Combined morphology and distribution features for risk stratification in pathology-enriched BI-RADS 4A/4B microcalcifications

Background Breast microcalcifications are a common indication for biopsy in breast cancer screening and often represent the earliest mammographic manifestation of malignancy. However, substantial heterogeneity within BI-RADS 4 A lesions presents challenges for clinical interpretation. This study aimed to evaluate the association of microcalcification morphology and distribution with pathological outcomes and to explore their potential role in characterizing heterogeneity within BI-RADS 4 A lesions. Methods A total of 327 patients with 356 pathologically confirmed microcalcification-positive breast lesions detected on digital mammography were retrospectively analyzed on a lesion-based basis. Calcification morphology and distribution were classified according to the BI-RADS 5th Edition. Associations with pathological outcomes were assessed using chi-square or Fisher’s exact tests. Multivariable logistic regression was performed to identify independent factors associated with malignancy. Receiver operating characteristic (ROC) curve analysis was used to evaluate diagnostic performance at BI-RADS thresholds ≥4 A and ≥4B. Results Among the 327 patients, 205 (62.7%) had malignant disease and 122 (37.3%) had benign disease. Fine pleomorphic (80.2%), amorphous (66.2%), and fine linear or branching (100%) morphologies, as well as clustered (71.2%), regional (96.6%), and linear (100%) distribution patterns, were significantly associated with malignancy ( P < 0.001). Multivariable analysis demonstrated that morphology and distribution remained independently associated with malignancy. Among lesions initially categorized as BI-RADS 4 A in institutional practice ( n = 58), substantial heterogeneity was observed. Lesions with high-risk morphology and distribution patterns were more frequently malignant, whereas punctate calcifications with scattered distribution showed a lower malignancy proportion. In this pathology-confirmed cohort, BI-RADS ≥4B showed higher specificity than BI-RADS ≥4 A, although this category-level finding should be interpreted descriptively. Conclusions In this retrospective cohort of biopsied BI-RADS 4 A/4B microcalcification-positive lesions, combined morphology and distribution features provided supplementary information for imaging–pathology correlation and helped illustrate risk heterogeneity within institutionally assigned BI-RADS 4 A lesions. These findings should be interpreted as confirmatory observations within a selected biopsy cohort rather than population-level malignancy estimates, and they are not intended to replace established BI-RADS-based management principles. Clinical implications These findings should be interpreted with caution due to the pathology-enriched design and should not be considered population-level risk estimates. They may provide supplementary information for understanding lesion heterogeneity but are not intended to replace established BI-RADS-based clinical decision-making.