[Natural history of human cancer and prognostic factors. The example of breast cancer].

The most important event during the growth of a human cancer is the occurrence of distant spread. A model of the natural history of breast cancer was developed which was used for several aims. It has been possible to study the relationship between the size of the breast tumor and the probability of dissemination either in regional lymph nodes or in distant sites. It was found that the distribution of tumor volume at metastatic dissemination is log normal with a median (termed V50) of 23.6 ml (diameter 3.5 cm). The number of involved axillary lymph nodes and the histological grade have both an impact on the dissemination probability and the size of the V50. It is also possible to estimate the median volume of the primary at the involvement of the first axillary node, of the second, of the third... On average, the V50 for distant metastatic dissemination is greater than the volume of the tumor at the initiation of the second axillary node but smaller than the volume at the initiation of the third. Another study was the assessment of the time at which the metastases were initiated. The results concerning age of metastases at the time of initial treatment allow to estimate the reduction in the incidence of distant metastases which can be obtained by an earlier diagnosis; these estimations are consistent with data of screening programs. The size of the occult metastases at the time of treatment of primary tumor was calculated. It was found that this size is much larger in the subset of patients with pejorative prognostic factors which probably explains the relatively low effectiveness of adjuvant chemotherapy in these patients. The maximum size of the metastases which are controlled by adjuvant chemotherapy was found equal to 1.5.10(6) cells. Finally the model was used to investigate the possible dissemination arising from loco-regional recurrence. The discrepancy between the great reduction in the incidence of loco-regional recurrences obtained with post-operative radiotherapy and its relatively small impact on survival is probably due to the fact that a recurrence has an impact on survival only in patients without distant metastases at the time of initial treatment and in whom the recurrence was not detected and treated before it reached the size at which it can initiate a metastasis. This situation is likely to occur only in patients with tumors of the inner quadrants in whom the internal mammary chain has a great likelihood of being involved.