Ductal carcinoma in situ (DCIS): are we overdetecting it?

DCIS constitutes approximately one-quarter of screen-detected cancer. The proportion of screen-detected cancer that is DCIS decreases with the age of the screening population. Whether detecting DCIS at mammographic screening is of benefit depends on the natural history of the DCIS detected at screening and whether DCIS detection enables the detection of invasive foci already present within DCIS lesions.

The natural history of DCIS varies according to the grade of the DCIS detected. The natural history of low grade DCIS is that approximately 60% of lesions will become invasive at 40 years follow-up. The natural history of high grade DCIS derived from local recurrence rates within high grade DCIS lesions, which have been inadequately resected and not given radiotherapy, suggests an invasive risk of at least 50% at 7 years follow-up. Given the average life expectancy of women with screen-detected DCIS is 25 years, this suggests high grade DCIS is an obligate precursor of invasive disease. There is a strong correlation between the grading of invasive cancer and the grade of DCIS from which it arose. This suggests that approximately 50% of low grade DCIS detected at screening will represent overdiagnosis while overdiagnosis of high grade DCIS would be rare. The natural history of intermediate grade DCIS is as yet unknown. Only 15% of screen-detected DCIS is low grade, suggesting that overdiagnosis is uncommon.

Approximately 60% of screen-detected DCIS is high grade and in the vast majority of these patients adequate treatment will be preventing the occurrence of high grade invasive breast cancer. Approximately one-third of malignant calcification clusters detected at screening mammography already has an invasive focus. These invasive foci are usually small and high grade. Data gathered from all the units in the National Health Service Breast Screening Programme indicate a strong correlation between the detection of DCIS and the detection of small invasive cancers. These data indicate that for every two cases of DCIS detected, one case of a small invasive cancer is also detected. These data indicate that DCIS detection within a mammographic screening programme prevents the occurrence of high grade invasive cancer and aides the detection of small invasive cancers at a size where treatment can be successful. Overdiagnosis will occur in approximately one-half of the cases of low grade DCIS detected. This represents approximately 2% of screen-detected lesions. If one assumes the natural history of intermediate grade DCIS is intermediate between low and high grade disease then overdiagnosis of DCIS within a screening programme may represent 3% of screen-detected lesions.