Utility of ¹⁸F-fluoroestradiol (¹⁸F-FES) PET/CT imaging as a pharmacodynamic marker in patients with refractory estrogen receptor-positive solid tumors receiving Z-endoxifen therapy

Abstract

Background

Z-endoxifen is the most potent of the metabolites of tamoxifen, and has the potential to be more effective than tamoxifen because it bypasses potential drug resistance mechanisms attributable to patient variability in the expression of the hepatic microsomal enzyme CYP2D6. ¹⁸F-FES is a positron emission tomography (PET) imaging agent which selectively binds to estrogen receptor alpha (ER-α) and has been used for non-invasive in vivo assessment of ER activity in tumors. This study utilizes ¹⁸F-FES PET imaging as a pharmacodynamic biomarker in patients with ER+ tumors treated with Z-endoxifen.

Methods

Fifteen patients were recruited from a parent therapeutic trial of Z-endoxifen and underwent imaging with ¹⁸FFES PET at baseline. Eight had positive lesions on the baseline scan and underwent follow-up imaging with ¹⁸F-FES 1– 5 days post administration of Z-endoxifen.

Results

Statistically significant changes (p = 0.0078) in standard uptake value (SUV)-Max were observed between the baseline and follow-up scans as early as 1 day post drug administration.

Conclusion

F-FES PET imaging could serve as a pharmacodynamic biomarker for patients treated with ER-directed therapy.

Introduction

Advances in our understanding of the tumor biology of estrogen receptor-positive (ER+) malignancies have led to the development of a wide array of targeted therapies for patients with ER+ disease, and the ER is a target in malignancies like breast cancer, for which agents such as selective estrogen receptor degraders (SERDs) are under development. A major therapeutic option in ER+ tumors is hormone therapy directed at interrupting the estrogen signaling that stimulates growth [1]. Tamoxifen is among the most commonly used drugs for treating ER+ malignancies. However, it must first be converted into active metabolites, mediated by the enzyme CYP2D6, to exert its pharmacologic effect. Significant inter-patient variability has been observed in the efficacy of tamoxifen, with recurrence rates of 30–50 % in patients receiving tamoxifen in the adjuvant setting [2]. Responsiveness to tamoxifen depends in part on the genetic variability in patient drug-metabolizing enzyme levels, particularly CYP2D6. Administration of endoxifen, the most potent tamoxifen metabolite, would circumvent the need for metabolic activation, potentially reducing or eliminating the variability in response rates due to differences in CYP2D6 levels. Z-endoxifen, the hydrochloride salt and the Z(cis)stereoisomer of endoxifen, has significant potential for antineoplastic activity and is being investigated in several clinical trials (ClinicalTrials.gov identifiers: NCT02311933, NCT01273168). First-in-human studies of Z-endoxifen published in 2010 demonstrated that Zendoxifen administered orally was rapidly absorbed by the body, with a T-max of approximately 6 h [3].

The level of ER expression has been shown to provide important prognostic information and also to predict the likelihood of a response to conventional hormone therapy [4, 5]. 16 alpha-[¹⁸F]fluoro-17 beta-estradiol (¹⁸F-FES) has been proposed as a non-invasive in vivo marker of tumor ER activity. There are two naturally occurring subtypes of the ER, α and β, which are encoded by the ESR1 and ESR2 genes, respectively [6]. While the role of ER-β is not fully understood and has not yet been investigated in clinical trials of breast cancer treatment [7, 8], ER-α has been studied extensively as a binding target for both therapeutic and diagnostic agents [9, 10]. Furthermore, ER-α has been the subject of several studies examining the use of ER-targeting imaging agents to provide pharmacodynamic information on ER-targeted agents [11–13]. Similar approaches of using imaging studies as pharmacodynamic markers for therapy has also been reported in other agents and models, such as the use of ¹¹C-choline in prostate cancer [14] and ⁸⁹Zr-transferrin for transcription inhibitors [15].

¹⁸F-FES is an investigational radiolabeled positron emission tomography (PET) imaging agent that selectively binds to ER-α, with an affinity for ER-α which is 6.3 times as high as that for ER-β [16]. Several studies have shown the reliability of ¹⁸F-FES in detecting ER+ lesions in humans, with uptake corresponding to the degree of immunohistochemical (IHC) staining for ER on tumor biopsies [4, 11, 17, 18]. Studies conducted in patients with locally advanced, metastatic, and pretreated breast cancer have demonstrated a direct relationship between response to endocrine therapy, primarily tamoxifen and aromatase inhibitors, and ¹⁸F-FES uptake on PET imaging [19]. Furthermore, prior research reported that baseline ¹⁸F-FES uptake corresponded with subsequent tumor response 6 months after hormone treatment [20].

