*3.2. Qualitative Analysis (Systematic Review)*

The seven articles included in the systematic review had been published between 2013 and 2021. All of them had a retrospective design. Three studies were carried out in China, while India, Republic of Korea, USA and Italy, contributed one study each. The characteristics of the studies and their patients' populations are summarized in Table 1.

#### 3.2.1. Technical Aspects

The imaging modality consisted of PET/CT with low-dose computed tomography settings in all cases. Information on fasting before 18F-FDG (4–6 h) injection were available in all articles. On the other hand, fasting was not required before 18F-FES injection.

The injected radiotracer activity ranged from 185 to 370 MBq for 18F-FDG PET/CT and from 111 to 222 for 18F-FES PET/CT. The time interval between radiotracer injection and PET/CT image acquisition was similar across the studies, being 60 min for both tracers in the majority of cases but one [19] in which 18F-FES PET/CT acquisition started 80–100 min after the tracer injection.

In all studies, PET image analysis was performed by a combination of qualitative (visual) and semi-quantitative analysis through the calculation of the maximum standardized uptake values (SUVmax).

On visual analysis, 18F-FDG focal uptake greater than the surrounding normal tissue that could not be explained by physiological activity was considered as positive in five studies [12,18,20,22,23]. In the remaining two studies, the criteria to classify 18F-FDG PET findings as positive were not clearly specified [17,21]. When 18F-FES PET/CT was considered, in three studies a SUVmax cut-off was introduced to interpret as positive each focal tracer uptake [18,20,22]. On the other hand, a visual interpretation (i.e., uptake higher than surrounding background) was used in another study [12], and in a further three analyses, the criteria were not reported [17,19,21]. All technical aspects are summarized in Table 2.

#### 3.2.2. Diagnostic Performance

The seven articles selected for the systematic review were published between 2013 and 2021 and included populations consisting of 7 to 49 patients affected by ER+ BC (Table 1). Table 3 details the rate of positive cases at the PBA and LBA.

Overall, 18F-FDG PET/CT and 18F-FES PET/CT showed very high sensitivity in detecting sites of disease in ER+ BC patients. Indeed, in two studies, 18F-FDG PET/CT identified more patients with BC lesions/metastases than 18F-FES PET/CT [19,22]. Conversely, in one study, 18F-FES PET/CT showed more BC lesions than 18F-FDG PET/CT [21]. In the other four studies [12,17,18,20], both diagnostic procedures identified at least one BC lesion in the 75 patients analysed. Overall, no significant differences between the two methods were reported.

When the ability for detecting each single lesion/metastasis was considered (i.e., LBA), 18F-FES PET/CT identified more BC lesions in three studies [18,21,22] and 18F-FDG PET/CT identified more BC lesions in another three studies [12,19,20]. In one study, including only patients with primary BC, both modalities identified the same number of lesions [17]. As the main finding, we can point out that the studies reporting a higher number of 18F-FDG-positive lesions included patients often affected by ductal carcinoma (i.e., 96%) [19] among which were also included those with ER+ HER2 + BC (11%) [19] and those with liver metastases (>10%) [12]. In addition, these patients were studied at a time of suspicious relapse when heterogeneity, due to the comparison of metastatic ERclones, is more frequent [12,20]. On the other hand, studies showing a higher number of 18F-FES-positive lesions often analysed patients affected by lobular BC [21] or patients with a high prevalence of bone metastases [22]. Moreover, patients were predominantly affected by ER+ HER- BC [18,22] and were often studied at the time of their first staging [18].

When the impact on clinical decision making was considered, only one study [18] reported that 18F-FES PET/CT was able to change therapeutic strategies in 5 out of the 19 (26.5%) patients analysed at the time of first staging. Indeed, this diagnostic procedure was able to properly downstage two patients and upstage three when compared with 18F-FDG PET/CT.



**Table 2.** *Cont.*


**Table 3.** Data available in the seven studies included in the present systematic review.

Legend. +ve: positive.

#### 3.2.3. Quality Assessment of the Studies

The risk of bias was assessed according to seven items, which are listed in Table 4. The overall bias score ranged from none to 2.5; therefore, no study had to be excluded because of high bias risk. The most frequent sources of possible bias were the "study test" and "reference standard categories" since, in some studies, it was unclear whether a blinded evaluation of the two methods had been performed. In particular, in the study by Ulaner et al. [21], the same reader assessed both examinations. In opposition to this, risks regarding the feasibility were almost never detected.

**Table 4.** Quality assessment of the studies and risk of bias in each study considered.


Legend: H = high, L = low, U = unclear.

