Volume 11 Supplement 2
Can we predict the likelihood of malignancy in mammographically indeterminate microcalcification?
© Khan et al; licensee BioMed Central Ltd. 2009
Published: 26 October 2009
The aim of this study was to identify features that will enable a radiologist to predict the likelihood of malignancy in mammographically indeterminate microcalcification (M3) before biopsy.
Retrospective review of patients with M3 microcalcification identified from our database. Patient age, cluster size, and site, and whether they had a new or increase in size of microcalcification was studied as features that can help in differentiating malignant from benign microcalcification. Data were analysed using SPSS using KruskalWallis and chi-square test.
The total number of patients with M3 microcalcification was 242. On final histology, 99 (41%) patients had breast cancer and 143 (59%) had benign breast disease. In the cancer group, the mean age was 57 years (95% confidence interval (CI) 55.7 to 58.4 years), the mean cluster size was 18.62 mm (95% CI 14.86 to 22.38), and 84% of patients had a new microcalcification and16% had an increase in size of microcalcification. Of the malignant M3 microcalcifications, 65% were in the outer quardrant, 26% were in the medial quardrant and 9% were central. In the benign group, the mean age was 54.5 years (95% CI 53.9 to 55.8), the mean size of microcalcification was 18.62 mm (95% CI 14.86 to 22.38), 76% had a new microcalcification and 22% had an increase in size of microcalcification. Of the benign M3 microcalcifications, 51% were in the outer quadrant, 35% were in the medial quadrant and 14% were central. Except for age, there was no statistically significant difference between the studied variables in the two groups. On the KruskalWallis test, patients with cancer were older compared to patients with benign disease (P = 0.01).
Increasing age is the only strong predictor of malignancy in M3 microcalcification. Size, site, new microcalcification and increasing microcalcification size were not related with the probability of cancer.
This article is published under license to BioMed Central Ltd.