Early outcome prediction on 18F-fluorocholine PET/CT in metastatic castration-resistant prostate cancer patients treated with abiraterone

Objective: We investigated the role of 18F-fluorocholine positron emission tomography/computed tomography (FCH-PET/CT) in the early evaluation of abiraterone and outcome prediction in patients with metastatic castration-resistant prostate cancer (CRPC). Patient and methods: Forty-three patients with metastatic CRPC progressing after docetaxel received abiraterone 1,000 mg daily with prednisone 5 mg twice daily. Patients were evaluated monthly for serological PSA response and safety. FCH-PET/CT was done at baseline and after 3 to 6 weeks. Univariate and multivariate Cox regression models addressed potential predictors of progression-free survival (PFS) and overall survival (OS). Results: Declines in PSA level of ≥50% were seen in 21 of 43 (49%) patients. Forty-two patients were evaluable for FCH-PET/CT response. FCH-PET/CT bone flare was observed in 4 of 42 (10%) evaluable patients. In univariate analysis, PSA decline and FCH-PET/CT response predicted PFS, while PSA decline and FCH-PET/CT (progression vs non progression) predicted OS. In multivariate analysis, only FCH-PET/CT (progression vs nonprogression) remained significant for PFS and OS (p = 0.022 and p = 0.027, respectively). Conclusion: Early FCH-PET/CT can predict clinical outcome in CRPC beyond PSA response. These data support further studies on FCH-PET/CT for abiraterone monitoring and outcome prediction in patients with CRPC.


INTRODUCTION
A well-recognized barrier to therapy management in castration-resistant prostate cancer (CRPC) is the lack of reliable surrogate markers of response to treatment coupled with a potentially exaggerated reliance on changes in serum prostate-specific antigen (PSA) as an indicator of treatment efficacy. This is particularly true with agents that have the potential for PSA modulation such as abiraterone, a selective inhibitor of androgen biosynthesis that potently and irreversibly blocks cytochrome P-450c17 (CYP17), which improved survival in metastatic CRPC [1][2][3]. Consequently, changes in radiological images have been advocated as important markers of disease response and progression to abiraterone [4,5]. www.impactjournals.com/oncotarget Altered choline metabolism is a characteristic of prostate cancer, and 18F-fluorocholine Positron emission tomography (PET) -computed tomography (CT) (FCH-PET/CT) is one of the most promising tools for prostate cancer imaging [6]. FCH-PET/CT has been shown to have a higher sensitivity and specificity than standard 99m Tc-MDP bone scans and 18F-fluorodeoxyglucose (FDG)-PET/CT for the detection of bone metastases from CRPC [7][8][9]. On the basis of these characteristics, this metabolic imaging might be considered for further evaluation for guiding prostate cancer management [10][11][12].
We investigated the role of FCH-PET/CT in the early treatment evaluation and outcome prediction in patients with CRPC treated with abiraterone.

