RESUMO
OBJECTIVE: To determine if there is a significant difference in the malignancy rates of thyroid follicular neoplasms, based on lesion size. DESIGN: Retrospective study. SETTING: Tertiary care institution. MATERIALS AND METHODS: We retrospectively reviewed charts from patients with a thyroid fine-needle biopsy diagnosis of "follicular neoplasm," who subsequently had thyroidectomy, from 2000 to 2005 at our institution. The size of the lesion was determined from pathology reports or preoperative sonograms. We divided the patients into "large nodule" and "small nodule" groups, based on the median size. MAIN OUTCOME MEASURE: Malignancy rate difference between thyroid follicular neoplasms in large nodule and small nodule size groups. RESULTS: One hundred seventeen patients were included. The median size was 2.2 cm; thus, our small nodule group included lesions < or = 2.2 cm with a malignancy rate of 16.7%. The large nodule group included lesions > 2.2 cm with a malignancy rate of 26.3%. Chi-square analysis showed no statistically significant difference between the groups (p = .20). A receiver operating characteristic (ROC) curve was generated, and all cutoff values of tumour size did not lie far from the line of no discrimination. The area under the ROC curve was 0.52 (95% confidence interval = 0.38-0.66), indicating that the use of increasing tumour size as a diagnostic tool may be worse than chance. CONCLUSION: There was no statistical difference in the malignancy rates of follicular neoplasms observed in the small nodule and large nodule groups. Our ROC curve also indicated that tumour size cannot be used to predict malignancy.