How significant is detection of ductal carcinoma in situ in a breast screening programme?

To compare the histological grades of screen detected and non-screen detected ductal carcinoma in situ (DCIS) and to identify any differences that might support the contention that DCIS found by breast screening represents an over-diagnosis. The aim was also to establish whether any particular mammographic features of DCIS can be used to predict tumour grade reliably.

Biopsy proven cases of DCIS (n=153) were reviewed with respect to grade and subdivided into high, intermediate and low grades using the Van Nuys classification. A more aggressive subset of DCIS (microinvasive and interval cancers) were similarly analysed. Mammograms were reviewed with regard to abnormal features and distribution, and the appearances correlated with grade.

Fifty-four percent (53/98) of screen detected and 62% (34/52) of non-screen detected DCIS were high grade. The rest were equally intermediate and low grade, with no statistical difference between the two groups. Eighty-four percent of the aggressive subset of tumours were high grade. Micro-calcification was present in 90% and in 10% there were soft tissue changes alone. Seventy-six percent of linear branching calcification was associated with high grade DCIS. Only 13% of high grade DCIS demonstrated punctate micro-calcification; however, 38% of cases of punctate micro-calfication were associated with high grade tumours and there was a great deal of overlap between the groups.

Most cases of DCIS in both screen and non-screen detected groups were high grade. Only one in five was low grade. Analysis of the aggressive subgroup underlines the significance of high grade DCIS. Mammographic patterns are not always reliable in the prediction of tumour grade. The detection of DCIS in screening programmes is important and should not be regarded as over-diagnosis.