Currently, invasive procedures including tissue biopsy are required to determine the degree to which Z-endoxifen might suppress ER activity in vivo. ¹⁸F-FES PET imaging could be useful for non-invasive detection of the response of refractory ER+ solid tumors to Z-endoxifen, and has in fact been shown to have added value in tumors such as uterine sarcoma and leiomyoma [21]. The objective of this study was to determine whether a very early (1–5 days post therapy) pharmacodynamic effect could be detected with ¹⁸F-FES PET after Z-endoxifen therapy.

Material and methods

Patient population

The imaging study was an optional component of a prospective National Cancer Institute Institutional Review Board (IRB)-approved study (CTEP study no. 8826, ClinicalTrials.gov ID: NCT01273168) of the therapeutic drug Z-endoxifen. Written informed consent was obtained from all patients, and all data were acquired and stored in compliance with the Health Insurance Portability and Accountability Act (HIPAA). The results of the therapeutic trial will be reported separately.

The ¹⁸F-FES PET imaging protocol was conducted in patients participating in a phase I treatment trial with Z-endoxifen, whose inclusion criteria included patients with refractory ER+ breast cancer, desmoid tumors, gynecologic cancers, or other presumed ER+ solid tumors. Patients with breast cancer had to have had at least one prior chemotherapy regimen and one prior hormone regimen for metastatic disease, and all other patients had disease that had progressed following at least one line of standard therapy. No major surgery, radiation, hormone therapy, or chemotherapy was permitted within 4 weeks prior to study enrollment. From December 2011 to November 2014, 15 of 29 eligible patients enrolled in the phase I treatment trial were recruited to the ¹⁸F-FES PET imaging study. All 15 patients completed a baseline ¹⁸F-FES PET/CT scan, and eight of the 15 were brought back for follow-up ¹⁸F-FES PET/CTscans 1 to 5 days after Z-endoxifen administration because the initial scan was positive for ¹⁸F-FES uptake.

Radiotracer

The ¹⁸F-FES (16 alpha-[¹⁸F]-fluoro-17 beta-estradiol) used in this study was prepared by the Leidos Biomedical Research, Inc. radiopharmacy in the Frederick National Laboratory for Cancer Research located in Frederick, Maryland, and was supplied as a sterile IV-injectable solution in a volume of ≤ 30 mL containing 0.15 M phosphate-buffered saline [22, 23]. The specific activity was 3500 ± 1500 Ci/mM at the end of synthesis, with radiochemical purity of > 95 %. The radiotracer was administered to patients by intravenous injection of 0.07 mCi/kg (2.6 Mbq/kg) of ¹⁸F-FES, with a maximum prescribed dose of 6.0 mCi (222 Mbq).

Z-endoxifen and clinical follow-up

Z-endoxifen HCl (Z-4-hydroxy-N-desmethyl-tamoxifen hydrochloride) was administered orally to patients once daily for 28 days per cycle. Due to the phase I dose-escalation nature of the parent therapy trial, patients in the PET imaging study received varied doses of Z-endoxifen ranging from 40 to 300 mg.

Imaging protocol

Patients underwent imaging from the base of the skull to the mid-thigh approximately 60 min after ¹⁸F-FES IV injection over 1 min. In some patients, the injection time was drawn out over several minutes due to transient pain from the injected solution. PET/CT images were acquired using a Philips Gemini Time-of-Flight (TOF) scanner (Philips Medical Systems, Cleveland, OH, USA) in 3D acquisition mode, with 2-min acquisition time per bed position. Images were reconstructed using a TOF LOR-based fully 3D iterative reconstruction using a blob-biased function [24]. The iterative 3D reconstruction was performed using two iterations and 33 subsets, with no post filtering. The scans were reconstructed into 4 × 4 × 4 mm voxel images. All PET data were reconstructed with and without attenuation correction using low-dose, unenhanced computed tomography (CT), which was also used for anatomical co-registration. The CT scan parameters were 1.438 pitch, 0.75 s rotation, 120 keV xray tube voltage, and 60 mAs tube current. The images were reconstructed into a 70-cm-diameter field of view (FOV), 512 × 512 image. Blood samples were collected within 24 h of each ¹⁸F-FES-PET/CT scan session to check for levels of serum estradiol and sex hormone-binding globulin (SHBG) to rule out inhibition of the ¹⁸F-FES by competition or carrier binding.

