Comparison of diagnostic sensitivity of [18F]fluoroestradiol and [18F]fluorodeoxyglucose positron emission tomography/computed tomography for breast cancer recurrence in patients with a history of estrogen receptor-positive primary breast cancer

Background To compare the diagnostic sensitivity of [18F]fluoroestradiol ([18F]FES) and [18F]fluorodeoxyglucose ([18F]FDG) positron emission tomography/computed tomography (PET/CT) for breast cancer recurrence in patients with estrogen receptor (ER)-positive primary breast cancer. Methods Our database of consecutive patients enrolled in a previous prospective cohort study to assess [18F]FES PET/CT was reviewed to identify eligible patients who had ER-positive primary breast cancer with suspected first recurrence at presentation and who underwent [18F]FDG PET/CT. The sensitivity of qualitative [18F]FES and [18F]FDG PET/CT interpretations was assessed, comparing them with histological diagnoses. Results Of the 46 enrolled patients, 45 were confirmed as having recurrent breast cancer, while one was diagnosed with chronic granulomatous inflammation. Forty (89%) patients were ER-positive, four (9%) were ER-negative, and one (2%) patient did not undergo an ER assay. The sensitivity of [18F]FES PET/CT was 71.1% (32/45, 95% CI, 55.7–83.6), while that of [18F]FDG PET/CT was 80.0% (36/45, 95% CI, 65.4–90.4) with a threshold of positive interpretation, and 93.3% (42/45, 95% CI, 81.7–98.6) when a threshold of equivocal was used. There was no significant difference in sensitivity between [18F]FES and [18F]FDG PET/CT (P = 0.48) with a threshold of positive [18F]FDG uptake, but the sensitivity of [18F]FDG was significantly higher than [18F]FES (P = 0.013) with a threshold of equivocal [18F]FDG uptake. One patient with a benign lesion showed negative [18F]FES but positive [18F]FDG uptake. Conclusions The restaging of patients who had ER-positive primary breast cancer and present with recurrent disease may include [18F]FES PET/CT as an initial test when standard imaging studies are equivocal or suspicious.


Background
Breast cancer is one of the most frequently diagnosed cancers and is the leading cause of cancer death in women [1]. About 75% of all breast cancer expresses estrogen receptor (ER) at the time of initial diagnosis [2]. The 20-year risk of distant recurrence after adjuvant endocrine therapy for ER-positive breast cancer ranges from 10 to 40% [3]. The diagnosis of recurrence and restaging are important for determining appropriate treatment [4].
Current guidelines on breast cancer recommend that the staging evaluation of women who present with recurrent breast cancer includes standard imaging studies [5,6]. Positron emission tomography/computed tomography (PET/ CT) with [ 18 F]fluorodeoxyglucose ([ 18 F]FDG) would only be used in situations where standard staging studies are equivocal or suspicious [5,6]. Although clinical studies have indicated high diagnostic accuracy with sensitivities ranging from 81 to 100% and specificities ranging from 52 to 91% and have suggested the superiority of [ 18 F]FDG PET/CT over standard imaging studies [7][8][9][10][11], these studies are largely retrospective with heterogeneous cohorts, and there are also methodological issues of concern [7, 10,11]. The diagnostic accuracy was not separately determined in patients who presented with equivocal or suspicious imaging studies [8,[12][13][14]. In addition, both false-negative and false-positive [ 18 F]FDG PET/CT results [7,10] are inherent in [ 18 F]FDG PET/CT [15,16]. The consensus is that biopsy is more likely to provide useful information [6]. Moreover, the reference standard used in previous studies has ranged from histopathological diagnosis to clinical or radiological follow-up, and misclassification of disease might have led to positive or negative biases. Finally, the positivity threshold of [ 18 F]FDG PET/CT may differ between studies, depending on how to determine the foci of increased [ 18 F]FDG uptake are related to benign conditions. Equivocal decisions are therefore likely to be inevitable [8,9,11,12,[17][18][19].
The histological and biological characteristics of breast cancer have an important impact on tumor visualization with [ 18 F]FDG PET/CT [20]. [ 18 F]FDG uptake correlates with histologic grade and tumor proliferation index [15,21,22], and [ 18 F]FDG uptake is higher in ER-negative tumors [15]. Accordingly, the relatively lower [ 18

Study design
The study patients were identified from our database of a prospective patient cohort who underwent [ 18 F]FES PET/CT for the assessment of ER status because of recurrent or metastatic breast cancer lesions [31]. The patients' medical histories and [ 18 F]FDG PET/CT studies performed to identify unsuspected regional nodal disease and/or distant metastases were retrospectively reviewed. Our institutional review board approved this study protocol and waived the requirement for informed consent. This study was conducted in accordance with the Declaration of Helsinki and our institutional guidelines.

