Thyroid nodule size calculated using ultrasound and gross pathology as predictors of cancer: A 23-year retrospective study.

Background Thyroid nodules are very common. Ultrasound (US) and fine needle aspiration (FNA) are both integral in cancer screening. This study investigated the concordance between thyroid nodule sizes measured by US and gross pathologic examination and their relationship with malignancy. Methods A retrospective design was used to select consecutive patients with proven carcinoma of the thyroid. The number and maximum diameter of nodules, rates and types of malignancy, as well as predictors of malignancy were determined. Results The 10 944 patients examined had 15 283 thyroid nodules, 44.6% of which were malignant. Of the 4449 nodules sampled by FNA and the 8748 not sampled by FNA, 76.5% and 30.5% were malignant, respectively. The sensitivity, specificity, positive and negative predictive values (NPVs), and overall accuracy of FNA based on final pathology were 97.9%, 96.3%, 98.8%, 93.5%, and 97.5%, respectively. Nodule sizes determined by US were comparable with most nodules having either the same size range (n = 2959, 77.7%) or differing only by one size range (n = 770, 20.5%). Conclusions Thyroid nodule size is inversely related to malignancy risk, as larger nodules have lower malignancy rates. Nodule size estimated by US shows relatively good correlation with final pathologic size. However, thyroid nodules should undergo FNA regardless of size. If the FNA is not benign, nodule size should influence therapeutic decision making.