Application of PET Tracers in Molecular Imaging for Breast Cancer

Purpose of Review Molecular imaging with positron emission tomography (PET) is a powerful tool to visualize breast cancer characteristics. Nonetheless, implementation of PET imaging into cancer care is challenging, and essential steps have been outlined in the international “imaging biomarker roadmap.” In this review, we identify hurdles and provide recommendations for implementation of PET biomarkers in breast cancer care, focusing on the PET tracers 2-[18F]-fluoro-2-deoxyglucose ([18F]-FDG), sodium [18F]-fluoride ([18F]-NaF), 16α-[18F]-fluoroestradiol ([18F]-FES), and [89Zr]-trastuzumab. Recent Findings Technical validity of [18F]-FDG, [18F]-NaF, and [18F]-FES is established and supported by international guidelines. However, support for clinical validity and utility is still pending for these PET tracers in breast cancer, due to variable endpoints and procedures in clinical studies. Summary Assessment of clinical validity and utility is essential towards implementation; however, these steps are still lacking for PET biomarkers in breast cancer. This could be solved by adding PET biomarkers to randomized trials, development of imaging data warehouses, and harmonization of endpoints and procedures.


Introduction
Over the last decade, there has been an increasing interest in molecular imaging with positron emission tomography (PET), in particular in the field of oncology. PET imaging is a noninvasive tool to obtain qualitative and quantitative wholebody information of biological processes. Molecular imaging in breast cancer (BC) is of particular interest, as it can visualize the estrogen receptor (ER), human epidermal growth factor receptor 2 (HER2), and proliferation. However, molecular imaging with PET has not been widely adopted in clinical practice of BC. Only two radiotracers (2-[ 18 F]-fluoro-2deoxyglucose ([ 18 F]-FDG) and sodium [ 18 F]-fluoride ([ 18 F]-NaF)) are incorporated in cancer management guidelines, such as National Comprehensive Cancer Network (NCCN) and European Society for Medical Oncology (ESMO). In order to improve successful implementation of PET imaging biomarkers into clinical practice, it is essential to identify potential hurdles. Recently, an international consensus meeting resulted in the "imaging biomarker roadmap," describing the steps of imaging biomarkers towards clinical practice [1••]. In this review, we describe the current status of PET biomarkers for BC, according to this roadmap. We identify specific challenges for each tracer individually and make recommendations for next steps towards clinical implementation. prognosis, or therapy response. Finally, clinical utility, i.e., whether the test improves patient outcome and is cost-effective, is determined by health-related measurements. Successful progress through these tracks is essential for a test to pass from analytical to clinical research stage, and subsequently to routine clinical practice [1••].

Development Stages of [ 18 F]-FDG-PET/CT
Technical Validity [ 18 F]-FDG-PET/computed tomography (CT) can detect increased glucose metabolism in cancer cells and is indicated for multiple oncological indications [2,3]. [ 18 F]-FDG is phosphorylated by the enzyme hexokinase and trapped inside (tumor) cells [4]. The reproducibility and repeatability of [ 18 F]-FDG-PET/CT were assessed for various cancer types (see Table 1 for overview) [58]. One meta-analysis of 5 studies, including 102 cancer patients of which 6 had metastatic BC (MBC), assessed the repeatability of [ 18 F]-FDG-PET(/ CT) by measuring the standardized uptake value (SUV) max/ mean in the same patient on two separate occasions with an interval of 1-4 days [5]. A high test-retest interclass correlation coefficient (ICC) of 0.90 and 0.91 was found for SUV max and SUV mean , respectively. Reproducibility across different scanners was assessed in 23 patients, 17 with BC [13]. Patients underwent two [ 18 F]-FDG-PET/CT scans within 15 days on the same scanner or on different scanners at different sites. Cross-calibration of PET/CT scanners and dose calibrator was performed. The average difference in SUV max between test-retest [ 18 F]-FDG-PET/CT, using the same scanner, was 8% versus 18% on different scanners. International standardization efforts to improve reproducibility resulted in the European Association of Nuclear Medicine (EANM) guideline for 18 F imaging procedures, followed in 2010 by the Research Ltd. (EARL) accreditation program to assure independent quality control, comparable scanner performance, and reproducible assessments [3,59]. Since 2010, the number of accredited centers has increased over time in Europe and beyond [60].

