Preoperative mammographic features predict clinicopathological risk factors for the development of local recurrence in breast cancer.

This study attempts to correlate preoperative mammographic features with established clinicopathological risk factors for local recurrence with a view to selecting patients who might be unsuitable for breast-conserving surgery. A total of 357 patients underwent surgery for invasive breast cancer. All patients' preoperative mammograms were reviewed by a single radiologist (LW) blinded to clinicopathological features. The independent sample t-test and chi-squared tests were used to correlate preoperative mammographic features with established clinicopathological risk factors for the development of local recurrence. The presence of extensive in situ disease within the tumour was associated with both casting calcification and absence of a mammographic nidus (P=0.0001 and 0.002 respectively) but not with mammographic density. Furthermore casting calcification (P=0.003), absence of a mammographic nidus (P=0.027) and maximum nidus size (P=0.014) were all predictors of tumour bed positivity. This study has demonstrated that preoperative mammograms can be used to identify those patients at risk of occult disease and help select those patients that may be unsuitable for breast-conserving surgery.