Patients
All eligible patients had ER-positive primary breast cancer with suspected first recurrence at presentation, were enrolled in a previous prospective cohort study to assess [ 18 F]FES PET/CT, and underwent [ 18 F]FDG PET/CT within 60 days before needle biopsy or surgery. Patients were excluded if they had ER-negative or ER-unknown primary breast cancer. The complete selection criteria, setting, location, and dates of the prospective cohort study were previously reported [31]. The patient population used in this study represented a consecutive sample of eligible patients. The production of [ 18 F]FES was described previously [33]. PET/CT imaging was performed from the skull base to upper thigh using a Discovery PET/CT 690 or 710 scanner (GE Healthcare), 80-100 min after intravenous injection of 111-222 MBq (3-6 mCi) of [ 18 F]FES [31]. PET/CT images were reconstructed using the manufacturer-provided iterative algorithm with 4 iterations and 18 subsets.

Histological assessment
Experienced breast pathologists who were blinded to the results of the [ 18 F]FES PET/CT reviewed core-needle biopsy and surgical samples in accordance with routine standard procedures. The recurrent or metastatic specimens taken from the breast, liver, ovary, uterus, or bone were excluded from the analysis. Clinical information and other laboratory test results were available to the assessors of the histological examinations. Immunostaining for ER was performed using the 6F11 mouse monoclonal antibody (NCL-ER-6F11, Novocastra Laboratories), as previously described [35]. Semi-quantitative ER receptor expression was evaluated according to the Allred score [36]. ER status was considered positive if Allred scores were 3 or higher.

Analysis of PET/CT imaging
The recurrent lesion from which a sample was taken for histological examination was selected as the reference lesion for the sensitivity analysis, as reported previously [31]. Three board-certified nuclear medicine physicians (S.B., J.Y.K., and S.J.N.) who were blinded to all patient information including histological diagnosis, independently assessed the [ 18 F]FES PET/CT images in a random sequence. The physicians qualitatively assessed the [ 18 F]FES uptake of reference lesions, as well as up to five additional non-reference lesions in the order of highest intensity relative to background activity [25,37]. Lesions were interpreted as ER-positive breast cancer when the majority of [ 18 F]FES assessments were positive.
Two board-certified nuclear medicine physicians (H.J.S. and D.Y.L.) interpreted the [ 18 F]FDG PET/CT images in consensus using a syngo.via image display system (VB20A, Siemens Healthineers), with the images displayed in a random sequence. These two physicians were also blinded to all patient information and the [ 18 F]FES PET/CT images; they only had knowledge of the location of the biopsy sites for [ 18 F]FDG PET/CT image evaluation. The lesions were classified as positive, equivocal, or negative lesions by qualitative analysis. An [ 18 F]FDG focus was interpreted as negative when it was related to a non-malignant process or the physiological biodistribution of [ 18 F]FDG. Focally abnormal [ 18 F]FDG uptake higher than that of surrounding tissues and that could not be related to non-malignant or physiological uptake was interpreted as positive.
[ 18 F]FDG uptake that could not be characterized was defined as equivocal [17].
For quantitative analysis, the readers calculated the maximal standardized uptake value (SUV) as SUV = activity (Bq/g)/[injected activity (Bq)/body weight (g)]. The [ 18 F]FDG SUV results were harmonized across different PET/CT scanners by matching the recovery coefficient profiles without partial volume correction, as described previously [38].

Statistical analysis
All data are presented as median values with interquartile range (IQR) for continuous variables, and numbers (%) for categorical variables. A two-sided P value of less than 0.05 was considered as the threshold for significance. Quantitative parameters were compared using the Mann-Whitney U test or Kruskal-Wallis H test and categorical data were compared using the McNemar test. All statistical tests were conducted using SPSS Statistics for Windows (version 21, IBM Company). The sensitivity of PET/CT was defined as the probability that a PET/ CT result would show positive uptake when the patient had histologically proven recurrent breast cancer. Estimates are presented with 95% Clopper-Pearson exact confidence intervals (CI).  Table 1. Fortyfive patients were confirmed as having breast cancer, but one patient was diagnosed with chronic granulomatous inflammation. Of the 44 patients with recurrent breast cancer who underwent an immunohistochemical study, four (9%) were ER-negative. One patient had a fine-needle aspiration biopsy that did not provide enough cells for immunohistochemical study.

