We agree with Leinung and colleagues that thyroid ultrasonography detects many nodules not palpated on physical examination. However, we disagree that the routine use of thyroid ultrasonography in patients with suspected nodular thyroid disease leads to unnecessary ultrasonography and surgeries.

We aspirated nodules that were 1 cm or larger. Previous studies have demonstrated that on physical examination, experienced clinicians often miss nodules that are larger than 1 cm on ultrasonography (1). There is no logic to the contention that a nonpalpable nodule larger than 1 cm is less likely to be malignant than a palpable 1-cm nodule. Our study, although small, confirms this. In addition, the incidence of malignancy in our study (6.7% in patients with solitary nodules and 8.9% in patients with multiple nodules) is in accord with incidence rates found in larger series and indicates that our approach does not detect clinically insignificant thyroid cancer (2).

Leinung and colleagues' statement that surgery should not be performed in patients who are expected to have only a 50% risk for carcinoma is at variance with most opinions in this field. The problem of indeterminate cytologic results is not specific to our approach, and the percentage of patients with indeterminate cytologic characteristics who required surgery is similar to that found in larger series (2). We are not aware of long-term studies of patients with suspicious results on fine-needle aspiration who were followed without surgery.

An assessment of the cost–benefit ratio for routine use of ultrasonography for fine-needle aspiration is complex. The Thyroid Nodule Clinic has eliminated multiple office visits and streamlined patient care by allowing both endocrine consultation and definitive evaluation of suspected nodular thyroid disease in a single visit. Ultrasound guidance ensures that the needle is in the nodule when the sample is obtained and allows direct sampling of the solid portions in cystic nodules, often preventing multiple aspirations. In addition, the rapidity of the diagnosis obviates the need for additional testing, other than a screening test for thyroid-stimulating hormone level before referral. This eliminates unnecessary testing by primary care providers evaluating suspected nodular disease (3).

We appreciate Dr. Gross's comments about the use of characteristics of thyroid nodule ultrasonography for predicting malignancy. Although fine-needle aspiration is the best test for diagnosis of malignancy in a thyroid nodule, several reports have suggested an association between intranodular microcalcifications and thyroid cancer. In addition, irregular nodule margins and hypoechogenicity have been reported to predict malignancy (4). The specificity and sensitivity of these findings vary, however, and the clinical utility of these ultrasonography characteristics remains unclear. We did not include ultrasonography characteristics in our analysis.