Clinical Validity
For [ 18 F]-FDG-PET/CT, we focused on clinical validity studies with at least 100 BC patients. A meta-analysis of 13 studies (see Table 1) reported incidental and unexpected breast uptake detected by [ 18 F]-FDG-PET(/CT) [23]. Overlap between SUVs in malignant and benign breast incidentalomas was found, and not all lesions were further histologically examined. Therefore, [ 18 F]-FDG-PET/CT is not routinely used for diagnosis of primary BC. With regard to diagnosis of axillary lymph node metastases in BC, a meta-analysis was performed of studies comparing [ 18 F]-FDG-PET(/CT) to the reference standard: axillary lymph node dissection (ALND) or sentinel lymph node biopsy (SLNB) [25]. In 7 out of 26 studies involving 862 BC patients, [ 18 F]-FDG-PET/CT sensitivity was 56% and specificity 96%, compared to 52% and 95% for ALND and/or SLNB [25]. Another meta-analysis (21 studies including 1887 BC patients), using ALND and/or SLNB as reference standard, showed a sensitivity and specificity of 64% and 93%, respectively, for detection of axillary lymph node metastases by [ [61]. With regard to [ 18 F]-FDG-PET/CT for diagnosis of recurrent or distant metastases in BC, two meta-analyses including a total of 2500 patients (2 studies with overlapping subjects) showed both high sensitivity (92-96%) and specificity (82-95%) [28,29]. For the detection of bone metastases, [ 18 F]-FDG-PET/CT showed a sensitivity and specificity of 93% and 99%, versus 81% and 96% respectively, for conventional bone scintigraphy, as determined in a meta-analysis involving 668 BC patients in 7 studies [30]. According to the EANM, ESMO, and NCCN guidelines, [ 18 F]-FDG-PET/ CT should be considered in cases of suspected recurrence or equivocal findings on standard imaging and can be used for staging in high-risk BC patients [2, 3, 62, 63••, 64, 65••].
Despite the non-specific uptake of [ 18 F]-FDG, preoperative [ 18 F]-FDG uptake, expressed as SUV max , was found to be related to prognostic pathological characteristics assessed on core biopsy in primary BC. SUV max was higher in ER− than ER+ tumors (7.6 versus 5.5); higher uptake was also observed in triple-negative tumors, tumor grade 3, ductal carcinoma, . Lower baseline SUV max predicted more favorable survival outcomes than higher SUV max (analyzed as a continuous variable) [34]. The lack of clear cutoff values has so far precluded the use of [ 18 F]-FDG-PET as a prognostic tool in BC. This is partly due to the fact that SUV calculations can depend on the PET camera systems used. To harmonize the acquisition protocols and the quantification process between different camera systems, the EARL harmonization program was introduced.
Clinical validity of serial [ 18 F]-FDG-PET/CT to monitor therapy response to neoadjuvant treatment was analyzed in two meta-analyses (see Table 1), showing a pooled sensitivity of 82-86% and specificity of 72-79%, using histopathology as reference standard for pathological (non-)response [35, 36•]. Possibly differences between the pace of disease response between BC subtypes may play a role in this setting. In the randomized neoadjuvant study AVATAXHER in 142 patients with HER2+ BC, [ 18 F]-FDG-PET/CT at baseline and after 1 cycle of docetaxel/trastuzumab was used for further treatment decisions [37]. Patients with a ΔSUV max of ≥ 70% (n = 69) continued docetaxel/trastuzumab. Patients with a ΔSUV max of < 70% (n = 73) were randomized for continued docetaxel/trastuzumab or addition of bevacizumab. In all patients receiving docetaxel/trastuzumab, this ΔSUV max cutoff of 70% showed a positive and negative predictive value of 53% and 75%, respectively, to detect pathological complete response. Recently, preliminary data from the neoadjuvant PREDIX HER2 trial showed that pathological response was related to decreased uptake on early [ 18 F]-FDG-PET/CT compared to baseline, in HER2+ primary BC [66]. For MBC, no well-designed large study to assess the clinical value of [ 18 F]-FDG-PET/CT has been performed, only small studies with varying endpoints [67,68]. The optimal cutoff value and interval between [ 18 F]-FDG-PET/CT scans for response measurement in BC are still unknown and may limit implementation of [