Image analysis

Images were reviewed and analyzed by two board-certified nuclear medicine physicians using commercial software (MIM, version 6.3.9). Volumes of interest (VOIs) were manually drawn over lesions by an experienced nuclear medicine physician using the PET images with the corresponding non-contrast CT serving as a guide, and contours were checked for concurrence by a second experienced nuclear medicine physician. The non-contrast CT component of the PET/CT scan and other imaging modalities (magnetic resonance imaging [MRI], ultrasound, ¹⁸F-FDG PET), when available, were used to better identify and localize tumors. Each visualized lesions was assessed qualitatively, with positive lesions being defined as visually distinct compared to background, and semi-quantitatively, using the standard uptake value (SUV). Histopathological information, when available, was used as the reference standard for lesion positivity/negativity. However, as this information was not available for many of the visualized lesions, serial imaging of all available modalities was used as the “imaging consensus” reference standard in these cases. For each tumor, SUV-Max (defined as the mean of the hottest 20 % of pixels in the VOI), SUV-Mean, and tumor volume (TV) of the manually drawn VOIs were calculated. Baseline SUV parameters and TVs, as well as percentage change in the parameters from baseline to post treatment, were computed.

Statistical analysis

Clinical and tumor lesion data were collected from all patients. Pre- and post-treatment percentage change in SUV uptake and TV were calculated for SUV-Max, SUV-Mean, and TV. For patients with multiple lesions, the mean and maximum percentage change across multiple lesions was taken to represent each patient’s average and worst outcomes, respectively. A subanalysis of bony lesions was conducted. The significance of per-patient percentage change in SUV uptake was evaluated by one-sample Wilcoxon tests. The difference in percentage change in SUV between patients with and without bone lesions was compared by the two-sample Wilcoxon test. The correlation between dose level and per-patient percentage change in SUV uptake was measured by the Spearman rank correlation.

Results

A total of 15 patients (13 females and 2 males) underwent ¹⁸F-FES scans, with an average age of 59.2 ± 13.9 years (range 22–75). Cancers included breast (n = 4, with one male breast cancer), gynecological malignancies (two endometrial, four ovarian, one cervical, one fallopian tube), and desmoid tumors (n = 3). Eight of 15 patients completed baseline and posttreatment follow-up ¹⁸F-FES PET imaging. Patients who did not have at least one positive lesion on the baseline scan were not asked to return for the follow-up study. Table 1 presents the complete demographic information on the enrolled patients.

Table 1.

Patient demographics with scan and treatment dose information

Patient no.	Age (years)	Sex	Tumor type	Baseline 18F-FES dose (mCi)	F/U scan time (days)	F/U 18F-FES dose (mCi)	Z-endoxifen dose (mg q.d.)
1	56	F	Endometrial adenocarcinoma	3.70	3	5.90	40
2	63	F	Invasive ductal carcinoma	5.77	5	5.87	40
3	46	F	Metastatic breast cancer	5.90	5	6.30	60
4	73	F	Serous papillary ovarian cancer	6.60	N/A		60
5	75	F	Ovarian adenocarcinoma	5.80	N/A		100
6	56	F	Cervical cancer	6.00	2	5.80	140
7	42	F	Desmoid tumor of shoulder	6.00	N/A		140
8	66	F	Serous ovarian carcinoma	6.00	2	6.00	140
9	61	F	Serous carcinoma of the fallopian tube	4.30	3	5.50	200
10	68	F	Poorly differentiated granulosa tumor of the ovary	3.90	N/A		200
11	62	M	Invasive ductal carcinoma of the male breast	4.50	N/A		200
12	67	F	Endometrial carcinoma	5.90	N/A		300
13	59	M	Desmoid tumor of abdominal mesentery	5.90	2	5.90	300
14	72	F	Infiltrating ductal carcinoma of breast	6.00	2	6.00	280
15	22	F	Desmoid tumor of shoulder axilla	5.90	N/A		280

The mean injected dose of ¹⁸F-FES was 5.5 ± 0.9 mCi (203 ± 33 Mbq), with a range of 3.7 to 6.6 mCi (137– 244 Mbq), for the baseline study, and 5.9 ± 0.2 mCi (219 ± 8 Mbq), with a range of 5.5 to 6.3 mCi (204– 233 Mbq) for the follow-up scan.

