disease, either at a scheduled review or at re-referral between such reviews.