290 chapter 3 [18F]FDG PET/CT was significantly associated with malignancy when qualitative [99mTc]Tc-MIBI scans were rated as negative, and (3) the association of a positive [18F]FDG PET/CT combined with a positive multiparametric US was significantly more specific than [18F]FDG PET/CT alone in identifying differentiated thyroid cancer. A combined approach by Trimboli et al. investigated whether [18F]FDG PET/CT could play a role in the stratification of nodules with an intermediate risk upon EU-TI-RADS in 93 unselected nodules with EU-TI-RADS 4 and 5, including 38 nodules with indeterminate cytology [462]. They found that thyroid lesions classified as EU-TI-RADS 4 and with no [18F]FDG uptake could be excluded from further examination. Another study by Piccardo et al. also investigated [18F]FDG PET/CT, EU-TI-RADS, and the Italian consensus for the classification and reporting of thyroid cytology (ICCRTC) to distinguish differentiated thyroid cancers and FNs from nodular hyperplasias in 201 Bethesda III and IV thyroid nodules [40]. On multivariate analysis, [18F]FDG PET/CT (OR 9.04), ICCRTC (OR 7.57), and EU-TI-RADS (OR 4.41) were all independent risk factors associated with differentiated thyroid carcinomas and FNs. These studies conclude that [18F]FDG PET/CT could serve as a reliable rule-out test in case of nodules with intermediate risk upon US stratification. Future perspectives Medical imaging plays an important role in the preoperative workup of cytologically indeterminate thyroid nodules. A comprehensive overview of imaging biomarkers exemplified in this chapter can be found in Table 2. Most biomarkers used in the clinical work-up are visual interpretation or basic quantitative metrics. AI applications and radiomic methodologies, on the other hand, are less well established, but are currently developed on a large scale. Extensive external validation should be performed in order to achieve implementation of AI-derived imaging biomarkers in clinical practice. Many of the imaging biomarkers have either an adequate rule-in or rule-out capacity, but no single biomarker seems to serve both purposes well. A multimodal stepwise approach using a sensitive rule-out test and a specific rule-in test complementing each other might provide the most conclusive diagnosis for indeterminate thyroid nodules [25]. It should be noted that test performance of a test depends on the patient population. With the introduction of US risk stratification systems, FNAC might be withheld more often for patients with a presumed benign nodule, thereby potentially changing the composition of patient population and increasing its associated risk of malignancy. The proportion of Hürthle cell nodules is additionally crucial, as these nodules should be considered a separate entity with varying diagnostic yield of the different imaging modalities [454]. In addition, the prevalence of malignancy and the performance, costs, and feasibility of the imaging techniques might vary globally. Clinical utility should be examined in local implementation studies.
RkJQdWJsaXNoZXIy MTk4NDMw