58 chapter 2 is not routinely recommended for the diagnostic workup of indeterminate thyroid nodules due to limited clinical validation, despite a 2011 meta-analysis by Vriens et al. that demonstrated 95% sensitivity and 96% NPV in indeterminate thyroid nodules larger than 15mm [17, 304]. Results of available individual studies were mutually consistent despite limited sample sizes. Especially the first studies showed extremely promising results, each reporting 100% sensitivity [37, 303, 305, 306]. De Geus-Oei et al. argued that implementation of [18F]FDG-PET could reduce the number of futile hemithyroidectomies for benign nodules by 66%, likely outweighing the costs of the extra scans and suggesting cost-effectiveness of this technique in the preoperative setting [37]. A subsequent study suggested a less optimistic 39% reduction in futile surgeries, following a lower benign call rate [306]. More recent studies demonstrated more modest performance of [18F] FDG-PET(/CT) [58, 59, 307-309]. Overall, reported sensitivity and specificity of [18F]FDG-PET(/CT) to detect thyroid carcinoma in indeterminate thyroid nodules ranged from 77% to 100% and from 33% to 64%, respectively. A negative index test was reported in approximately 40% of patients [37, 58, 59, 303, 305-310]. Several reasons for false-negativity were proposed, foremost small nodule size. It is how Traugott et al. explained their 20% false negative [18F]FDG-PET scans: eight lesions were histopathologically smaller than 1 cm. Excluding these, sensitivity and NPV increased to 100% [307]. [18F]FDG-avidity in very small nodules may be missed on [18F]FDG-PET due to the low volume of malignant cells and due to the partial volume effect: the detected [18F]FDG-concentration is underestimated dependent on nodule size in relation to the (limited) spatial resolution of the scanner. In larger nodules, this effect is negligible [37, 59]. Although the improving resolution of state-of-the-art PET scanners pushes the detection limit towards 10 mm, PET is less sensitive in lesions smaller than 15 mm on US. It less reliable to rule-out microcarcinomas [304]. Theoretically, the improving spatial resolution could also become a limitation of the technique: not only will there be less false-negatives, but likely also more false-positive results - leading to a decrease in the already limited specificity over time. In the currently available literature no such downward trend is noted, but future studies should monitor this possibility. Semi-quantitative [18F]FDG-PET Semi-quantitative analysis of [18F]FDG-PET is performed using the maximum standardized uptake value (SUVmax): the ratio between the maximum radioactivity concentration measured within a region of interest on the PET image (the ‘hottest’ voxel) and the decay-corrected amount injected radiotracer per unit of body mass. It reflects the [18F]FDG-concentration factor compared to a homogenous distribution of the radiotracer [311]. The SUVmax is generally significantly higher in malignant than in benign lesions [37, 59, 306, 308, 309, 311, 312]. There is a possible correlation between higher SUVmax values and increasing size in nodules, insufficiently explained by the abovementioned partial volume effect [59, 312]. Also, in FTC a higher SUV was associated with capsular or vascular invasion [310]. Nonetheless, even though Kresnik et al. demonstrated that all carcinoma and Hürthle cell
RkJQdWJsaXNoZXIy MTk4NDMw