Pathology-related medical malpractice claims frequently concern fine-needle aspirations (FNAs) of breast lesions, and diagnostic errors have been attributed in part to the inadequacy of the specimens. Cytologic criteria for adequate FNA specimens, specifically in cases without malignancy, have not been clearly defined. From January 1988, to August 1995, 669 ultrasonographic-guided FNAs of nonpalpable, solid breast lesions with subsequent histologic examination were performed at our institution. From these, 54 cases with cytologic diagnoses of insufficient or nonspecific benign findings were identified. All aspirates were reviewed, and the number and size of the epithelial cell groups were quantitated in each case. By using criteria for adequate aspirates of palpable breast lesions (six or more epithelial cell groups per case with a minimum of 5-10 cells per group), 23 of the 54 aspirates were deemed inadequate and 31 adequate. Eleven (48%) of the inadequate aspirates and 17 (55%) of the adequate aspirates were from histologically confirmed carcinomas (ductal carcinoma in situ, 6; invasive carcinoma, 22, of which 12 were ductal, 7, lobular, and 3, mixed ductal and lobular). For the mammographic diagnoses "probably benign," "indeterminate," and "suggestive of malignancy or malignant," the probability of malignancy in aspirates of adequate cellularity (eg, > 6 epithelial groups) was 9%, 40%, and 93%, respectively. These findings indicate that a significant proportion of breast aspirates still may yield false-negative results despite adequate to high cellularity. Although a definition of adequacy based on cellularity is useful in reducing false-negative results, cellularity alone cannot be relied on in the management of nonpalpable lesions. For mammographic findings that are indeterminate or suggestive of malignancy or malignant, nonspecific FNA findings should be followed by core or excisional biopsy to exclude carcinoma.