Abstract

Tumour imprints were made from 311 breast tumours at the time of frozen section. Rapid and permanent staining methods were employed and the accuracy of diagnosis was compared with that obtained from frozen and paraffin sections. Two false positives (0·65%) and 16 false negatives (5·15%) were found in the imprint group compared with no false positives and five false negatives (1·6%) in the frozen section group.

The cytology of breast imprints is described and the false results are analysed. The possible uses of this technique, both alone and in combination with frozen sections, are discussed.