Breast cancer metastatic development is commonly considered as resulting from continuous tumour growth from tumour seeding until clinical recurrence is documented. Continuous growth model inferences, however, fail to explain clinical findings concerning local recurrences, as well as the time-distribution of first treatment failure and mortality for patients undergoing mastectomy. The tumour dormancy hypothesis is considered to provide a more reasonable description of the natural history of breast cancer, while primary tumour removal is believed to be a potential perturbing factor for metastasis development. A new model of the natural history of operable breast cancer, incorporating tumour dormancy and starting signals from surgery for micrometastatic growth is proposed.