Probabilistic approach to the reporting of fine-needle aspiration cytology of the breast

The use of fine-needle aspiration cytology (FNAC) in the clinical management of breast disease has grown steadily in the U.S. As proficiency and experience with this technique have increased, the need for the standardization of terminology and approach is clear. Simple and easily applied diagnostic categories are needed to facilitate communication with clinical colleagues and other pathologists and to allow outcomes assessment, both intraand interinstitutionally. Most important is that cytopathologists understand the impact of diagnostic terms on their clinicians. When greater certainty than is warranted can lead to an irretrievable error (i.e., mastectomy or chemotherapy), mutual understanding of diagnostic certainty is mandatory. Although the terms may vary, a consistent system of presentation is essential. Progress in the clinical and pathologic diagnosis of breast tumors and their management demands a more complex and sophisticated approach than one presupposing that most specimens may be accorded a cancer yes/cancer no, dichotomous diagnosis. The philosophy described herein follows that of the 1996 Breast FNA Conference held in Bethesda, Maryland, which was supported by the National Cancer Institute and many other professional associations. At that conference, the subcommittee on terminology and other attendees entered into lengthy and animated discussion of the number and names of proposed categories to be applied to the interpretation of See editorial on pages 1–2, this issue. breast aspirates. This presentation reflects the importance of this basic framework, and endeavors to explain the probabilistic versus diAddresss for correspondence: David L. Page, M.D., Department of Pathology, Vanderbilt Unichotomous approach to diagnostic reporting in this important area. versity, C3321 Medical Center North, 1161 21st There are two major competing philosophic bases to the reporting Avenue South, Nashville, TN 37232. Reprints of FNAC in the breast. The first is predicated on the assumption that not available. this is a laboratory test that can be evaluated with regard to specificity, sensitivity, predictive values, etc. Basically, this is done in a 2 1 2 Received December 4, 1996; revision received table format in which each sample is assigned to one of the four cells December 11, 1996; accepted December 13, 1996. labeled as follows: true-positive, true-negative, false-positive, and