In this case-control study we describe how often thyroid cancers and occult cancers are diagnosed or not diagnosed by fine-needle aspiration (FNA) in patients with thyroid nodules and a family history of nonmedullary thyroid cancers (FNMTC). Our hypothesis is that patients with thyroid nodules and a family history of FNMTC seem to be similar to patients with thyroid nodules and a history of exposure to low-dose therapeutic radiation. Both have been reported to have multifocal thyroid neoplasms and malignant tumors are common. Cytological examination may therefore be less accurate. From 1979 to 1996, 27 patients from 24 families with FNMTC were examined histologically after a preoperative cytological examination in all of them. A positive cytology examination was defined when biopsy documented thyroid cancer. It was interpreted as a false-negative study when a benign diagnosis was made and thyroid cancer was present anywhere within the thyroid, including in areas sampled or not sampled by FNA and not palpable preoperatively. A randomized control group, matched for age and gender, contained 27 patients with papillary thyroid cancer without familial disease. In our study group, 25 patients were treated with total thyroidectomy, including 7 with neck dissection, and 2 by thyroid lobectomy. At final histological examination 17 of 27 patients (63%) in this study group had multiple nodules and 25 of 27 (92.6%) had thyroid cancer. Thyroid cancer was diagnosed by FNA in 22 of 25 patients (88%), with 3 (12%) false-negative biopsies due to sampling errors (thyroid cancer not in the index nodule), versus 1 (3.7%) false-negative biopsy in the control group. Two patients in the study group with benign nodules were accurately diagnosed. In patients with false-negative biopsies and a history of FNMTC, the cancer was situated in one or more small nodules. Only one cancer was occult (< 1.0 cm). One-third of the patients in our study group (33%) had a history of radiation; 44% of the irradiated group had a single nodule; 56% had multiple nodules. In the control group, 9 of 27 patients (33%) also had a history of radiation; 33% of the irradiated group had a single nodule, 67% had multiple nodules. In conclusion, the reliability of FNA in patients with FNMTC appears to be less accurate than it is for other patients because of the high incidence of multifocal thyroid cancer and coexistence of benign nodules. Patients with thyroid nodules and a family history of thyroid cancer are more likely to have thyroid cancer and because they also have more coexistent benign nodules, they must be followed closely or treated with total or near-total thyroidectomy.