A retrospective review of thyroidectomies to rule out cancer at Columbia Presbyterian Medical Center indicated that the maximum cancer incidence achievable in surgical specimens was only 57% if all microfollicular adenomas and thyroid cancers were removed. The size distribution of follicular cancer and microfollicular adenoma was similar and it does not appear to be useful in deciding which microfollicular lesions need to be surgically excised to exclude carcinoma. The incidence of thyroid cancer in patients undergoing thyroidectomy was not influenced by the introduction of fine-needle biopsy (FNB) at Columbia Presbyterian Medical Center. FNB did appear to increase the incidence of cancer at an affiliated community hospital. Coarse-needle biopsy definitely improved the incidence of cancer in patients undergoing thyroidectomy and was superior to FNB and clinical selection alone. A combination of FNB and coarse-needle biopsy when possible yielded a 41% incidence of cancer and appears to be the most useful technique for patient selection.