52 chapter 2 are likely benign and categorized as TIRADS 3. Their risk of malignancy is ~1.7% in a cytologically unselected population. TIRADS 4 includes suspicious nodules, which are further classified according to an increasing malignancy risk into 4a (one suspicious US feature), 4b (two suspicious features) and 4c (three or four suspicious features). Nodules with all five suspicious US features are classified as TIRADS 5 and associated with a high 88% risk of cancer in an unselected population [274]. Studies that validated the TIRADS specifically in indeterminate thyroid nodules, showed that diagnostic accuracy depended on the chosen cut-off score and type of cytology [240, 273, 276-279]. Although TIRADS 5 scores were infrequently assigned in indeterminate nodules, a higher TIRADS score (4b/4c/5) was an accurate predictor of malignancy, especially in Bethesda IV cytology [240, 276, 278]. In Bethesda III nodules, lower TIRADS scores (3/4a) could also rule out malignancy [276, 279]. Prospective validation studies applying the TIRADS in indeterminate cytology are warranted to assess its possible clinical utility in indeterminate nodules. US performance in Hürthle cell nodules Cytological Hürthle cell nodules expressed a large variation of US characteristics [241, 246, 250, 261]. Many malignant and most benign Hürthle cell nodules had a benign US appearance [246, 261]. Only three US features possibly predictive of malignancy were reported in individual studies: both hypoechogenicity and hyperechogenicity (as opposed to isoechogenicity) [246], large nodule size [261], and microcalcifications [241]. Despite limited evidence, US evaluation does not seem reliable to differentiate Hürthle cell lesions. Availability and limitations of ultrasonography The major advantages of ultrasound over other additional diagnostics are its already permanent position in the workup of thyroid nodules, global availability and low costs. No additional resources nor hospital visits are needed to include US interpretations in preoperative management decisions and the investigation is noninvasive. Nonetheless, besides known limitations concerning interobserver variability and less reliable interpretation of small nodules, US feasibility in indeterminate thyroid nodules is limited by the presumed differences in US appearance of papillary and follicular thyroid malignancies, illustrated by the conflicting results for nodule shape and hypoechoic halo in indeterminate nodules. Consequently, local diagnostic accuracy likely follows variations in the local histopathological case mix. In addition, many of the available ultrasound studies are retrospective, limiting the power of the evidence. As the decision to perform FNAC is customarily based on the results of the prior US, the prevalence of suspicious US features in indeterminate cytology in these studies is presumably overestimated. Nonetheless, several individual US characteristics seem to have reasonable specificity in indeterminate nodules, although insufficient for accurate diagnosis. A combination of US features is likely more accurate, although current evidence does not support US-based surgical decision-
RkJQdWJsaXNoZXIy MTk4NDMw