Objective: To evaluate histological needle biopsy in breast lesions classified radiologically as 'non-benign'.
Design: Prospective, descriptive.
Setting: Hospital Velp, Velp, the Netherlands.
Methods: 232 women with an 'uncertain', 'suspicious' or 'malignant' result of mammography, if necessary supplemented by echography, were subjected to histological biopsy from a breast between 1 January 1994 and 1 January 1997. The earlier biopsies were made with a 16 Gauge (G) needle, those after April 1996 with a 14 G needle, as a rule under stereotactic control. In principle, operation was performed after a positive result. Concerning the women operated after the biopsy, the results of the histological examinations were compared in a 2 x 2 table.
Results: 165 of the 232 patients (71%) had breast cancer. Of the 59 patients classified roentgenologically as 'uncertain', 15 (25%) had breast cancer, of the 'suspicious' cases this ratio was 44/67 (66%) and of the 'malignant' results it was 106/106 (100%). Operation was performed in 186 women. The biopsy findings and the surgical preparation were in agreement in 169 patients (91%). The sensitivity of the stereotactic biopsy was 90%, its specificity 93%. One woman was over-treated (axillary lymph node resection) because of a biopsy classified as malignant performed on a rare tumour ultimately diagnosed as 'adenomyo-epithelioma with epithelial atypia'. The proportion of false-negative results was 36%. The predictive value of a positive result was 99%, that of a negative result 63%.
Conclusion: A diagnostic stereotactic biopsy after roentgenological classification based on mammography and echography had a good predictive value regarding the probability of breast cancer.