Technical Validity
Bone is the most common site of metastasis in BC. Two PET tracers ([ 18 F]-FDG and [ 18 F]-NaF) are included in EANM and NCCN guidelines to identify bone metastases in BC patients. [ 18 F]-NaF, approved by the FDA in 1972, reflects enhanced bone metabolism due to bone metastases but also due to degeneration, arthritis, or fractures [71,72]. The repeatability of [ 18 F]-NaF-PET/CT was evaluated in a prospective multicenter study by Lin et al. in 35 prostate cancer patients with bone metastases who underwent two pretreatment [ 18 F]-NaF-PET/CT scans (test-retest interval 3 ± 2 days), with SUV mean as most re pe a t ab le e n dp oin t (o ve rv ie w: Tab l e 1 ) [ 1 8 ]. Repeatability of SUV mean/max , functional tumor volume (FTV 50% ), and total lesion [ 18 F]-fluoride uptake (TLF) measured with [ 18 F]-NaF-PET/CT was confirmed by Wassberg et al. [19]. Moreover, a high inter-observer agreement at the patient level was found by using three scales to define [ 18 F]-NaF-PET/CT findings [21]. How to correctly perform and interpret [ 18 F]-NaF-PET/CT scans is published in EANM and Society of Nuclear Medicine and Molecular Imaging (SNMMI) guidelines, supporting technical standardization and harmonization [73,74].

Clinical Validity
At present, no comparison has been performed of [ 18 F]-NaF-PET/CT with a bone biopsy as the gold standard for the entire study population, but it has been compared with other imaging modalities. [ 18 F]-NaF-PET/CT has a higher sensitivity to detect bone metastases than either [ 18 F]-FDG-PET/CT or conventional bone scintigraphy with 99m Tc-labeled diphosphonates (planar and SPECT) (97-100% versus 74% versus 91%, respectively). However, although the specificity of [ 18 F]-NaF-PET/CT was higher than that of bone scintigraphy, it was slightly lower than [ 18 F]-FDG-PET/CT (71-85% versus 63% and 97%, respectively) [39, 40•]. In general, a negative [ 18 F]-NaF-PET/CT can be used to exclude bone metastases, but in case of positive findings, [ 18 F]-NaF-PET/CT should be carefully interpreted and correlated with CT findings. With regard to the prognostic value of [ 18 F]-NaF-PET/ CT, one prospective study was performed in 28 BC patients with bone-dominant disease, showing no correlation between baseline SUV max and skeletal-related events, time-to-progression or overall survival (OS) [41]. However, ΔSUV max of 5 lesions between baseline and 4 months of systemic treatment was associated with OS [41]. With regard to the predictive value of [ 18 F]-NaF-PET/CT, two small studies showed that lack of endocrine treatment efficacy was related to an increase in metabolic flux to mineral bone or SUV max in BC patients with bone only disease (see Table 1) [42,44]. The national prospective oncologic PET registry of the USA showed that [ 18 F]-NaF-PET/CT altered the treatment plan in 39% of BC patients [75]. However, the impact of [ 18 F]-NaF-PET/CT for therapy response on clinical decision-making remains unclear due to varying endpoints and experimental procedures.

Clinical Utility
The cost-effectiveness of [ 18 F]-NaF-PET(/CT) to detect bone metastases was assessed in a meta-analysis of 11 trials, including 425 patients (7 BC patients) [76]. It was concluded that the average cost-effective ratio was less favorable for [ 18 F]-NaF-PET(/CT) than for conventional bone scintigraphy.

Conclusions and Recommendations of [ 18 F] -NaF-PET/CT
While the technical validation of [ 18 F]-NaF-PET/CT is completed, clinical validation with comparison to a biopsy as reference standard is still warranted. Also, clinical validity of [ 18 F]-NaF-PET/CT should be further assessed with uniform endpoints. Therefore, [ 18 F]-NaF-PET/CT has not yet passed through the necessary steps towards routine clinical practice according to the imaging biomarker roadmap. Although in bone-trope cancers such as BC, an optimal tool for diagnosis and treatment evaluation is still needed and it is unclear whether this tool could be [ 18 F]-NaF-PET/CT.

