Discussion:

In our study 28.8% of patients were in the higher density group (BI-RADS 3 or 4), comparable to recent data from an urban screening centre in the UK which found that 32% of women had heterogeneously dense breasts [2].

A meta-analysis of MRI use in ILC found that in 32% of patients there were additional ipsilateral lesions and contralateral malignancy in 7% [5].  Our findings are similar with additional ipsilateral lesions in 28.8% and contralateral malignancy in 4.1%. 

Additional disease was seen in all 4 breast density groups.  Although a higher proportion of the dense breasts showed additional disease compared to the less dense breasts this was not statistically significant.  However a study by Wong et al found that in multivariable analyses a factor significantly associated with additional suspicious findings on MRI in ILC was higher breast density [10]. 

We had a slightly lower percentage of change in surgical management following MRI of 21.9% compared to literature figures of 25 to 28.3% of cases [5-7].

Our re-excision rate to achieve clear margins of 16.4% was similar between low and high density groups and to published work (18.3%) [7].

Primary mastectomy rates following pre-operative MRI are 31-32% in ILC [7,11].  Our higher primary mastectomy rate of 41.1% is attributed to several patients with previous radiotherapy and further patients who chose mastectomy over conservation.

Many patients required a second look US and in 60% this documented a benign lesion highlighting the issue of low specificity of MRI.

Conclusion:

Additional disease was identified across all breast density categories and we thus feel MRI has a role in all pre-operative ILC patients though is most useful in patients for BCS.