Our findings suggest that PSMA expression and [
18F]FDG PET/CT findings play a complementary role in TC risk stratification. The majority of patients who expressed PSMA recurred (15/18). [
18F]FDG positive scan was related to high recurrence rate (11/18), risk of RAI-R development (9/12), and presence of disease at LFU (10/13). Moreover, patients with positive scan exhibited higher PSMA expression than patients with negative [
18F]FDG PET/CT (Fig.
1e, f).. The expression of PSMA, a marker of neovasculature formation [
10,
11], has been reported to be related to tumor recurrence in TC [
8,
12]; while high [
18F]FDG uptake, a marker of less differentiated thyroid tumors [
13], indicates poorer prognosis [
4,
14]. Ciappuccini et al. [
7] found a significantly different immunoreactive score in patients with positive PET/CT compared to those with negative scan, and a lower disease-free survival in patients who presented [
18F]FDG uptake. Collectively, our cohort included a higher proportion of patients who expressed PSMA than those reported by Ciappuccini et al. [
7] (17/23 vs. 30/44), especially in case of negative scan (6/12 versus 3/30). Nonetheless, as previously shown [
7], we confirmed that a higher number of recurrent patients with positive [
18F]FDG PET/CT expressed PSMA at a higher level (79% vs. 25%) than those with a negative scan (11/18 versus 4/18). Preliminary data showed that PSMA-targeting imaging might add information to [
18F]FDG PET/CT, potentially impacting on patient management [
6,
15], although detection rates ranging from 25 to 100%, were overall inferior to [
18F]FDG PET/CT, when compared [
16]. We found an absolute agreement between [
18F]FDG positivity and PSMA expression, while we can speculate a little more in patients with negative [
18F]FDG PET/CT, with a potential impact of PSMA positivity on disease recurrence rate. However, the numbers are small to draw conclusions and taking into consideration potential selection bias of this retrospective study. Indeed, [
18F]FDG PET/CT is generally performed in clinical practice in patients with suspected recurrence, as also suggested by thyroglobulin levels in our cohort. Therefore, a high prevalence of recurrence was expected in both patients with positive and negative scan. However, when considering only patients with negative scan, the probability of recurrence in case of PSMA expression was double than in patients with both negative immunostaining and [
18F]FDG PET/CT (4 vs. 2 and 3 vs. 3, respectively).
Our results suggest that PSMA expression might be an alternative to [
18F]FDG PET/CT as prognostic biomarker. Moreover, PSMA expression assessment has the advantage of being not expensive, and available at diagnosis with a minimal additional effort from pathologists. Therefore, differently from [
18F]FDG that—even if recognized as prognostic TC biomarker—is not routinary used probably because of costs and limited direct impact on patient management, PSMA expression assessment can be easily implemented in clinical practice. This would allow to set up at baseline a tailored treatment and follow-up according to PSMA staining findings. Moreover, radioligand therapy might become a valuable therapeutic option to offer to recurrent patients, upon in vivo target confirmation by PET/CT. Notably, in our cohort both patients classified as low risk of recurrence according to ATA Guidelines [
4] expressed PSMA (40% and 70%, respectively) and recurred. Additionally, patients with intermediate risk and negative PSMA immunostaining recurred less than patients belonging to the same class risk and PSMA expression (1 versus 3, respectively). If PSMA positivity will be confirmed in dedicated prospective trials as an outcome predictor, immunohistochemistry for PSMA assessment could be implemented in the clinics alongside other pathological data. We have to acknowledge some limitations. Firstly, the retrospective design and the criteria used for patients’ selection possibly affected our results. For some patients, a limited follow-up time was available.