Molecular Imaging of Prostate Cancer: PET Radiotracers

Abstract

OBJECTIVE

Recent advances in the fundamental understanding of the complex biology of prostate cancer have provided an increasing number of potential targets for imaging and treatment. The imaging evaluation of prostate cancer needs to be tailored to the various phases of this remarkably heterogeneous disease.

CONCLUSION

In this article, I review the current state of affairs on a range of PET radiotracers for potential use in the imaging evaluation of men with prostate cancer.

The complex and heterogeneous biology of prostate cancer poses major challenges and opportunities. The disease is a major public health problem and is associated with significant emotional, comorbid practical, and economic costs. There is also an evolution in the development and use of specific endpoints for the emerging therapeutic approaches with the general goal to control, relieve, or eliminate disease manifestations (e.g., prostate-specific antigen [PSA], imaging findings, and symptoms) and to delay or prevent future disease manifestations [1]. Currently, imaging plays an important role in many aspects of this disease, but its role will need to evolve to accurately answer key clinical questions at various phases of the disease in a cost-effective manner. These clinical decision-making landmarks include accurate primary diagnosis, characterization and staging of cancer at the time of initial presentation, determination of local recurrence or distant disease at the time of biochemical recurrence of prostate cancer to select the most appropriate therapy, accurate assessment of therapy response to various treatment regimen under the new practice paradigm, and prediction of patient outcome (e.g., time-to-event endpoints such as time to hormone refractoriness in castrate-sensitive disease, time to progression, and overall survival).

The parallel advances in deciphering the molecular biology underpinnings of prostate cancer and the development of new imaging biomarkers and integrated imaging systems have provided unprecedented new and exciting opportunities to address these pivotal clinical needs. The biologically relevant targets for imaging may include metabolites (e.g., glucose, fatty acids, and amino acids), antigens (e.g., prostate-specific membrane antigen and prostate-specific stem cell antigen), androgen signaling (e.g., androgen receptor), angiogenesis, hypoxia, and gene-based pathways [2–21].

PET is fundamentally suited for the imaging evaluation of biologic targets and events. With the ever-increasing availability and access to hybrid PET/CT scanners, as well as the recent emergence of PET/MRI systems coupled with exceptional ongoing research and development in radiochemistry, the future for noninvasive molecular imaging–based examination of prostate cancer biology not only will lead to new enlightening levels of understanding of the disease but also will be clinically useful. With these introductory remarks in mind, I review the current evidence on the potential utility of a number of PET radiotracers in the imaging evaluation of prostate cancer.

Glucose Metabolism: ¹⁸F-FDG

A recognized hallmark of malignancy is the Warburg effect, which involves complex adaptive biologic mechanisms for enhanced cancer cell survival [22, 23]. PET with FDG takes advantage of this phenomenon and has excelled in the imaging evaluation of cancer [24]. Increased expression of cellular membrane glucose transporters and enhanced hexokinase enzymatic activity have been noted in many cancers that display augmented glucose metabolism [25, 26].

Effert et al. [27] assessed the expression of glucose transporter 1 (GLUT1) messenger RNA in hormone-independent and -sensitive prostate cancer cell lines. GLUT1 expression was higher in poorly differentiated hormone-independent cell lines than in well-differentiated hormone-sensitive cell lines. GLUT1 gene expression has also been shown to be generally higher in prostate cancer than in benign prostatic hyperplasia [28]. A recent study showed that overexpression of GLUT1 was observed in only some of the highly proliferative intraductal prostate cancers and that the expression level was generally low in most prostate tumors, probably reflecting the well-recognized heterogeneity of prostate tumor biology [29]. Xenograft prostate cancer mouse models have shown that androgen deprivation can significantly decrease the accumulation of FDG in hormone-sensitive tumors but it may not exert a significant effect on hormone-independent tumors [30–32].