Relationship between PSA decline and therapy response at FCH-PET/CT
Declines in PSA level of 50% or more were observed in 21 of 43 (49%) patients. The CT response at 3 months of follow-up was assessed in 42 patients, while one case with renal dysfunction did not perform it. 3-month CT results were: complete response (n = 1, 2%), partial response (n = 1, 2%), stable disease (n = 28, 67%), PD (n = 12, 29%). Forty-two patients were evaluable for FCH-PET/CT response after a median time between baseline and follow-up FCH-PET/CT scanning of 40 days (range, 19 to 69 days). One of 43 patients did not perform the follow-up FCH-PET/CT due to rapid PD with clinical deterioration and decline in performance status. FCH-PET/CT scanning showed 3 CR, 11 PR, 6 SD, and 22 PD, but in 6 of 12 pts with initial PD at FCH-PET/CT a bone flare phenomenon was suspected due to a stable disease at 3-month CT scan associated with a progressive decline of PSA. In one case suspected for bone flare, the PET scan was not repeated due to severe intestinal infection with treatment interruption and following clinical deterioration. In another patient who had a 3-month PSA decline of 60% with SD at 3-month CT scan, the followup FCH-PET/CT scanning after 11 weeks confirmed PD with rapidly increasing PSA levels. Thus, overall, bone scan flare, as defined by the combination of PSA decline, initial flare, and subsequent improvement or stability in the following FCH-PET/CT scanning was observed in 4 of 42 (10%) evaluable patients and 4 of 24 (17%) responsive cases. The clinical characteristics of these 4 cases were not significantly different from the study population as a whole, median age was 71 years (range, 57 to 78), median PSA level at baseline was 15.5 ng/mL (range, 1.5-92.2), the maximal decline of PSA was 91%, 73%, 64% and 37%, respectively. The latter case who had a very low baseline PSA level of 1.5 ng/mL and PSA decline to 0.9 ng/mL (37%, lower than the threshold of 50%) was also considered to be of interest and is reported as bone flare. A PSA flare, defined as an initial PSA increment followed by a decline below the PSA baseline level, was reported in one case of bone flare, but also in another patient without bone flare. Of 4 patients with FCH-PET/CT bone flare, 3 had PD after 9, 11 and 13 months, respectively, while one is continuously progression free after 22+ months. A CR was achieved at follow-up FCH-PET/CT in 3 of the 26 (12%) evaluable patients, and was associated with undetectable PSA levels (< 0.03 ng/mL) in 1 case after 5 months on abiraterone. Of 3 patients with nodal disease only, one obtained a CR at first FCH-PET/CT and 2 a PR with a mean SUVmax reduction of 55% and 70%, respectively. Figure 1 to 3 show examples of early FCH-PET/CT effects of abiraterone, including an example of bone flare.
Changes in PSA levels were concordant with the FCH PET/CT response/non-response in 30 of 42 patients (71%) ( Table 2), and with the FCH-PET/CT progression/ non-progression in 34 of 42 (81%) ( Table 3).  Univariate and multivariate analyses were performed to assess the association between factors of interest and PFS and OS. In univariate analysis, PSA decline, FCH-PET/CT (SD/PD vs PR/CR) and FCH-PET/CT (progression vs non progression) predicted PFS, while PSA decline and FCH-PET/CT (progression vs non progression) predicted OS (Table 4). In multivariate analysis, only FCH-PET/CT (progression vs nonprogression) remained significant for PFS and OS (p = 0.022 and p = 0.027, respectively) ( Table 4).