Sensitivity of [ 18 F]FES and [ 18 F]FDG PET/CT
The median time between PET/CT and biopsy or surgery was 5 days (IQR 1-11) for [ 18 (Fig. 2). The other patient with ER-positive invasive ductal primary cancer had a biopsy-confirmed benign lesion that showed negative [ 18 F]FES but positive [ 18 F]FDG uptake.

Assessment of whole-body tumor burden by [ 18 F]FES PET/CT
Of the 45 patients with recurrent breast cancer, 25 patients (56%) had a total of 47 [ 18 F]FES-positive non-reference lesions (Fig. 2). All these patients had ER-positive reference lesions. These [ 18 F]FES-positive lesions were located in the bone (n = 10), regional lymph nodes (n = 10), distant lymph nodes (n = 11), lung (n = 11), pleura (n = 4), and breast (n = 1). Interestingly, three of nine patients who had ER-positive/[ 18 F]FES-negative reference lesions had [ 18 F]FES-positive non-reference lesions. One non-reference lesion of the three patients was confirmed as ER-positive metastatic breast cancer by histologic analysis (Fig. 3). If these three patients were to be added to the calculation of the sensitivity of [ 18 F]FES, the figure would be 77.8% (35/ 45, 95% CI, 62.9-88.8).

Discussion
In this study, we showed that [ 18   PET/CT results were of relevance, because discordant ER expression between primary and recurrent breast cancer was accurately determined. We also had nine falsenegative patients who displayed ER-positive immunohistochemistry but negative [ 18 F]FES. As seven of the nine [ 18 F]FES-negative patients showed a positive [ 18 F]FDG uptake, the false-negative results cannot be explained by the limited spatial resolution of PET/CT, but rather reflect the low ER expression levels in tumor, or inherent differences between ER assays and [ 18 F]FES uptake [31,39]. ER-positive but [ 18 F]FES-negative tumors may represent a functionally endocrine therapy-resistant breast cancer [23,24]. Further studies are required to clearly characterize those patients with recurrent breast cancer who are ER-positive but [ 18 F]FES-negative.
An [ 18 F]FES PET/CT strategy for diagnosing recurrent breast cancer would have false-negative results, which would generally lead to delayed treatment [7]. Therefore, Diagnostic accuracy studies are at risk of bias due to shortcomings in their design and conduct, and the results may not be applicable to other patient populations [42]. We retrospectively reviewed our database of a prospective cohort who underwent [ 18 F]FES PET/CT for the assessment of ER status in recurrent lesions, and included consecutive patients who were typical of those presenting for first breast cancer recurrence. However, our patient inclusion and exclusion criteria may have resulted in potential risks and limited applicability. We enrolled only those patients who had histological confirmation and who underwent [ 18 F]FES PET/CT for restaging to guide optimal therapy. The diagnostic sensitivity may be higher if lesions are large enough for a tissue biopsy to be attained. In addition, PET/CT studies might have not been performed in situations where standard staging studies are equivocal or suspicious, as recommended by the guidelines. Nevertheless, we believe that the primary purpose of this study, which was to compare sensitivity between [ 18 F]FDG and [ 18 F]FES PET/CT, is unlikely to be influenced by our patient selection. We excluded only one patient who had no [ 18 F]FDG PET/CT. In most patients, the [ 18 F]FDG studies were performed for clinical reasons. There are no other potential risks of bias or applicability regarding the quality of the diagnostic sensitivity data [43]. Furthermore, the qualitative [ 18 F]FES positivity threshold could be precisely determined, as suggested by a very low P

Conclusion
The sensitivity of [ 18 F]FES PET/CT in patients who had ER-positive breast cancer and presented with recurrent disease was comparable to or lower than that of 18 F-FDG, depending on the threshold used. Although we were not able to directly demonstrate a high specificity in this study, the well-established high specificity indicates that positive [ 18