The Impact of Thyroid Nodule Size on the Risk of Malignancy and Accuracy of Fine-Needle Aspiration: A 10-Year Study from a Single Institution

Abstract

False-negative rates for thyroid fine-needle aspiration (FNA) vary from 0.4% to 13%, but the effect of nodule size on the accuracy of thyroid FNA remains controversial. We hypothesized that large thyroid nodule size does not contribute to the risk of malignancy or the risk of a false-negative FNA. All thyroid FNAs performed at the Walter Reed Army Medical Center during September 2001-August 2011 were reviewed. A strict correlation between the biopsy site, location, and size of nodule on ultrasound (US) and pathology report was ensured. FNA results were classified as benign, atypical, follicular neoplasm/suspicious for follicular neoplasm (FN/SFN), suspicious for malignancy (SM), or malignant, and the pathology result was categorized as either benign or malignant. Nodules were analyzed by size: 0.5-0.9 cm (group A), 1.0-3.9 cm (group B), and ≥ 4 cm (group C). Incidental thyroid cancer was not included. Of 3013 patients undergoing FNA, 667 (22.1%) had surgery. Patients were excluded for nodules <0.5 cm, nondiagnostic FNA, or no preoperative US, leaving 540 patients with 695 nodules. Among patients referred for surgery, FNA results were benign in 417 nodules (60%), atypical in 22 (3.2%), FN/SFN in 122 (17.6%), SM in 77 (11.1%), and malignant in 57 (8.2%). Postoperative malignancy rates by FNA result were 7% if benign, 4.5% if atypical, 23% if FN/SFN, 33.8% if SM, and 78.9% if malignant. FNA accuracy was 60% in group A, 68.5% in group B, and 80.3% in group C (p=0.01). False-negative rates for FNA were 7.0% overall, 15.8% in group A, 6.3% in group B, and 7.1% in group C (p=0.25). Sensitivity and negative predictive value were highest in group B at 81.6% and 93.7%, respectively. The prevalence of malignancy was not different between groups. Our results show that the thyroid nodule size ≥ 4 cm increases neither the risk of false-negative FNA results, nor the overall risk of malignancy. We also show a tendency toward a higher false-negative rate in subcentimeter nodules. We conclude that a large nodule size should not prompt automatic referral for thyroidectomy. An increased awareness of potential sampling error in subcentimeter nodules is warranted.