DISCUSSION
To the best of our knowledge, this is the first study that used FCH-PET/CT for outcome prediction and tumor monitoring in CRPC. The correlation between the prognostic value of PSA and FCH-PET/CT response to therapy suggests that PSA levels are linked to a tumor metabolic activity. Furthermore, in multivariate analysis FCH-PET/CT (progression vs nonprogression) was superior to PSA with respect to PFS and OS prediction, supporting the hypothesis of an essential role of FCH-PET/CT to improve therapy management.
PSA decline ≥50% concurred with the FCH-PET/ CT response/non-response in 71% of patients (Table 2), and with the FCH-PET/CT progression/non-progression in 22 out of 26 81% cases (Table 3). Moreover, in univariate analysis, FCH-PET/CT (SD/PD vs PR/CR) and FCH-PET/CT (progression vs non progression) predicted PFS, whereas in multivariate analysis only FCH-PET/CT (progression vs nonprogression) remained significant for PFS and OS (p = 0.022 and p = 0.027, respectively) ( Table 4). However, a main limitation of this study is represented by the criteria used for FCH-PET/CT classification of response, based on fairly arbitrary thresholds for change in SUV (as specified in EORTC guidelines). In this regard the differentiation between responders and non-responders results much less reliable than the FCH-PET/CT progression/non-progression evaluation.
PET/CT progression/non-progression proved capable of predicting outcome in patients with other tumors, including bone metastases [13,14]. As abiraterone in CRPC is given continuously until progression, it is essential to predict and monitor progression rather than the response to treatment. PET/CT criteria for monitoring tumor progression are simpler and more widely accepted than those for monitoring tumor response. The relationship of PSA kinetics to radiological PFS has been recently evaluated in an exploratory analysis of patients treated with abiraterone in a randomized phase III study [15]. Radiological PFS was positively associated with the magnitude of PSA decline, and the effect remained after correcting for covariates. However, in this trial the PSA level was monitored every 2-3 months, whereas the kinetics of serum PSA analyzed monthly during the first few months of abiraterone is associated with PSA flare in nearly 10% of these patients, which could impair the early evaluation of PSA response [16], The impact of FCH-PET/CT could be even more relevant for CRPC with no reliable serum biomarker, as is the case of bulky metastatic disease with low serum PSA levels. However, in a few of these cases, elevated circulating Chromogranin A levels have been reported as possible expression of a CRPC with neuroendocrine differentiation in which the role of FCH-PET/CT is not known [17,18]. Thus, the integrated use of PSA and FCH-PET/CT could potentially lead to a significant improvement in outcome prediction and in the therapeutic monitoring of CRPC patients.
Clinical studies have demonstrated that PET can predict response to chemotherapy and targeted therapy in cases of several tumors [14,[19][20][21]. The early identification of therapeutic response supports the clinician in the decision making. Moreover, new functional imaging techniques such as whole-body diffusion weighted magnetic resonance imaging (MRI) can improve the evaluation of therapy response in patients with metastatic bone disease with a potential significant impact in the therapy assessments of patients with prostate cancer because of the high prevalence of bone-only disease [23,24].
Controversy still exists on the applicability of PET/CT and other new imaging tools in clinical practice, mainly because of its expense. However apart from additional costs related to PET/CT examinations, potential savings associated with PET/CT as a result of avoiding additional imaging examinations and by helping clinicians make the optimum treatment decisions avoiding that multiple doses of high-cost antineoplastic agents are administered, even if they are ineffective. Thus, the potential implementation in clinical practice of FCH-PET/ CT as well as predictor of clinical outcome needs further evaluation and validation with larger patient population including cost-benefit analyses. Additionally alternative imagines such as whole-body diffusion weighted magnetic MRI should be also evaluated in the same contest and for the same purpose.
An important observation that emerges from our study is the evidence of the mismatch highlighted in 4 patients between the reduction of the value of serum PSA and the detection of an increased uptake of bone lesions shown in the PET/CT after 1-2 months. This mismatch was evaluated as related to the phenomena of "bone flare", already reported in patients after treatment with abiraterone in chemotherapy-naive CRPC who had undergone bone scintigraphy, but that stands out for the first time also with FCH-PET/CT [25]. FCH-PET/ CT bone flare predicted a favorable long-term outcome similar to that of FCH-PET/CT non progressive patients. The PCWG2 consensus criteria considered bone flare and included the recommendation to perform a first follow-up bone scan 12 or more weeks after the start of treatment, and furthermore defines progression in bone when a minimum of 2 new lesions are observed and confirmed on a second scan done 6 or more weeks later [5].
The metabolic tumor changes as detected with FCH-PET/CT make possible new ways of managing and investigating CRPC. However, the early FCH-PET/ CT bone flare phenomenon should be considered as a limitation ofits utility as an early indicator marker ,or thefirst follow-up FCH-PET/CT should be carried out 3 months after the start of therapy. Moreover, FCH uptake may be influenced by several factors, including the number of viable cells per unit volume, the tumor vascular supply, the sites of metastases and the neuroendocrine differentiation [17,18,26,27]. Recently, new PET tracers such as radiolabeled dihydrotestosterone and radiolabeled antibody to prostate-specific membrane antigen (PSMA) have shown great potential for assessing patients' early biologic response to hormonal therapies in prostate cancer [28][29][30].
In conclusion, in this pilot study early FCH-PET/ CT was able to predict clinical outcome in CRPC beyond PSA response. Further larger studies are needed to better define the most sensitive and cost-effective modality for assessment of therapy efficacy in CRPC.

Patients
Forty-three consecutive patients with advanced prostate cancer, without neuroendocrine differentiation, after radically prostatectomy and in progression on  Response was defined as complete response or partial remission; nonresponse was defined as stable disease or progressive disease. b Including 2 patients with bone flare at FCH PET/CT with SUV decline at second-follow-up PET >25%, and 1 patient with partial remission at PET but liver progression at 3-month CT only. androgen deprivation, were enrolled. Patients previously received docetaxel and not more than two chemotherapy regimens for metastatic CRPC. Prior ketoconazole therapy was not permitted. Additional eligibility criteria included adequate performance status, cardiac, renal, hepatic and bone marrow function, serum potassium level 3.5 mmol/L or more, and ongoing androgen deprivation with serum testosterone <50 ng/dL. The protocol was approved by the Institutional Review Board. Written informed consent was obtained from all patients.