Evaluation of the prostate gland on FDG PET/CT studies is challenging because of overlap of FDG uptake in normal, benign, and malignant tissues; the multifocal distribution of cancer deposits mixed with noncancerous cells; and the proximity of the gland to the urinary bladder. Despite these general limitations, one retrospective study assessed the level of FDG uptake in the “normal” prostate gland of men who had undergone FDG PET/CT without known or clinically suspected prostate disease [33]. Mean and maximum standardized uptake values (SUVs) were (mean ± SD) 1.3 ± 0.4 (range, 0.1–2.7) and 1.6 ± 0.4 (range, 1.1–3.7), respectively.

FDG PET generally has a limited role in the primary diagnosis and staging of prostate cancer in view of overlap of tracer accumulation in normal and abnormal prostate tissues [34]. Shiiba et al. [35] used a time-of-flight PET/CT to correlate the FDG uptake in primary prostate cancer with the biopsy specimen’s Gleason score. The cutoff SUV_max, sensitivity, and specificity for differentiating between biopsy specimens with a summed Gleason score of 5 or less and specimens with a summed Gleason score of 6 or greater were 2.8, 61.7%, and 80.0%, respectively. Another study reported a sensitivity of 80% and a positive predictive value of 87% for detection of prostate tumors with Gleason score of 7 and greater in men who presented with more than an intermediate risk of prostate cancer according to elevated serum PSA level [36]. Han and colleagues [37] evaluated the clinical significance of incidental focal prostate FDG uptake in men without known prostate cancer. The incidence of focal prostatic FDG uptake was 1.2%. Of the 55 cases that had clinical follow-up, only three had confirmed malignancy by biopsy, with SUV_max values of 2.3, 3.3, and 3.6; the remaining 52 cases were declared benign, with a mean SUV_max of 3.2 ± 1.7, which overlapped significantly with the malignant cases. These studies suggest that, in some cases, FDG PET may be able to characterize prostate tumors of sufficient size, which then can be helpful in imaging-directed biopsy of clinically relevant lesions [38].

The experience with FDG PET in the imaging evaluation of men with biochemical failure after definitive therapy for primary prostate cancer is quite limited. Biochemical failure is defined as an increase in serum PSA level with negative standard imaging studies after definitive therapy for primary prostate cancer. The American Urologic Association defines biochemical recurrence in postprostatectomy patients as an initial serum PSA level of 0.2 ng/mL or higher, with a second confirmatory level higher than 0.2 ng/mL [39]. In patients with prior primary external beam radiotherapy, the American Society for Therapeutic Radiology and Oncology consensus definition for biochemical failure recommends an increase of 2 ng/ mL or more above the nadir PSA level, regardless of hormonal therapy [40]. In one investigation, of 91 men with PSA relapse after prostatectomy, FDG PET detected local or systemic disease in 31% of patients with PSA relapse [41]. However, the validation criteria used in some cases included other positive standard imaging that violated the definition of PSA relapse–only clinical state [11]. In another study, FDG PET had sensitivity, specificity, positive predictive value, and negative predictive value of 75%, 100%, 100%, and 67.7%, respectively, for the detection of pelvic lymph node metastases, with histopathologic examination of the surgically harvested nodes for validation [42]. Recently, my colleagues and I [43] provided evidence in a prospective clinical study that FDG PET/CT plays a much more limited role than previously reported in the imaging evaluation of men with PSA relapse (range, 0.5–40.2 ng/mL) and strictly negative standard imaging studies.

In men with metastatic prostate cancer, FDG PET/CT may distinguish between metabolically active osseous lesions and metabolically dormant lesions, and the concordance of FDG PET/CT with other imaging studies appears to be higher in castrate-resistant metastatic disease than in castrate-sensitive disease [44–47]. FDG PET/CT may be useful in the evaluation of treatment response in metastatic prostate cancer, similar to the experience with many other cancers, with favorable response portrayed as therapy-induced decline in FDG uptake of tumor sites in comparison with the pretreatment scan [48, 49] (Fig. 1). However, again similar to the experience with other cancers, the response may be heterogeneous with the finding of significant decline in metabolic activity at some tumor sites but not in others. There may also be differences in imaging-based treatment response assessment depending on the response criteria used [50]. With regard to clinical impact, recent data analysis from the National Oncologic PET Registry in the United States has found that FDG PET/CT may indeed have a major impact on the clinical management of up to 35% of men with prostate cancer, mostly prompting treatment when none was planned before the PET scan [51]. Similar to other cancers, higher tumor FDG uptake in prostate tumor appears to be associated with poorer prognosis in comparison with tumors with lower metabolic activity [52, 53].