Seven patients underwent scanning only at baseline, with a total of 21 distinct non-FES-avid lesions identified using other available imaging sources including diagnostic CT or MRI. Of these 21 lesions, six were located in the bone, five in the chest, one in the arm, one in the liver, and eight elsewhere within the abdomen and pelvis. The lesion in the liver had an SUV of 9.3, but since the background liver SUV was 10.1 due to physiologic tracer metabolism, the liver lesion was considered not to be evaluable. Not counting this liver lesion, the average SUVMax of all lesions in this group was 2.0 ± 0.7 (range 0.8 to 3.7), while the average SUV-Max of the bone lesions was 2.0 ± 0.4 (range 1.3–2.3). The average TV in mm³ was 44.6 ± 151.6 (range 0.7–700.1).

For the eight patients who completed both pre- and posttherapy ¹⁸F-FES PET scans, 41 distinct lesions were identified, among which 19 were located in the bone, six within the chest, one in the breast, three in the liver, and 12 elsewhere within the abdomen and pelvis. Liver lesions were considered not evaluable for this study due to the high ¹⁸F-FES uptake in the background liver parenchyma. The number of lesions found per patient ranged from 1 to 11. On the baseline scan, the average SUV-Max among all lesions was 4.8 ± 3.4 (range 0.7–14.1), while on the follow-up scan, the average SUV-Max was 2.5 ± 2.3 (range 0.7–9.9). The average percentage change from baseline among all lesions was −38.9 % ± 31.7 % (range −86.1 to +29.9 %), while the average absolute change in SUVMax was −2.3 ± 1.1 (range −10.9 to +2.0). The average TV in mm3 at baseline and follow-up and the average percentage change from baseline to follow-up were 27.5 ± 62.4 (range 0.8–293.0), 29.8 ± 70.7 (range 0.4–376.2), and 11.1 ± 63.5 % (range −77.4 to +236.2 %), respectively. For bone lesions, the average SUV-Max on the baseline and follow-up scans and the average percentage change were 5.3 ± 3.2 (range 1.0–12.6), 1.6 ± 0.4 (range 0.9–2.7), and −58.9 ± 22.2 % (range −4.0 to −86.1 %). For TV in bone lesions, the average SUV-Max on the baseline and follow-up scans and the average percentage change were 16.8 ± 14.7 (range 2.2–50.7), 17.5 ± 16.0 (range 2.3–58.9), and 11.2 ± 63.5 % (range −49.0 to +202.4 %), respectively. These results are summarized in Table 2.

Table 2.

Correlation of treatment dose with change in SUV and tumor volume

Patient no.	Dose (mg)	Mean % change in SUV-Mean	Mean % change in SUV-Max	Mean % change in tumor volume
1	40	−17.00 %	−33.19 %	−20.30 %
2	40	−16.21 %	−15.23 %	24.85 %
3	60	−75.81 %	−76.41 %	43.30 %
6	140	−2.09 %	−13.19 %	6.20 %
8	140	−19.49 %	−20.97 %	8.62 %
9	200	−35.29 %	−32.01 %	48.65 %
13	300	40.24 %	−31.91 %	−6.91 %
14	280	−34.68 %	−46.19 %	−10.51 %

On a per-patient level, the mean percentage change in SUV-Max, SUV-Mean, and TV from baseline to follow-up scan was −33.6, −20.0, and 11.7 %, respectively, with the change in SUV-Max reaching statistical significance (P = 0.0078). In patients with more than one lesion, the maximum percentage change (representing the worst disease in the patient) for SUV-Max, SUV-Mean, and TV was 8.7, −4.9, and 82.2 %, respectively. These results are summarized in Table 3.

Table 3.

Per-patient mean and maximum percentage change of SUV and tumor volume after treatment

	Mean % change (per-patient average outcome)			Maximum % change (per-patient worst outcome)
	Mean	SD	P value	Mean	SD	P value
SUV-Mean	−20.04	0.33	0.1484	8.7	0.4	0.4609
SUV-Max	−33.59	0.2	0.0078	−4.88	0.3	0.8438
Tumor volume	11.69	0.25	0.3828	82.15	0.91	0.0391

The relationship of SUV-Max, SUV-Mean, and TV in patients with primarily bone lesions was also explored. Patients with bone lesions exhibited a trend toward higher mean percentage reduction in SUV-Max than patients who did not have bone disease (see Fig. 1), but the difference was not statistically significant (p = 0.0714). Because of the phase I dose-escalation nature of the parent therapeutic trial, patients in this imaging study received varied doses of the therapeutic drug Zendoxifen. Among the eight patients who had both baseline and post-therapy scans, the doses of Z-endoxifen administered were 40 mg (n = 2), 60 mg (n = 1), 140mg (n = 2), 200 mg (n = 1), 280mg (n = 1), and 300 mg (n = 1). The SUVresponse did not correlate with the dose level of Z-endoxifen (see Fig. 2). However, this may be related to the small sample size.