Treatment and evaluation
Treatment consisted of abiraterone acetate 1,000 mg daily with prednisone 5 mg twice daily in 28-day cycles. Treatment was given continuously until there was evidence of progressive disease (PD) or unacceptable toxicity. PD was defined according to Prostate Cancer Working Group 2 (PCWG2) criteria [5]. Baseline examinations included PSA, CT and FCH-PET/CT scanning. Patients were evaluated monthly for serological PSA response and safety. FCH-PET/CT was repeated after 3 to 6 weeks. A CT scan was done after 3 months of abiraterone therapy.
Adverse events were graded using the National Cancer Institute Common Terminology Criteria for Adverse Events (CTCAE), version 4.

PET/CT Imaging Protocol
FCH-PET/CT scans were performed on an integrated PET/CT system (Discovery Discovery LS camera, General Electric Medical Systems, Waukesha, WI) in 2D acquisition mode for 3 minutes per bed position. PET scanning was performed 45 minutes after intravenous injection of 18F-Methilcholine (3,7 MBq/Kg of body weight, AAA-Advanced Application Accellerator, Meldola, Italy). The field of view included the skull to mid-femurs. Low dose CT (120 kV, 80 mA) without contrast agents was performed for attenuation correction and as an anatomical map. The emission data were corrected for scatter, random coincidence events, and system dead time using the provided software.

Image Reading and Interpretation
PET/CT scans were evaluated by two nuclear medicine physicians with consolidate experience. In case of disagreement between the readers, studies were reexamined and a consensus was reached. Criteria to define PET/CT positivity included the presence of focal areas of increased tracer uptake with or without any underlying lesion identified using CT. Semiquantitative criteria based on the maximum standardized uptake value (SUVmax) and the target to background ratio were used to aid the visual analysis [31]. Images were read sequentially by using a Xeleris III workstation (General Electric Medical Systems). PET, CT, and PET/CT fused images were used to correctly evaluate the scans in various cuts (axial, sagittal, and coronal)

Measurement of treatment effect by FCH-PET/CT
Baseline FCH-PET/CT was performed in all patients and a minimum of 2 target lesions were defined as a references. Patients scanned 3-6 weeks after abiraterone therapy first cycle, were considered responders according to guidelines for PET on the basis of a metabolic activity decrease more than 25% of target lesions depicted at baseline images, coupled with or without evidence of change/decrease in size [32,33]. A complete disappearance of FCH uptake was considered indicative of a metabolic complete response (CR), whereas less than 100% but more than 25% of FCH uptake was considered a metabolic partial remission (PR). Non-responders were the patients with similar or higher FCH uptake in the target lesions, and/or with a substantial increase of the target lesions size (ie, 20% increase in sums of longest diameters) [33]. Among nonresponders, increased FCH uptake in a target lesion that was also enlarged on CT imaging (ie, 20% increase in the longest diameter) and/or a new metabolically active lesion were considered indicative of progressive disease (PD), whereas a substantially unchanged FCH uptake without substantial increase in size was considered stable disease (SD).

FCH-PET/CT bone flare
FCH-PET/CT bone flare was defined as the combination of increased >25% uptake of known lesions at early FCH-PET/CT after 3-6 weeks of therapy, which occurred in the context of a ≥50% decline in PSA and no evidence of PD at 3-month CT scan, with FCH-PET/ CT repeated 2-3 months later showing improvement or stability. An initial PET scan flare followed by continued declines in PSA levels below the threshold of 50% and stable uptake at the third FCH PET-TC was also considered to be of clinical interest.

Statistical analysis
PFS was defined as the time elapsed between the date of the start of abiraterone and the date of PD or death or the date of the last follow-up visit [5]. OS was defined as the elapsed time between the date of the start of abiraterone and the date of either death or the last follow-up. Kaplan-Meier survival plots were generated on the basis of FCH-PET/CT response, and the curves were compared using log-rank test. Cox proportional hazards regression was used to determine univariate and multivariate hazard ratios for selected potential predictors of PFS and OS. A p value of < 0.05 was considered statistically significant.