ObjectiveTo establish a nomogram to predict the risk of malignancy of suspicious calcifications in the breast for guiding clinical decision making.
MethodsA total of 178 patients with breast microcalcifications detected by mammography and localized with image guidance was retrospectively analyzed.All calcifications were classified by a breast radiologist according to the fifth edition of BI-RADS.Univariate and multivariate logistic regression were used to analyze the relationship between the imaging features of calcifications and the clinical features of patients.Based on multivariate logistic regression analysis, a nomogram was developed to predict the risk of malignancy, and the corresponding diagnostic efficiency of the model was evaluated using ROC curves.
ResultsAmong the 178 patients with microcalcification, 114 were suspected of calcification.In the suspected calcification group, there were statistically significant differences in malignancy rates among the < 50 years old and ≥50 years old groups(P < 0.01), amorphous/coarse heterogeneous calcification and fine pleomorphic/fine linear or subdivided dendritic calcification, regional/clustered calcification and linear/segmental calcification in the distribution of calcification(P < 0.01).Age≥50 years old, calcification type of fine pleomorphism/fine line or fine branch, and calcification distribution of line segment were independent risk factors of malignant lesions.Based on these three factors, a predictive nomogram was established, and the corresponding area under the curve was 0.747 (95%CI: 0.658-0.836).
ConclusionsIn early diagnosis of breast cancer, image-guided localized biopsy of breast microcalcifications has certain value.Morphological descriptors of suspicious microcalcifications are required for subclassification, and a nomogram combining morphology and distribution descriptors of suspicious microcalcifications can accurately predict the risk of malignancy.
DownLoad: