Can serial FCH-PET / CT scans predict the response to treatments in castrate resistant prostate cancer ?

20 Abstract. To evaluate the use of serial F-choline (FCH) PET/CT scans in order to monitor the response to chemotherapy in patients (pts) with metastatic castrate-resistant prostate cancer (mCRPCa). Materials and methods. We collected FCH-PET/CT data from a consecutive series of 21 mCRPCa pre-treated pts (median age:73years) who met the inclusion criteria: such as visceral and non-visceral metastasis with clear uptake on a baseline FCH-PET/CT scan and no fewer than two scans separated by at least one month and by no more than 12 months whilst receiving chemotherapy. Eighteen pts were treated by taxanes and 3 with abiraterone. Change in FCH uptake (SUVmax of the index lesion; ΔSUVmax) during the course of treatment was compared with the clinical assessment response and PSA change (ΔPSA). The correlations were evaluated by using non-parametric tests, as appropriate. Results. The median interval time between PET/CT scans was 5 months (interquartile rangeIQR: 3-7). Median (IQR) PSA value and SUVmax were 25(7.22-141) ng/mL, 15.7(11.3-17.83), 22.9(7.7-92.5) ng/mL and 5.5(7.21-19.1) ng/mL respectively at the time of baseline and 2 PET/CT. The median ΔPSA and ΔSUVmax were -49.9% and 8.3%, respectively. No correlation between ΔPSA and ΔSUVmax was found (r=0.046; p=0.843). In accordance with the clinical assessment, 8 pts were responders, 4 had a stable and 9 a progressive disease. A significant overlap was reported for ΔPSA and clinical categories of response (p=0.685). Conversely, ΔSUVmax was significantly different in pts with a response to chemotherapy as compared to those with stable and progressive disease (p=0.005). Conclusions. ΔSUVmax, rather than ΔPSA, represents an important measure of response to chemotherapy and should be useful to stratify the prognosis of mCRPCa pts. A large multicenter prospective study is necessary to confirm these findings.


INTRODUCTION
Prostate cancer is the most frequent tumor in males and both the availability of prostate specific antigen (PSA) testing and the extension of median life expectancy are responsible for a significant increase of incidence registered in the last decades.
Endocrine therapies (either luteinizing hormone-releasing hormone analogues, anti-androgens or combined) represent the first-choice for patients found with advanced disease at diagnosis or relapsing after local treatments, with an average time to progression (TTP) of 18-24 months to the so-called "Castrate-resistant" phase when diseases progresses notwithstanding low levels of circulating testosterone [1]. Approximately 10-20% of patients with localized prostate cancer developed metastatic castrateresistant prostate cancer (mCRPCa) within 5-years of follow-up after initial therapy, while the majority of patients with metastatic disease at diagnosis become castrate-resistant within a shorter interval time [2]. More than 80% of mCRCPa patients have bone metastases, with or without the coexistence of nodal or visceral metastases [3].

Evangelista et al.
The process by which prostate cancer cells become castrate resistant is unclear, but it is characterized by overexpression or hyperactivation of the androgen receptor despite castrate levels of androgen [4].
Docetaxel has long been the only treatment able to improve survival of mCRPCa and still represents the first cytotoxic option. Following docetaxel, patients have currently now at least three further options with the ability to impact on survival: namely cabazitaxel, abiterone acetate and enzalutamide. However, to date there are no clear guidelines indicating which treatment should be prescribed first or which patients would respond best to the different treatment options. Finally, 223-radium has been a proven treatment for mCRPCa, with excellent palliative effects as well as increases in overall survival [5,6].
In the clinical practice setting, there is a strong need to assess the response to chemotherapy for several reasons: (1) to determine the real efficacy of treatments;(2) to decide to stop or to continue the ongoing therapeutic scheme and (3) to assess the benefit of treatments on the disease course (how the disease behaves over time).
However, PSA is not always the most reliable method to understand the response to systemic therapy and therefore alternatives are required. A study by Al-Saeedi et al. [7] suggested the hypothesis that choline uptake could be changed before and after therapy and consequently, it may be indicative of tumor response. Since serum PSA alone is not a reliable marker for quantifying response to therapy, radiolabelled choline positron emission tomography (PET)/computed tomography (CT) could be proposed as an alternative imaging tool for the assessment of response to therapy.
While the European Organization for Research and Treatment of Cancer Criteria based on tumor glucose metabolism variations using 18 F-fluorodoxyglucose PET/CT are well standardized, the role of alternative PET compounds, such as radiolabelled choline (either 11 C and 18 F-choline), in the assessment and prediction of responses to cancer treatment has not yet been thoroughly investigated in large series, but in a small cohort of patients during abiraterone and enzalutamide treatments [8,10].
There is a lack in the current literature on this topic. Therefore, we performed a retrospective study on mCRPCa patients followed at our Institution in order to evaluate the usefulness of serial 18 F-fluoromethylcholine (FCH) PET/CT scans to monitor the response to chemotherapy. Moreover, as secondary end-point, we assessed the correlation between standardized uptake value (SUV) changes and the final outcome of the patients.

Patient population
A retrospective analysis was performed on medical records of mCRPCa patients undergoing systemic treatments and serial FCH-PET/CT between March 2011 and May 2014. Inclusion criteria for this study were: (1) patients with histological diagnosis of prostate cancer, any Gleason, any type of local treatment; (2) mCRPCa disease with bone and/or visceral metastases showing a pathologic uptake on a baseline FCH-PET/CT scan which is confirmed by biopsy or conventional imaging (such as bone scan, magnetic resonance imaging (MRI), CT and radiography); (3) no fewer than two FCH-PET/CT scans separated by at least one month and by no more than 12 months whilst receiving therapy; (4) treatment with systemic anticancer agents for mCRPCa, first or subsequent line; 4) no therapy at the time of baseline FCH-PET/CT; 4) no granulocyte colonystimulating factor for at least two weeks before FCH-PET/CT. Twenty subjects fulfilled the inclusion criteria. All patients provided written informed consent for access to their medical records and use of aggregated clinical data for scientific purposes before their initial FCH-PET/CT examination according to our Institutional guidelines. The research was performed in accordance with the Declaration of Helsinki.

FCH PET/CT examination
Prior to PET/CT examination with FCH, the patient was kept fasting for at least 6 hours without any restrictions for water intake. FCH (provided by IASON GmbH Graz, Austria), was administered intravenously with a scheduled activity of 3 MBq/kg body mass administered intravenously. To prevent incomplete administration and subcutaneous accumulation and to ensure good hydration, the radiopharmaceutical agent was injected in an infusion line connected to saline. A PET/CT camera (Biograph 16 enriched with high definition software, by Siemens) was used for PET imaging, with low dose CT acquired first. PET/CT whole-body acquisition was commenced 60 minutes after the FCH injection, covering a field of view from the skull to the knees. No dynamic PET was used. The PET data were reconstructed over a 128 x 128 matrix with 4.75 mm pixel size and 2 mm slice thickness. Processed images were displayed in coronal, transverse, and sagittal planes. PET/CT images were also assessed quantitatively using the maximum and average standardized uptake value (SUVmax and SUVavg, respectively) that was computed according to the following formula: K (SUV) = K (Bq/cc) x [Weight (Kg)/Dose (Bq)] x 1000 cc/Kg (where: K (Bq/cc) = calibrated and scaled Pixel volume and Dose (Bq) = injected dose in Bq at injection time with decay corrected).
All PET images were read by two trained nuclear medicine specialists, independently. At visual analysis, increased FCH uptake on the basis of either highly suspicious or definite CT morphologic changes and not corresponding to physiological uptake patterns were recorded as positive for metastases. On the contrary, the absence of uptake out of normal physiological sites, was used to define a negative PET/CT finding. SUVmax and SUVavg at the site of the most prominent FCH uptake was prospectively recorded on the baseline scan, and this site was designated as the index lesion. The SUVs of the index lesion served as the primary means for evaluating the change in FCH uptake over serial scans. The SUVs of the index lesion were recovered in each site (bone, lymph node, prostate gland or prostatic fossae, lung etc.).

Treatment
Patients were treated according to both normal practice of care and the National and International guidelines [11,12] with anticancer drugs approved for mCRPCa. Treatment was administered until there was progression or unacceptable toxicity.

Assessment of response
Change in FCH uptake during the course of treatment was compared with clinical assessment of response (as reference). Clinical information included serial measures of PSA values, conventional images (such as CT, bone scan and MRI) and subjective symptoms changes (pain and other disease-related symptoms). Response to treatment was categorized as response, stable disease, or progression according to the Prostate Cancer Clinical Trials Working Group 2 criteria [13]. Criteria for classification into the categories included size and number of metastatic sites at conventional imaging, change in tumor marker and symptom changes. Due to the retrospective nature of this study, there were no standard time intervals between PET scans, but most patients were expected to undergo a second PET/CT within 4 months from the first assessment. To compare FCH PET/CT results with clinical parameters, the response to therapy was first evaluated separately by two board-certified medical oncologists specialized in the treatment of prostate cancer and one board-certified nuclear medicine physician, and then a comparison made.

Clinical outcomes
Progression was defined as either an increase in PSA value(s) (>50%), and/or symptomatic progression, and/or progression determined by an imaging modality other than FCH PET/CT (CT, MRI, bone scan, radiography). Due to the relatively low number of patients, we did not assess separately the type of progression (biochemical, clinical or radiological).
Subsequently, as with the decline in PSA, we divided the patients as metabolic responders and non-metabolic responders: if the decline in SUVmax were >50% or <50%, respectively. Moreover, TTP and overall survival (OS) were assessed and defined as the time from the start of treatment to the evidence of progressive disease and allcauses of mortality, respectively. Death was confirmed by review of death certificate, hospital chart or physician's records.

Statistical analysis
Data were expressed as median and interquartile range(s) (IQR). The change in FCH uptake for each treatment interval was measured as the change in SUVmax and SUVavg of the index lesions. Continuous variables were compared by using the Kolmogorov Smirnov Z test and the Kruskal-Wallis test for K independent variables, as appropriate. Linear regression analysis was used to test the relation between ΔPSA and ΔPSA. The association between the change in SUV of the index lesion and the subjective category of clinical response (regression, stable disease, or progression) were assessed by using the Kendall rank correlation coefficient (Tau_b). K statistics were used to determine the level of concordance of clinical, PSA and PET/CT responses. K value of >0.4, >0.6, and >0.8 indicate fair, good and excellent agreement, respectively. A two-side analysis was used and p<0.05 value was considered statistically significant (SPSS Inc., Advanced Models 20.0, Chicago, Illinois).

RESULTS
There were 21 eligible patients who underwent 55 PET/CT scans, (8 had only 2 scans while 13 had more than two scans). In Table 1 are reported the characteristics of the study population. The median age was 73 years (range: 58-84 years). The majority of patients were treated by radical prostatectomy and lymph node dissection (n=9; 42.9%), followed or not by adjuvant radiotherapy. The median time from the commencement of the LH-RH analogue and mCRPCa phase was 39.5 months (IQR: 27.25-53 months).
In 18 of the subjects, the Gleason score was more often ≥7 (n=15; 83.3%). The interval between PET scans ranged from 3 to 11 months, with an average of 5 months between scans. Of 21 patients, 9 (42.9%) had scans 5 or fewer months apart, and the residual 12 (57.1%) had scans 6 or more months apart. Among patients with shorter intervals between FCH PET/CT scans compared to those with longer intervals there was no difference in FCH uptake at baseline (p=0.153), nor in percentage change in SUVmax (p=0.822). Figure 1 and Figure 2 show two representative examples of FCH-PET/CT images.    In accordance with the clinical assessment, 8 (38%) patients were responders, 4 (19%) had a stable and 9 (43%) a progressive disease. The concordance between clinical response (response vs. no response) and PSA (<50% vs. >50%) and between clinical response (response vs. no response) and SUVmax (<50% vs. >50%) were found in 12/21 (57%) and in 16/21 (76%), respectively (K value: 0.087±SE 0.209; p=0.676 and 0.493±SE 0.187; p=0.018). The median changes in PSA level were 22.3%, -68.5% and -49.9%, respectively for partial response, stable and progressive disease (p=0.685). Conversely, the median changes of SUVmax in the index lesion were significantly different among three categories (52.8%, -10.4% and -12.2%, respectively; p<0.001) ( Table 2). Moreover, also the changes in SUVmax of bone and lymph node index lesions were significantly different among the subsets of patients. As illustrated in Figure 3, a significant overlap was reported for ΔPSA and clinical categories of response. On the other hand, ΔSUVmax was significantly different in patients with a partial response to chemotherapy as compared to those with stable and progressive disease (p=0.005), although stable and progressive disease patients showed a similar ΔSUVmax (p=0.940).
Three out of 21 patients treated with abiraterone acetate showed a partial (n=1), stable (n=1) and progressive response (n=1). Conversely, 7 (38.9%) patients who underwent taxane-based treatment showed a partial response, 3 (16.7%) had a stable disease and 8 (44.4%) had a progression of disease. In these latter subsets of patients, ΔSUVmax resulted more accurately than ΔPSA in defining the clinical response to therapy (p=0.012 and p=0.420, respectively; see Figure 5, Table 1 and Figure 1).

Outcome measures
Median time to progression disease time was 12 months from the time of the start of treatment (range 5-33 months). Eleven patients (52.4%) experienced clinical progression. No differences among ΔSUVmax, ΔPSA and long-term progressive disease were found (all p>0.05). After 16 months (IQR 12-24) from the start of treatment and 108 months (IQR 78-150) from the initial diagnosis, 3 patients died. As illustrated in Table 3, both the changes in PSA and in SUVmax were lower in progressive and dead patients as compared to the counterpart, although not statistically significant. Figure 4 illustrates the distribution of ΔPSA and ΔSUVmax in accordance with the progression and survival of patients. Finally, we found that 100% vs. 67%, and 55% vs. 82% of patients with a decrease of SUV<50% vs. PSA<50%, respectively, showed both a low survival and low TTP period.

DISCUSSION
Over the past 15 years, substantial progress has been achieved in the treatment of patients with mCRPCa with chemotherapy. Primary androgen deprivation therapy is the standard of care, usually with a long-lasting release formulation of LHRH analogue or antagonist [14]. Throughout these years, many patients developed a mCRPCa which progressed despite castrate level of testosterone [15]. Metastatic CRPCa consists in a heterogeneous population of patients. Table 2. Correlation between change in biochemical-semiquantitative data and response to therapy The values are expressed as median (minimum; maximum); LN= lymph node; *available in 6 patients; available in 5 patients; †available in 2 patients; † †available in 3 patients.
Evangelista et al. The values are expressed as median (minimum; maximum); LN: lymph node;*available in 5 patients;**available in 9 patients; †available in 8 patients; † †available in 12 patients Various prognostic factors have been suggested, including general prognostic factors, such as pain, ECOG score, Gleason score or biological information, which still needs to be confirmed in larger trials [1]. PSA response -defined as a decrease in serum PSA of >50% for at least 4 weekshas been proposed as a surrogate marker for radiological and clinical progression [16]. A serum PSA is much simpler to obtain than traditional measurements of treatment outcome and can also be obtained more frequently. It should be noted, however, that PSA responses have not been validated as surrogates for OS in advanced prostate cancer [17]. In the present preliminary study, we aimed to demonstrate that serial FCH-PET/CT correlate with therapeutic response as assessed by PSA value, symptoms and conventional imaging. In our retrospective cohort, the change in SUVmax was more accurate in differentiating a partial response from stable and progressive disease as compared to a change in PSA during or after chemotherapy. During cytotoxic treatments, the decline of PSA reflects a significant cell kill in response to agents that cause a reduction in the overall tumor burden. Therefore, in the case of an absent decline of PSA during this type of treatment, the drug is considered inactive. In an Editorial published in 2008, Collette [21] reported that the noncytotoxic drugs could induce changes in PSA that may not be reflected by similar changes in the tumor. Therefore, similarly to PSA, the evaluation of response by 18 F or 11 Ccholine PET/CT could be different in patients who are treated by cytotoxic or non-cytotoxic drugs. Herein, we found that, during cytotoxic treatment, ΔSUVmax is more accurate than ΔPSA in distinguishing patients with a partial response to therapy from those with stable or progressive disease. In fact, the concordance between clinical response to therapy was good for the absolute change of SUVmax (K=0.493; p=0.018), but it resulted fair (K=0.087; p=0.209) or the change of PSA lower of upper 50%.
Moreover, we confirmed the same results also in patients who were treated by taxane-based chemotherapy alone (see Figure 5).
Pantaleo et al [8] examined a clinical case of a patient with metastatic prostate cancer being treated with taxotere and zolendronic acid who underwent 11 C-choline PET/CT in the pre and post-therapy setting. The examinations indicated a partial remission of disease which was in accordance with the decreased in serum PSA level, while a total bone scan showed stable disease. The authors stated that the correlation between metabolic activity and serum PSA levels suggests that 11 C-choline PET/CT may be more predictive of treatment response than a bone scan in case of bone metastases because it offer more specific data about the metabolic activity of disease and not only about the bone remodeling.
The findings of this study demonstrate the prognostic potential of the changes in SUVs during treatment for the prediction of OS, although both the values were not statistically significant. However, 100% of patients with a decrease in SUVmax>50% resulted in still living still after 16 months from the start of treatment. Giovacchini et al . [22] suggested that 11 C-Choline PET/CT might be a suitable technique to identify patients with different prognosis for potential stratification for salvage lymphadenectomy or radiotherapy in oncologic trials. Conversely, Breeuwsma et al. [23] did not find any significant differences in prostate cancer-specific survival on the basis of 11 C-choline PET/CT results. In our opinion, and by sharing the results by De Giorgi et al. [9,10], radiolabelled choline PET, either 11 C or 18 F, might also be useful for monitoring the response to new therapies (such as taxanes or new anti-hormonal agents).Therefore, providing an appropriate support for future clinical trials.    This study's limitations are typical for a retrospective study.
Treatment regimens, duration of treatment intervals, intervals between serial FCH-PET and the use of conventional imaging and tumor markers all varied widely in this patient cohort. Incidentally, the comparison between FCH-PET/CT results and conventional imaging did not represent an end-point of the present study. A bias in selection of patients for inclusion in this analysis was minimized by careful review of consecutive patients referred for FCH-PET/CT imaging and by use of specific inclusion criteria to define a more homogenous group of patients with respect to disease biology and prognosis. The small number of patients in this study as well as insufficient information on assessment of possible change in management represents important study's limitations. The phenomena of "bone flare", has already been reported in patients after treatment with abiraterone in chemotherapynaive mCRPCa who had undergone bone scintigraphy, but that stands out for the first time also with FCH-PET/CT [24]. Three patients who underwent FCH-PET/CT during abiraterone acetate showed a partial response (n=1) , a stable disease (n=1) and a progressive disease (n=1) at imaging and also at PSA; however we did not find any flare phenomena in this subset of patients; in particular in the patient with tumor progression (the time between therapy and PET/CT scan was 4 months).

CONCLUSION
In conclusion, our results suggest that the change in SUVmax, rather than change in PSA, represents an important measure of response to chemotherapy and should be useful to stratify the prognosis of mCRPCa patients. The high cost of a PET/CT scan and the limited available data about the role of 11 C/ 18 F-choline PET/CT in the definition of response to therapy, requires a comprehensive review and cost-effective studies in order to approve this new indication.

CONFLICTS OF INTEREST
The authors report no conflicts of interest.