Prevalence and Risk of Malignancy of Thyroid Incidentalomas Detected by 18 F-Fluorodeoxyglucose Positron-Emission Tomography

Dear Editor:

We read with great interest the recent article by Soelberg et al. (1) concerning the proportion of incidental thyroid cancers found by ¹⁸F-fluorodeoxyglucose positron-emission tomography or positron-emission tomography/computed tomography (¹⁸F-FDG PET or PET/CT).

The authors performed a systematic review of the literature to calculate the risk of malignancy in thyroid incidentalomas detected by ¹⁸F-FDG PET or PET/CT. We have really appreciated the valuable systematic review performed by these authors and the important information retrieved about the risk of thyroid malignancy, in particular those regarding the importance of the different uptake pattern (focal or diffuse) in the thyroid. Furthermore, the discussion about the possible sources of heterogeneity between studies was really relevant.

Nevertheless, we would like to comment some methodological aspects of this article, adding further evidence-based data about the prevalence and the risk of malignancy of thyroid incidentalomas detected by ¹⁸F-FDG PET or PET/CT.

Prevalence of focal and diffuse thyroid incidentalomas detected by ¹⁸F-FDG PET or PET/CT. First of all, to find the overall prevalence for unexpected focal as well as diffuse uptake in the thyroid gland, the ratio of patients with positive thyroid findings to the total number of patients evaluated using PET or PET/CT was calculated by these authors, adding the data retrieved from each included articles. This was done separately for focal and diffuse uptake (1). We questioned the accuracy of this statistical approach; in fact, the authors obtained the overall prevalence of thyroid incidentalomas by adding the cases retrieved from the different included articles, but these studies were very heterogeneous, as recognized by the authors.

We believe that a meta-analytic approach could be more adequate in this case. A meta-analysis could be performed to calculate a pooled prevalence of focal or diffuse thyroid incidentalomas detected by ¹⁸F-FDG PET or PET/CT, taking into account the different weight of each included study (depending on the sample size). Furthermore, a meta-analysis could demonstrate the statistical heterogeneity between the included studies.

Therefore, to calculate a pooled prevalence of focal and diffuse thyroid incidentalomas detected by ¹⁸F-FDG PET or PET/CT, we performed a proportion meta-analysis (by means of StatsDirect statistical software) as an alternative method compared to that of Soelberg et al. (1) by using the data retrieved by these authors. The prevalence of focal and diffuse thyroid incidentalomas by each included study was computed; subsequently, a pooled proportion was obtained using the statistical software, taking into account the different weight of each study. A random-effects model has been used for statistical pooling of the data; this model is suggested in the case of heterogeneity between the studies. Pooled data are presented with 95% confidence intervals [CIs]. An I-square statistic was performed to test for heterogeneity between studies.

Performing this meta-analysis, the pooled prevalence of patients with incidental focal uptake in the thyroid detected by ¹⁸F-FDG PET or PET/CT was 1.78% [CI 1.21%–2.47%] (see Supplementary Data, available online at www.liebertpub.com/thy). The pooled prevalence of patients with incidental diffuse ¹⁸F-FDG uptake in the thyroid was 1.51% [CI 0.62%–2.79%] (Supplementary Data). Both for focal and diffuse thyroid incidentalomas, a great statistical heterogeneity between studies was found (Supplementary Data).

Risk of malignancy of thyroid incidentalomas detected by ¹⁸F-FDG PET or PET/CT. Soelberg et al. (1) separately calculated the overall prevalence of malignancy in patients with a focal or diffuse uptake by the ratio of assigned cases with malignancy and the total number of patients with follow-up in each group (1). These data were obtained by adding the cases retrieved from the included articles. Again, we believe that a meta-analytic approach could be a useful alternative method to calculate the pooled proportion of malignancy in thyroid incidentalomas detected by ¹⁸F-FDG PET or PET/CT for the reasons already mentioned above.

