Sir James Young Simpson Memorial Lecture 1995. Breast cancer prevention: a pathologist's approach

Few would argue that there is not substantial room for an improvement in breast cancer practice. It has to be borne in mind that the United Kingdom experience in the 1980s was a 50:50 survival to death ratio in the 10 years after diagnosis. Preliminary analysis of the effects of mammographic screening suggests that there will be a real but small fall in overall mortality. Existing practice involves three stages: first, the 'earliest' detection of a lump by palpation or imaging; second, diagnosis by histopathology; and third, treatment by surgery, etc. Evidence is given that the limited success of existing practice could be due, in part, to a failure to recognize the precancerous state of the mammary tissue as a whole in cancer cases; and a failure to exploit this state for earlier diagnosis. In support of these contentions, comparative data from the microscopy of cancer-associated breasts and age-matched normal breast are given. There is a gross excess of focal hyperplasia in premenopausal cancer-associated breast tissue. Further, epidemiological data are consistent in that the tissue is subject to a sixfold increase in the risk of further primary carcinogenesis. A method is presented for detecting the cancer-associated breast. It exploits the breast menstrual cycle, a subject which is reviewed in extenso. Physiologically the premenopausal mammary tissue goes into a monthly pregnancy rehearsal with glandular proliferation and increased blood supply. The latter effects a luteal heat cycle, which can be measured readily by an electronic thermometric bra as increased breast surface temperature (1 degree C). Data are presented in terms of 50 normal breasts and 41 cancer-associated breasts studied daily (with progesterone assays) for one menstrual cycle. The cancer-associated breasts exhibit an absent or altered response to endogenous progesterone during the luteal phase of the menstrual cycle. The abnormality in the luteal heat cycle is maximal during the few days just after ovulation. Our data indicate that a 1-h clinical test at this time achieves a sensitivity of 71% and a specificity of 80% for "clinically normal' yet cancer-associated breast tissue. Such patients would be candidates for increased surveillance and chemoprevention.