Abstract
Although screening substantially reduces breast cancer mortality, 1 the mammographic screening test to detect asymptomatic breast cancer is an imperfect means of early detection. Not every breast cancer is detected earlier and not every death is prevented, even among women who dedicatedly participate in screening. This imperfection of mammographic screening is partly caused by its moderate test sensitivity. In the Netherlands, the current sensitivity of mammographic screening is around 70%. 2 One main reason for this, to a certain extent disappointing sensitivity, is the reduced ability of mammography to detect invasive lobular carcinoma (ILC). ILC tends to grow more diffusely than invasive ductal carcinoma (IDC). 3 This behaviour hampers the ability of mammography to detect ILC at an early stage. It has been reported that the false-negative mammography rates are up to 19% more detrimental in ILC compared to IDC. 4
IDC accounts for about 80% of all cases, and ILC for 10–15% of all breast cancers. Screening of the general population reduces breast cancer mortality by about 50%; 1 we were specifically interested in comparing the extent of mortality reduction in patients diagnosed with IDC and ILC. Therefore, we conducted a case – referent study within the population-based screening programme in Nijmegen, the Netherlands. A detailed description of this study has been published. 1
There were 70 ILC and 411 IDC cases in our study. The distribution of ILC and IDC cases did not change over time between 1975 and 2008.
The odds ratio (OR) for the effect of screening on ILC mortality was 1.00 (95% confidence interval [CI] = 0.56–1.78; Table 1). The OR calculated for the IDC group was 0.63 (95% CI = 0.50–0.80).
The effect of screening on ILC and IDC mortality expressed by odds ratio, after adjustment for age
These results show that there is no effect of screening on reduction of mortality in ILC patients; this compared with a significant reduction measured in IDC. If detected at an early stage, ILC has a similar 5 or possibly an even better prognosis than IDC. 6 This may be due to a more favourable biological profile of ILC; these tumours are more likely to contain oestrogen receptor and progesterone receptor expressions, and normal expressions of Her2/Neu and p53. 7 Previous results from the Swedish Two-County Trial have indicated that there is a potential for a mortality reduction from ILC tumours by screening. 8 It is thus a challenge for service screening programmes to detect ILC at an early stage.
On a mammogram, and in particular on the mediolateral oblique views, ILC is usually difficult to identify. 3 In subsequent screening examinations in the Dutch screening programme a mediolateral oblique view is standard, whereas a craniocaudal view is only performed on indication. In most cases, however, signs of ILC are more visible on craniocaudal views. 3
Digital mammography has recently been implemented in the Netherlands, making computer-aided detection relatively easier to apply. In programmes that have applied this technique for some time, a promising mammographic sensitivity of more than 90% for detecting both ILC and IDC at an early stage has been reported. 9 The application of computer aided detection in the Dutch service screening programme will probably increase referral rates, 10 however the increase in the number of false-positive referrals will probably not outweigh the benefit of the increase in the number of true-positive referrals. 11 We expect that routine use of two-view digital mammography and computer-aided detection systems will contribute to further reduction in breast cancer mortality.