Therefore, we performed a proportion meta-analysis, calculating the pooled prevalence of malignancy both in focal and diffuse thyroid incidentalomas detected by ¹⁸F-FDG PET or PET/CT. Using the same methods described above, we found that the pooled prevalence of malignancy in patients with incidental focal ¹⁸F-FDG uptake in the thyroid was 35.4% [CI 30.2%–40.8%]. Heterogeneity between studies was found (Supplementary Data). The pooled prevalence of malignancy in patients with incidental diffuse ¹⁸F-FDG uptake in the thyroid was 4.8% [CI 2.1%–8.4%]. In this case, a significant heterogeneity between studies was not revealed (Supplementary Data).

Overall, the results obtained by our meta-analysis about the prevalence and the risk of malignancy of thyroid incidentalomas detected by ¹⁸F-FDG PET or PET/CT were quite similar to those obtained by Soelberg et al. (1), but the methodological approach was different.

The meta-analytical approach is usually subject to some relative limitations, namely the small number of included studies, the possible publication bias, and the presence of heterogeneity between studies.

The small number of included studies and the small sample size of the included studies may limit the statistical power of the pooled analysis; in our case, a significant number of studies were published about thyroid incidentalomas detected by ¹⁸F-FDG PET or PET/CT, increasing the power of the pooled analysis.

Publication bias is a major concern in all forms of pooled analyses, as studies reporting significant findings are more likely to be published than those reporting nonsignificant results. Indeed, it is not unusual for small-sized early studies to report a positive relationship that subsequent larger studies fail to replicate. We have demonstrated that there was not significant publication bias regarding the articles that evaluated the risk of malignancy of focal thyroid incidentalomas, by using a funnel plot and Egger's test (Supplementary Data).

The heterogeneity between studies remains the main factor limiting our pooled analysis, as it was in the study of Soelberg et al. (1). This heterogeneity is likely to arise through diversity in methodological aspects between different studies. The baseline differences among the patients in the included studies may also have contributed to the observed heterogeneity of the results. However, such variability was accounted for in the random-effects model of the pooled analysis.

Anyway, it should also be considered that in most studies, follow-up and verification of ¹⁸F-FDG uptake in the thyroid were only available for some patients. It is therefore likely that the true risk for malignancy in solitary ¹⁸F-FDG-positive thyroid nodules is overstated in most publications (2).

Difference in standardized uptake value between benign and malignant thyroid incidentalomas detected by ¹⁸F-FDG PET or PET/CT. We have some considerations about the maximum standardized uptake values (SUVmax) of thyroid incidentalomas. Soelberg et al. performed a subanalysis calculating the mean SUVmax of benign and malignant thyroid incidentalomas detected by ¹⁸F-FDG PET or PET/CT, retrieving the data by different studies. The authors reported a mean SUVmax of 4.8 and 6.9 in benign and malignant lesions, respectively, with a statistically significant difference (1). This approach is not only heavily biased as stated by the authors, but also inappropriate from a nuclear medicine point of view.

SUV is a semiquantitative measure of ¹⁸F-FDG uptake calculated using the reported formula (3): \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document} \begin{align*}{\rm SUV} = \frac {\hbox{activity measured in a volume of interest}} {\hbox{activity administered/body weight}}\end{align*} \end{document}

This value is influenced by several factors related to the patient and by various technical aspects and procedures. In practice, there are several sources of bias and variance that are introduced in the measurement of SUVmax (3,4). Therefore, we believe that it is not only highly biased but also inappropriate to calculate a mean SUVmax obtained by different studies in which PET scans were performed with different tomographs, protocols, injected activity, and different patient characteristics.

Furthermore, although some authors reported that SUVmax was significantly lower in benign than malignant ¹⁸F-FDG-positive thyroid incidentalomas, there is a considerable overlap in SUVmax between benign and malignant thyroid incidentalomas, and in the clinical practice, SUVmax cannot discriminate between the two groups (2). Therefore, caution against using SUVmax as the sole parameter on which to base management of thyroid incidentalomas should be recommended (2), as also stated by the authors in the discussion (1).

In conclusion, we agree with Soelberg et al. (1) that incidental focal ¹⁸F-FDG uptake in the thyroid is associated with a significant risk of malignancy. Cytologic and/or histologic diagnosis should therefore be obtained for focal lesions if the nature of the thyroid disease has the potential to change patient management.