#### *3.3. Quantitative Analysis (Meta-Analysis)*

The pooled sensitivity of 18F-FES PET/CT and 18F-FDG PET/CT in terms of PBA and LBA was computed (Table 5 and Figures 2–5) based on the available data (see Table 3). Regarding PBA, the pooled sensitivity of 18F-FDG PET/CT and 18F-FES PET/CT was 97% and 94%, respectively, with overlapping 95% confidence intervals. In the LBA, however, we observed a wider difference between the pooled sensitivity of the two methods (95% for 18F-FES vs. 85% for 18F-FDG). Although the 95% confidence intervals of the two methods were overlapping, the values of 18F-FES PET/CT were consistently at the higher end of the spectrum (93–97%), while those of 18F-FDG PET/CT showed a much larger spread (68–100%). In all the analyses, high heterogeneity was found (Table 5).

On the basis of the above results, the heterogeneity of PBA and LBA was explored by using the following covariates: timing of the studies (i.e., PET/CT assessment on staging and on time of relapse), their sample sizes, the prevalence of ductal or lobular BC, and the prevalence of bone and liver metastases.

2 YHUDOO,A 3 

6XEJUR X S 5 H VWDJ LQ J ,A 3  ʝ

2 YHUDOO,A 3 

6 WX G LHV

6 X E J UR X S 6 WDJ LQ J ,A 3 

6XEJUR X S 5 H VWDJ LQ J ,A 3  ʝ

2 YHUDOO,A 3 


**163/171**

**0.97 (0.94, 0.99)**

**Table 5.** Pooled sensitivity for PBA and LBA of 18F-FES PET/CT and 18F-FDG PET/CT.

**Figure 2.** Sensitivity of 18F-FDG PET/CT at the level of patients across studies.

 3URSRUWLRQ

**Figure 3.** Sensitivity of 18F-FES PET/CT at the level of patients across studies.

6 WX G LHV

6 X E J UR X S 6 WDJ LQ J ,A 3 

6XEJUR X S 5 H VWDJ LQ J ,A 3  ʝ

2 YHUDOO,A 3 

**Figure 5.** Sensitivity of 18F-FES PET/CT at the level of the lesions across studies.

However, these last variables could not be tested, given incomplete and inconsistent data, and only the variable "timing of the studies" was explored. As illustrated in Figures 2 and 3, the heterogeneity of the 18F-FES PET/CT and 18F-FDG PET/CT results was no longer present in the PBA when we considered these two groups separately. In fact, in the staging scenario, the sensitivity in the PBA of both the 18F-FES PET/CT and 18F-FDG PET/CT was 97% (Figures 2 and 3). Conversely, at the time of restaging, the patient-based sensitivity of 18F-FES PET/CT and 18F-FDG PET/CT was 90% and 95% respectively (Figures 2 and 3).

When LBA was conducted, the sensitivity of 18F-FES PET/CT and 18F-FDG PET/CT at the time of first staging was 88% and 89% respectively, while at the time of restaging it was 81% and 98%, respectively (Figures 4 and 5). Interestingly, the sensitivity of 18F-FES PET/CT at the time of restaging was significantly higher than that of the same procedure at the time of staging (Figure 4).

#### **4. Discussion**

In this systematic review and meta-analysis, we aimed to clarify the diagnostic role of 18F-FES PET/CT in BC patients compared to 18F-FDG PET/CT. In fact, 18F-FES PET/CT has gained significant attraction as a non-invasive means to predict the effectiveness of hormone blockade. However, its potential in the mere diagnostic evaluation is still debated [1,8,21]. In this study, we gathered all the available evidence of its sensitivity in ER+ BC patients when compared with the more commonly used 18F-FDG PET/CT.

Our qualitative and quantitative assessment in this particular BC subpopulation showed that there are no significant differences in terms of sensitivity between the two PET tracers at the PBA. Indeed, both molecular imaging modalities proved able to detect patients affected by ER+ BC with similarly high sensitivity. The lack of significant difference between these two modalities can, however, be explained by the selection of patients included in the analysis. These patients were evaluated at the time of first staging for an already known primary tumour or who underwent PET at the time of relapse to evaluate the extension or the ER expression of the metastatic disease. This population is indeed characterized by a very high prevalence of a true positive BC lesion; the probability that at least one these lesions was detected by one of the two molecular imaging procedures was very high. The small diagnostic advantage of the 18F-FDG PET/CT over 18F-FES PET/CT (97% vs. 94%) reported in our analysis seems related to the variability in terms of the clinical characteristics of the patients. Indeed, the high prevalence of metastatic heterogeneity, often present at the time of restaging, can be associated to false negative 18F-FES PET/CT results. When the diagnostic performances of these two imaging procedures at the time of recurrence was explored by means of a PBA, the sensitivity of 18F-FDG PET/CT and that of 18F-FES PET/CT was 95% and 90%, respectively.