Fig. 1.

¹⁸F-Fluoroestradiol (¹⁸FFES) PET images before and 3 days after treatment with Zendoxifen administered orally in a breast cancer patient with bone metastasis. Focal tracer uptake in ER+ bone lesions is seen before treatment, but is blocked on the subsequent scan. Physiologic activity is seen in liver, GI tract, bladder, and vascular collateral vessels from the left arm injection

Fig. 2.

Dose–response analysis of changes seen on the FES PET scan vs. Z-endoxifen dose. Panels A–C show the percentage change in the SUVMax, SUV-Mean, and tumor volume, respectively, in all eight patients who received both a baseline and follow-up FES PET scan. Panels D–F show the same data but only for patients with breast cancer (three patients with 21 lesions), and panels G–I are patients with other gynecological cancers (four patients with 17 lesions). One patient with a desmoid tumor (with one lesion) is not plotted separately

The maximum serum estradiol was 74.0 pg/mL on the baseline scans and 20.9 pg/mL on the follow-up scans, which is within the normal physiologic range, and should not have adversely affected the ¹⁸F-FES PET scans. The highest SHBG levels at baseline and follow-up were 186 nmol/L (in a patient with a desmoid tumor) and 69 nmol/L, respectively, which is also not expected to have significantly affected the scans.

Of the four patients with breast cancers, all but one had focally positive ¹⁸F-FES tumors. One male patient with breast cancer had a negative ¹⁸F-FES scan of his bone metastasis. One breast cancer patient had bone metastases that were both ER+ and ER− (patient 14). Among the eight patients with gynecologic malignancies, four were ¹⁸F-FES-negative (three ovarian, one endometrial cancer), and two had both ER+ and ER− tumors (cervical and endometrial cancers). The three patients with desmoid tumors had very low ¹⁸F-FES uptake.

Discussion

¹⁸F-FES imaging was helpful in identifying ER+ tumors that could be targeted by anti-estrogen agents. As a metabolite of tamoxifen, endoxifen may improve response to treatment by circumventing possible genetic resistance mechanisms related to inherent differences in tamoxifen metabolism. This study revealed little association between ¹⁸F-FES uptake and Zendoxifen treatment. However, Z-endoxifen induced statistically significant changes in SUV that were demonstrated as early as 24 h post administration of Z-endoxifen (see Figs. 3 and 4). While this effect has been studied previously with tamoxifen [5, 13], fulvestrant [13, 25], and aromatase inhibitors [13], to the best of our knowledge, this is the first study to demonstrate this effect with Z-endoxifen. The ability to examine and quantitate ER expression of lesions at a whole-body level is important, as this can be used as an indicator of a drug’s ability to block tumor ER. In our study, we found that the average change in FES uptake after Z-endoxifen administration was a 33.6 % decline on a per-patient level, which is comparable to numbers reported in the literature. For instance, the Mortimer study published in 2001 showed a mean percentage decrease in SUV-Max of 54.8 % for responders vs. 19.4 % for non-responders on tamoxifen therapy [5], and the Linden study published in 2011 demonstrated a similar average decline of 54 % with tamoxifen and fulvestrant therapy [13].

Fig. 3.

Axial ¹⁸F-FES PET and PET/CT fusion images of a patient with cervical cancer at baseline and after three doses of Z-endoxifen, demonstrating heterogeneous ER status. Top images show a peri-rectal mass with focal uptake at baseline and significantly decreased activity at follow-up. Physiologic bowel activity is seen in the upper-right images at follow-up. The lower images show known lung metastases without significant uptake at baseline and follow-up suggesting ER expression

Fig. 4.

