Accuracy of touch imprint cytology of image-directed breast core needle biopsies

OBJECTIVE: To evaluate the accuracy of touch imprint (TI) cytology of image -directed core needle biopsy (CNB) specimens of nonpalpable breast lesions. STUDY DESIGN: Fifty-two consecutive CNBs from 44 patients were performed un der mammographic or ultrasound guidance. Air-dried TIs of CNBs were stained with Diff-Quik. TI cellularity was considered adequate if six or more duct al cell groups were identified. CNBs and TIs were interpreted in a blinded fashion. RESULTS: TI cellularity was adequate in 37/52 (71%) cases, including 17/20 carcinomas and 20/32 benign lesions. Among 17 carcinomas, TIs were positive in 12, suspicious in 4 and atypical in 1. One case of lactational change w as suspicious on TI, and 5/8 fibroadenomas were atypical. No benign lesions were called "carcinoma" on cytology. When lesions categorized its "carcino ma" or "suspicious" were considered positive and those classified as "atypi cal" or "benign" were scored as negative, TI sensitivity and specificity we re 94% and 95%, respectively. When the "atypical" cases were grouped with t he positive cases, TI sensitivity was 100%, with 75% specificity.; CONCLUSION With satisfactory cellularity, TIs of CNBs are highly accurate. When immediate evaluation of CNB specimens is important, TIs can potentiall y decrease the number of biopsy passes required and provide preliminary dia gnoses.