To determine, by means of long-term follow-up evaluation, the outcome and accuracy of stereotactic core-needle biopsy (SCNB) of non-mass calcifications observed at mammography, and to analyze the factors contributing to false-negative findings.
Materials and Methods
Using a 14-gauge needle, SCNB was performed in cases involving 271 non-mass calcified lesions observed at mammography in 267 patients aged 23-72 (mean, 47) years. We compared the SCNB results with those of long-term follow-up which included surgery, mammography performed for at least six months, and reference to Korean Cancer Registry listings. We investigated the retrieval rate for calcifications observed at specimen mammography and histologic evaluation, and determined the incidence rate of cancer, sensitivity, and the underestimation rate for SCNB. False-negative cases were evaluated in terms of their mammographic findings, the effect of the operators' experience, and the retrieval rate for calcifications.
For specimen mammography and histologic evaluation of SCNB, the retrieval rate for calcifications was, respectively, 84% and 77%. At SCNB, 54 of 271 lesions (19.9%) were malignant [carcinoma in situ, 45/54 (83%)], 16 were borderline, and 201 were benign. SCNB showed that the incidence of cancer was 5.0% (6/120) in the benign mammographic category and 31.8% (48/151) in the malignant category. The findings revealed by immediate surgery and by long-term follow-up showed, respectively, that the sensitivity of SCNB was 90% and 82%. For borderline lesions, the underestimation rate was 10%. For false-negative cases, which were more frequent among the first ten cases we studied (p = 0.01), the most frequent mammographic finding was clustered amorphous calcifications. For true-negative and false-negative cases, the retrieval rate for calcifications was similar at specimen mammography (83% and 67%, respectively; p = 0.14) and histologic evaluation (79% and 75%, respectively; p = 0.47).
In this study group, most diagnosed cancers were in-situ lesions, and long-term follow-up showed that the sensitivity of SCNB was 82%. False-negative findings were frequent during the operators' learning period.