Response to Stack et al.

Dear Editor:

In reviewing the issue of whether incidentally detected thyroid nodules deserve to receive any particular attention when they have been detected by ¹⁸F-fluorodeoxyglucose positron emission tomography (¹⁸F-FDG PET), apart from that agreed upon for thyroid nodules in general (1,2), our original interest was based on the overwhelming increase in number of referrals to our endocrine unit within the last five years, and our lack of experience in the field (3). The more or less simultaneous appearance of several other recent reports (4 –6), and the interest shown in our review, as exemplified by Letters to the Editor (7,8), suggests that this topic has ignited curiosity simultaneously amongst many researchers from a variety of specialities, including endocrinology, surgery, otorhinolaryngology, and nuclear medicine. Despite this, the overall findings, and thus the recommendations, are remarkably similar. First, an incidental focal uptake in the thyroid, detected by ¹⁸F-FDG PET or PET/CT in a patient without known thyroid disease, is seen in 2.5% of cases (range 0.2–8.9% [95% confidence interval (CI) 1.7–3.4%]) (2). Second, around one-third of these nodules are malignant, with the predominant (∼84%) histotype being papillary thyroid cancer (2). Third, there is no safe maximum standardized uptake value (SUVmax)—the higher the SUVmax, the higher the risk of malignancy—which allows disregarding further investigations. Fourth, incidentalomas diagnosed with this modality should receive the same attention, and follow the same diagnostic algorithm, as if diagnosed in any other way.

With minor differences—mainly explained by considerable heterogeneity in study populations, standardization of investigations, and the number of individuals where surgical confirmation of the final diagnosis has been obtained—the above conclusions are agreed upon. When Stack et al. (8) focus our attention on their own data with respect to SUVmax, it could optimistically be hoped that a noninvasive test is imminent and could discriminate between malignant and benign incidentally diagnosed thyroid nodules employing ¹⁸F-FDG PET. Looking at their data, two main observations hinder such a progress in our handling of these patients. First, even if we disregard the retrospective nature of their study, the heterogeneity in study population, and potentially the methods of investigations, as well as the changes in referral pattern, a serious shortcoming persists. Over a 10-year period, 690 cases of incidental focal uptake (2.95% of those having had a PET scan) were found. However, complete data were only available for 359 of these and obtained in a total of 103 patients. Second, accepting that the SUVmax values, reported from 14 patients (although the figure seems to show 16 patients) with malignant and 50 patients with benign nodules, are representative and can be generalized to a larger population, the interpretation is not straightforward. Although the mean SUVmax was statistically significantly higher in those patients harboring malignancy [as also reported in our review (2)] than in those who had benign disease, the overlap is considerable. Based on this, and in agreement with Stack et al. (8), we cannot recommend basing the final diagnosis of thyroid malignancy solely on an SUVmax determination.

Therefore, the latest studies and the present correspondence (7 –9) do not advance our interpretation but reinforce the present consensus. While diagnostic imaging advances and standardization of SUVmax determinations may well increase sensitivity and specificity of discriminating between malignant and benign thyroid nodules diagnosed incidentally via ¹⁸F-PDG PET, major progress is more likely to come from other sources. The last few years have pointed at the possibility of a preoperative diagnosis of various thyroid malignancies using molecular analyses, even in routine air-dried fine needle aspiration (FNA) smears (10). An alternative strategy of ruling out malignancy is based on using gene-expression classifiers applied to FNA material (11). Both principles are already in use, especially in cases where FNA is nondiagnostic. Based on the aforementioned, we find that SUVmax determination plays a minor, if not negligible, role in the diagnostic algorithm for incidental thyroid nodules found by ¹⁸F-FDG PET. Independent of securing a diagnosis, which is currently the primary focus, it is equally important to obtain data on the consequences of a detected malignancy. It may well be that the natural history of thyroid malignancy diagnosed via ¹⁸F-FDG PET differs from that based on other routes of diagnosis. This information is pertinent in order to clarify the aggressiveness with which to pursue nodules diagnosed by ¹⁸F-FDG PET.