A Practical Score Model for Detecting Malignant Thyroid Nodules by Using Ultrasonographic Findings



Sodagari N¹ ; Jafari M¹ ; Nasiri S¹ ; Molavi B¹ ; Notash AY¹ ; Mir A¹ ; Hedayat A¹ ; Shojaiefard A¹ ; Abdehgah AG¹ ; Raouf MT² ; Mosadegh S³ ; Soroush A¹ Show Affiliations
Source: Acta Medica Iranica Published:2018

Abstract

About 10% of thyroid nodule Fine Needle Aspirations (FNAs) are reported indeterminate, and this number is much higher in Iran. All of these patients undergo thyroid surgery, whereas, only about 20% of them are malignant and they did not need surgery. Therefore, we decided to evaluate the ultrasonographic features of malignant thyroid nodules to find a predictive scoring model for thyroid nodules and consequently reduce the unnecessary thyroid surgeries. This prospective cross-sectional study was conducted on 114 patients with thyroid nodule who were candidate for thyroid surgery. All the patients were assessed by ultrasonography of thyroid before surgery, and after surgery by the gold standard of permanent pathology, the sensitivity, specificity, positive predictive value, and negative predictive value of each parameter of ultrasonography were determined. Finally, by using the logistic regression analysis, a predictive scoring model was suggested. A total of 114 patients with mean age of 43.13±13 years (90 females (78.9%) were studied. Prevalence of malignancy in final pathology was 75.9%, 39.0%, and 15.9% in FNA biopsies which were malignant, undetermined and benign, respectively. FNA correctly verified 48.9% of the malignant pathologies. In all, smaller nodules, hypoechogenicity, metastatic lymphadenopathy, oval shape, thick or incomplete halo, ill margins, microcalcification, and heterogenicity are associated with malignancy (All, P<0.05). We recommend that complementary tools upon ultrasonographic data along with FNA can be helpful for more accuracy and early diagnosis of malignant thyroid nodules. © 2018 Tehran University of Medical Sciences.