Conversely, when we investigated the diagnostic sensitivity of these diagnostic tools by means of an LBA, we found that the sensitivity of 18F-FDG PET/CT and that of 18F-FES PET/CT was 85% and 95%, respectively. Indeed, when the analysis was focused on the time of disease relapse, 18F-FES PET/CT was more sensitive than 18F-FDG PET/CT (98% vs. 81%) with a trend towards significance (95% CI: 97–100% and 56–100%, respectively).

Overall, the use 18F-FDG PET/CT as a first-line examination appears to be the best strategy to stage and restage the ER+ BC subjects, being able to identify the highest number of true positive patients. However, if this approach is applied to a population of ER+ BC with low metastatic heterogeneity (i.e., lobular breast cancer, or ductal breast cancer with a high percentage of ER positive and HER2 negative cells) it could underestimate the actual disease burden since such a clone could show low glucose avidity. Indeed, although 18F-FES PET/CT has low sensitivity in detecting liver metastases (given the intense background tracer uptake in this organ), it has a very high sensitivity in disclosing peripheral lesions in other organs or tissues, such as bone lesions, which represent the most frequent sites of disease in ER+ BC [22]. In addition, as reported by Gupta et al. [12], one of the diagnostic advantages of this receptor PET tracer is its higher specificity in characterizing cervical, axillary and mediastinal lymph nodes, which may often present unspecific uptake at 18F-FDG PET/CT and thus be misinterpreted on 18F-FDG PET/CT.

Although our systematic review and meta-analysis reported interesting and useful results to understand the weaknesses and strengths of the two imaging procedures, some limitations should also be borne in mind. First, only seven studies, investigating relatively small patient populations, were fit for inclusions in this meta-analysis. In addition, all these seven studies showed a retrospective design, and three out of seven included a low number of ER+ BC patients. However, the selection criteria limited us to studies that tested both 18F-FDG PET/CT and 18F-FES PET/CT within a restricted time frame. Second, a computation of specificity was not possible since none of the studies reported the true negative rate. Third, the absence of histological confirmation of suspected distant metastases detected by both modalities is an important bias that could have affected some of the studies included in this analysis [12,20–22], and we cannot exclude the possibility that some of the metastases detected by PET tracers may have been false-positive findings. However, given the elevated prevalence of patients with disseminated disease, the likelihood of any given finding to be a false negative was relatively low. Moreover, ethical and practical reasons prevented the execution of a biopsy evaluation of each single lesion. In addition, in at least 3 out of 7 studies, a multidisciplinary follow-up (consisting of clinical and imaging evaluation) was available. Fourth, only two out of the seven studies reported information regarding the menopausal status of the patients. Indeed, these data could be of value to correctly estimate the real sensitivity of 18F-FES PET/CT that could theoretically be affected by competitive binding of high oestradiol concentration. However, the exact endogenous oestrogen concentration that can have a measurable effect on tumour 18F-FES uptake remains hitherto unexplored [9]. Statistically significant heterogeneity of the 18F-FES PET/CT pooled heterogeneity was found across studies. Regrettably, the sparse data available in the studies did not allow exploration of the heterogeneity with the exception of the variable "timing of the studies".

Finally, at this time, data about the cost/effectiveness of 18F-FES PET/CT in comparison with that of 18F-FDG PET/CT do not exist. However, it is conceivable that, when the overall expense is considered, the additional information provided by FES might help in optimizing the cost by guiding the choice of the most appropriate therapeutic protocol.

#### **5. Conclusions**

The present data show that both 18F-FDG PET/CT and 18F-FES PET/CT represent accurate imaging procedures in patients with ER+ BC, providing comparable results in terms of sensitivity. However, in the field of lesion detection, we observed a non-negligible difference in favour of 18F-FES PET/CT. Thus, the use of 18F-FES PET/CT as a first-line molecular imaging procedure might be considered in lobular breast cancer or ductal breast cancer with a high percentage of ER and HER2 negative. However, larger, multicentre and prospective studies are required to confirm these preliminary indications.

**Author Contributions:** Conceiving the study and writing the manuscript draft, A.P., G.B., P.T.; performing the literature search and selecting the studies, F.F., G.B.; performing the meta-analysis, P.T.; reviewing the manuscript draft, F.F., G.T. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Not applicable.

**Conflicts of Interest:** The authors declare no conflict of interest.