Axial ¹⁸F-FES PET and PET/CT fusion images of a patient with serous carcinoma of the fallopian tube at baseline and following three doses of Zendoxifen. Arrows point to the right iliopsoas mass with ER+ expression at baseline and blockade of receptors after initiating therapy

In the literature, ¹⁸F-FES activity has been shown to predict response to endocrine therapy with selective ER modulators in breast cancer [17, 26]. In examining specific tumors, bone disease appeared to respond earlier or faster than soft tissue lesions. This suggests underlying differences in biology or tumor environment that may need to be taken into account in evaluating treatment. In breast cancer patients, most lesions detected by ¹⁸F-FES are bone lesions, but heterogeneity of ER expression can also be a factor in the varied response to therapy [27]. For instance, we found three patients exhibiting apparent mixed ER expression among their known metastases as determined by the ¹⁸F-FES PET scan.

Only patients with distinctly positive focal ¹⁸F-FES uptake in tumors on baseline imaging had follow-up imaging after initial Z-endoxifen dosing. Patients without focal ¹⁸F-FES-positive tumors were considered not to have functional ER, and thus repeat imaging after treatment with an ER blocker would not be expected to change scan results. Furthermore, if ¹⁸F-FES PET imaging can be shown to reliably represent the functional state of ER expression in tumors, a completely negative ¹⁸F-FES PET scan may be sufficient justification to spare patients the potential toxicity and side effects of futile ER-targeted treatment.

There is limited experience with ¹⁸F-FES in gynecologic malignancies [21, 28, 29]. While an SUV-Max of greater than 1.5 is sometimes considered the threshold of positivity for ER expression in breast cancer, the average SUV of the negative lesions (by qualitative analysis) in our study was 1.95. One possible explanation for this slightly higher observed lesion SUV could be that many of the lesions seen in this study were close to areas of high physiologic ¹⁸F-FES activity, and that some of this surrounding activity was included in the lesion VOIs. Another explanation is that our study included many tumor types in addition to breast cancer, and that different malignancies may have different levels of ER expression. For instance, a recent study in epithelial ovarian cancer found an SUV-Max of > 1.8 to be the optimal threshold for discriminating between ER+ and ER− lesions [28]. In a recent study, ¹⁸F-FES PET demonstrated positive predictive value (PPV) of 60 % and negative predictive value (NPV) of 80 % for predicting response to estradiol therapy [30]. Although some trends were seen with regard to the predictive value of ¹⁸FFES for Z-endoxifen therapy, our study cohort was too small to draw statistically valid conclusions. Furthermore, some malignancies, such as desmoid tumors, may preferentially exhibit more ER-β expression, for which ¹⁸F-FES has a much lower binding affinity (see Fig. 5)[31–33].

Fig. 5.

Axial ¹⁸F-FES PET and PET/CT fusion images of a desmoid tumor in the left subpectoral region without significant radiotracer uptake at baseline

Serum estradiol levels were within physiologic ranges and did not appear to affect ¹⁸F-FES uptake. Levels of SHBG were also measured, but even at the highest level, found in a patient with a desmoid tumor, the possible reduction in SUV still would likely not have changed the overall negative uptake status. Peterson et al. did not observe an effect with ¹⁸F-FES SUV and estradiol, but did note an inverse relationship with SHBG and SUV uptake, and suggested measuring levels in patients [34].

The TVs used in this study were obtained from VOIs that were manually drawn on the PET images in conjunction with the corresponding non-contrast CT. Due to the intrinsic resolution limitations of PET, it is likely that the TVs contain both inaccuracies and challenges in reproducibility that prevent us from drawing meaningful conclusions regarding the anatomic volumes of the lesions. However, just as some value has been found in the metabolic tumor volume (MTV) and total lesion glycolysis (TGL) measurements with ¹⁸F-FDG PET scans [35, 36], there may be some value in providing the total ¹⁸FFES uptake from this study in a volume format, and this should be investigated in future studies.

There are some limitations to this study. A larger number of patients would afford greater statistical power in determining the relationship between Z-endoxifen and ¹⁸F-FES. The decision to include all ER+ tumor types was made because the FES PET scan should, from a mechanistic point of view, be able to visualize all tumors that over-express ER. However, the wide variety in tumor histology also made it challenging to determine definitive histological associations with ¹⁸F-FES uptake, although ours is the first study to describe ¹⁸F-FES imaging in desmoid tumors.

Conclusion

In this study, we demonstrated the usefulness of ¹⁸F-FES as a pharmacodynamic marker for Z-endoxifen and perhaps for other novel ER-directed therapies. Future studies could evaluate the use of ¹⁸F-FES PET imaging both as a predictive marker for ER-based therapies and for treatment response.