Clinical Validity
A meta-analysis of 9 studies (all prospective, except one) involving 238 patients reported a pooled sensitivity of 82% and specificity of 95% to detect ER+ tumor lesions by quantitative assessment of [ 18 F]-FES uptake (overview: Table 1) [45]. A similar sensitivity and specificity was found in direct comparison of [ 18 F]-FES uptake and ER expression on biopsy (in 5 studies including 158 BC patients) [45]. Recently, a large prospective cohort study was published involving 90 BC patients with first recurrence/ metastatic disease, comparing the correlation between qualitative [ 18 F]-FES-PET/CT results and immunohistochemistry (IHC) of ER status of the same metastatic lesion. This resulted in a positive and negative predictive value of 100% and 78%, respectively [22••]. A quantitative analysis was also performed, showing a positive and negative agreement of [ 18 F]-FES-PET/CT (threshold SUV max 1.5) with ER IHC equaling 85% and 79%, respectively. Despite the importance of this well-defined prospective vertebrae, costae, pelvis, and proximal femora. The increased uptake in the joint was related to degeneration cohort trial, its impact is likely limited due to exclusion of bone metastases, the most common metastatic site in ER+ MBC. Furthermore, an optimal SUV max cutoff to distinguish benign from malignant lesions by [ 18 F]-FES-PET/ CT has not been established. Although SUV max 1.5 is most commonly used for this distinction, ranges of 1.0 to 2.0 have also been described. Yang et al. determined an ROC curve in 46 ER+ BC patients, showing an optimal SUV max cutoff of 1.8, with a sensitivity of 88% and specificity of 88% (optimal SUV mean cutoff: 1.2) [79]. The study of Nienhuis et al. in 91 ER+ MBC patients found that physiological background uptake could exceed SUV max 1.5, for example, in the lumbar spine [80]. [ 18 F]-FES-PET/CT scans performed in 108 individuals showed that irradiation could induce atypical (non-malignant) enhanced [ 18 F]-FES uptake in the lungs [81]. These issues should be taken into account in interpreting [ 18 F]-FES-PET/CT scans for the diagnosis of BC. However, these data are retrospective and should be interpreted with caution. Nonetheless, two trials have indicated usefulness of [ 18 F]-FES-PET(/CT) for the physician by improving diagnostic understanding compared to conventional assessments in 88% of patients, and causing a treatment change in 48-49% of patients enrolled in the studies [82,83]. Therefore, [ 18 F]-FES-PET/CT may be a useful diagnostic tool in exceptional diagnostic dilemmas when added to a conventional workup. A prospective study involving 90 ER+ BC patients treated with endocrine therapy found that [ 18 F]-FES-PET(/CT) may be a useful prognostic biomarker for [ 18 F]-FDG avid tumors, demonstrating a higher median progression-free survival (PFS) in the high [ 18 F]-FES uptake group compared to low [ 18 F]-FES uptake group (7.9 versus 3.3 months, respectively) [48]. With regard to response prediction, a meta-analysis including 6 prospective trials and 183 patients found a pooled sensitivity of 64% and specificity of 29% to predict early or late response to hormonal therapy, with an SUV max cutoff of 1.5, and a sensitivity of 67% and specificity of 62% with SUV max of 2.0 [45]. In 26 patients with primary ER+ BC, randomized to neoadjuvant chemotherapy or endocrine treatment, no differences in baseline SUV max were found between post-treatment pathological (non-) responders [49•]. In another small trial (including 18 patients), pathological response to neoadjuvant chemotherapy was related to low rather than high baseline SUV max (1.8 versus 4.4) [84]. Overall, it is difficult to compare this data due to the heterogeneity of the trials, i.e., different endpoints, and imaging procedures.

Clinical Utility
Two computer simulation studies described the impact of [ 18 F]-FES-PET/CT on health-related measurements, such as life years gained (LYG), ICER, and total costs (Table 1) [55,56]. One study selected first-line treatment in MBC patients based on biopsy results or [ 18 F]-FES-PET/CT imaging findings and showed higher diagnostic and treatment costs in the PET/CT imaging group [56]. A second study determined the number of avoided biopsies to assess MBC after the introduction of [ 18 F]-FES-PET/CT and showed that the number of biopsies (39 ± 9%) was lower in the [ 18 F]-FES-PET/CT imaging group [55].

Conclusions and Recommendations of [ 18 F] -FES-PET/CT
While [ 18 F]-FES-PET/CT is currently used in a limited number of hospitals worldwide, mostly in a research setting, but also as a diagnostic tool in exceptional diagnostic dilemmas, consistent data to support its clinical validity and utility are still lacking. Only in France is [ 18 F]-FES approved for routine clinical use to determine ER status in MBC. In order to implement [ 18 F]-FES-PET/CT more broadly in routine clinical practice, additional studies are needed. Within two prospective cohort trials, the multicenter IMPACT breast trial and the ECOG-ACRIN trial (NCT02398773; 99 newly diagnosed MBC patients), the analysis of baseline [ 18 F]-FES uptake related to treatment response or PFS is ongoing. In the ongoing ET-FES TRANSCAN trial (EUDRACT 2013-000-287-29), the treatment choice is based on [ 18 F]-FES-PET/CT (high versus low 18 F-FES uptake) [85]. [ 18 F]-FES-PET/CT is also added as integrated biomarker to another randomized controlled trial, the SONImage trial (NCT04125277). With these additional studies, sufficient evidence could potentially be generated to support implementation of [ 18 F]-FES-PET/CT in routine clinical practice.

