Breast cancers invisible on mammography.

Background: A proportion of the cancers 'missed' at mammography are invisible even with the benefit of hindsight. The aim of the present study was to identify a group of women with proven breast cancer whose mammograms did not show a suspicious lesion even in retrospect (i.e. the truly mammographically occult cancers), and to compare them with mammogram-positive cancers.

Methods: A total of 1757 breast cancers was diagnosed at the Wesley Breast Clinic's Screening or Diagnostic Services between July 1987 and August 1997. One hundred and twenty cases were identified where, after independent review by two of the authors, no mammographic abnormality could be found in the region where the cancer was subsequently found. These 120 cases were compared with 1548 cancers considered to have a lesion visible on mammography, whether benign, indeterminate, suspicious or malignant in appearance.

Results: In 90% of the mammogram-negative cancers, a clinical abnormality led to further investigation, while the remainder were found incidentally on ultrasound. There were a higher proportion of dense breasts, and of women aged 40-49, in the mammogram-negative cancers than in the mammogram-positive cancers. The mammogram-negative cancers were of smaller size overall, but three of them were surprisingly large (7-11 cm). In both the mammogram-positive and -negative cancers approximately 60% were ductal invasive cancers.

Conclusions: Where factors are present that make mammographically occult malignancy more likely (e.g. age 40-49 and dense breasts), women may be targeted for further investigation by other modalities. This is essential in the presence of a clinical abnormality.