Does breast cancer grade worsen with time? Evidence from breast screening

Abstract

It has been suggested that alteration of the distribution of histological grades, that has been found in screening programmes, is evidence for progression in histological grade with increasing size. A predictive model, that was based upon estimated growth data from an unselected series of 98 new primary breast cancers, is used to estimate the proportions of tumors that would be diagnosed according to limiting screen diagnostic and clinical diagnostic sizes after particular screening intervals; windows of opportunity are created. The results show that the limitations imposed by time and size criteria alter the distribution of growth rates of tumours that may appear in the windows. Small screen diagnostic sizes and short screening intervals allow only the most rapidly growing tumours to reach large sizes. This produces an apparent association of grade 3 tumours with large size. Interval cancers are also likely to be more rapidly growing while the more slowly growing tumours will be diagnosed at the subsequent screen to produce a spurious association of slowly growing grade 1 tumours with small size. We conclude that the evidence from screening does not support the thesis of progression of histological grade with the ageing of the tumour, since the changes that have been observed are predictable from a simple model based upon patterns of tumour growth rates and the relationships between growth rate and histology.