Technical Validity
The [ 89 Zr]-labeled antibody trastuzumab binds to the HER2receptor and has a relatively long half-life (t ½ = 78 h). This enables imaging at late time points but also limits repeatability testing as radiation dose is high and repeated scans would require a 2-week interval [86]. To optimize the acquisition protocol, imaging at multiple time points (after 1-7 days) was performed after a single tracer injection [87,88]. The optimal time point was found after 4-5 days, due to lower background uptake and higher contrast. No comparison of [ 89 Zr]-trastuzumab-PET/CT with biopsy has been performed so far. In a prospective study including 34 HER2+ and 16 HER2− BC patients, an SUV max cutoff of 3.2 showed a sensitivity of 76% and specificity of 62% to distinguish HER2+ from HER2− lesions [53]. The HER2 status was based on the primary tumor or metastatic lesion; however, a recent biopsy of a tumor lesion was not performed in all patients. Despite this relatively low discriminative value, [ 89 Zr]-trastuzumab-PET/CT did support diagnostic understanding and resulted in a treatment change in 90% and 40% of patients respectively, in whom HER2 status could not be determined by standard workup [91]. With regard to the prognostic value of [ 89 Zr]-trastuzumab-PET/CT no data are available, but its value to predict therapy response was assessed in the ZEPHIR trial (see Table 1) [54]. In 56 HER2+ MBC patients, qualitative analysis of baseline PET/CT scans indicated that [ 89 Zr]-trastuzumab uptake was related to longer trastuzumab emtansine treatment duration, compared to no uptake (11.2 versus 3.5 months) [54].

Clinical Utility
A computer simulated study of a hypothetical cohort of 1000 MBC patients assessed whether [ 89 Zr]-trastuzumab-PET/CT could replace biopsy [56]. This study concluded that total costs were higher with [ 89 Zr]-trastuzumab-PET/CT. However, biopsy effects on quality of life were not included in the analysis.

Conclusions and Recommendations of [ 89 Zr] -Trastuzumab-PET/CT
Although technical standardization and harmonization is supported by the recently introduced [ 89 Zr]-PET/CT EARL accreditation program, at present, still significant knowledge gaps exist (for instance regarding the relation between biopsy and uptake on [ 89 Zr]-trastuzumab-PET/CT) [89]. Therefore, multiple steps according to the imaging biomarker roadmap have to be taken before [ 89 Zr]-trastuzumab-PET/CT can be implemented in clinical practice. It is expected that the previously mentioned multicenter IMPACT breast study will provide information that can advance the validation of [ 89 Zr]trastuzumab-PET/CT.

Other PET Tracers for Molecular Imaging in BC
Multiple new tracers of potential interest in BC can be identified (see Table 2). PET imaging of additional receptors may be the next step, for example, the hormone receptor tracer [ 18 F]-dihydrotesterone ([ 18 F]-FDHT)-PET, which is commonly used in prostate cancer trials. This tracer provides information about androgen receptor (AR) expression, which is a potential new target for BC treatment [46]. Moreover, cell proliferation can be detected by [ [48,54,94]. This could help to select the best therapeutic strategy.

Conclusions
In this review, we identified hurdles based on the biomarker roadmap for the four most commonly used PET tracers in BC and made recommendations for the next steps towards clinical implementation. This review has summarized several important steps to be considered to successfully implement molecular biomarkers for BC patients in clinical practice. In general, support for clinical utility is still pending for PET tracers in BC, but also assessment of clinical validity is hampered by varying endpoints and procedures. Improving trial designs can contribute to solve this matter; for instance, multicenter trials require standardization and harmonization of procedures. International collaboration is essential, as this would also potentially allow building warehouses of data to overcome a plethora of small solitary single center studies. Based on these warehouses, clinical validation can be established in line with the RECIST guidelines. In this setting, considering all aspects of the biomarker roadmap at an early stage is important. Smart trial designs adding imaging biomarkers to randomized controlled trials (integrated biomarker) are desirable, as imaging biomarker-based randomized controlled trials (integral biomarker) are usually not feasible due to the large numbers of patients required [95]. From a regulatory point of view, the evidence required for implementation is still Human and Animal Rights and Informed Consent This article does not contain any studies with human or animal subjects performed by any of the authors.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.