Fig. 1.

60-year-old man. CT (left), ¹⁸F-FDG PET (middle), and fused PET/CT (right) images show how androgen deprivation therapy induced decline in FDG uptake in osseous metastatic sites in posterior right ilium and left L5, with concomitant increase in lesion sclerosis. Top panels are pretreatment scans (prostate-specific antigen [PSA] level, 67.4 ng/mL), and bottom panels are posttreatment scans (PSA level, 0.3 ng/mL).

In summary, FDG PET/CT is generally limited in the primary diagnosis, staging, and restaging of prostate cancer. However, it may be useful in tumor characterization and in assessment of treatment response and prognosis in castrate-sensitive and castrate-resistant metastatic prostate cancer.

Lipogenesis

Acetate

Acetate is a simple metabolite that is preferentially transported across the cellular membrane through the monocarboxylate transporter. The two major sources of acetate consumption are the Krebs cycle and the metabolic pathways related to the production of phospholipids in cellular membranes facilitated by the fatty acid synthase reaction [54–56]. Preclinical studies have confirmed malignancy-induced up-regulation of fatty acid synthase as the underlying biologic basis for the enhanced ¹¹C-acetate uptake in prostate cancer [57, 58].

Normal biodistribution of ¹¹C-acetate displays high accumulation in the pancreas, variable uptake in the liver and bowel, and some renal uptake with little urinary excretion [59]. A three-compartment three-parameter model describes the acetate kinetics in prostate cancer adequately [60]. Similar to FDG and choline, there can be a considerable overlap among the uptake levels in primary cancer, benign prostatic hyperplasia, and normal prostate gland, but generally, the tracer uptake appears to be greater in the tumor than in the normal and benign prostatic hyperplasia tissue [61].

Limited number of comparative studies have shown that ¹¹C-acetate may be more sensitive than FDG in detection of primary and metastatic prostate cancer, although overall the tracers appear to be complementary [62, 63]. ¹¹C-acetate may be useful in the detection of tumor recurrence in men who had been previously treated for their primary disease, with lesion detectability of up to 75% [64, 65]. The success rate for lesion detection by ¹¹C-acetate appeared to be related to serum PSA level, with a 59% positive rate in patients with serum PSA level greater than 3 ng/mL that declined significantly to 4% in patients with serum PSA levels 3 ng/mL or lower [66].

An ¹⁸F-labeled formulation of acetate has also been reported with potential use in prostate cancer, although experience with this tracer remains scant [67]. A comparative animal study of ¹¹C-acetate and ¹⁸F-fluoroacetate showed that, for most organs, the tumor-to-organ uptake ratios at 30 minutes after tracer administration were higher with ¹⁸F-fluoroacetate than with ¹¹C-acetate, whereas the tumor-to-heart and tumor-to-prostate ratios were similar [68]. A recent investigation in cynomolgus monkeys and pigs showed that ¹⁸F-fluoroacetate is not a functional analog of ¹¹C-acetate in normal physiology, with ¹⁸F-fluoroacetate showing relatively protracted blood retention, rapid clearance from liver, excretion in bile and urine, and defluorination (i.e., high bone uptake) [69]. Finally, in a limited study that compared ¹¹C-acetate and ¹¹C-choline, the tracers appeared to be similarly useful in imaging prostate cancer in individual patients [70].

Choline

Choline is a water-soluble essential nutrient that was discovered in 1864 by Adolph Strecker. Choline enters the cell through choline transporters and is the precursor for the biosynthesis of phospholipids, which are major components of the cellular membrane. The biologic basis for the accumulation of radiolabeled choline in tumors is, in part, due to overexpression of choline kinase in support of malignancy-induced increased demand for cellular membrane synthesis. Choline kinase catalyzes the phosphorylation of choline to form phosphorylcholine, followed by generation of phosphatidylcholine in the tumor cell membrane [71, 72]. Choline uptake in prostate tumor appears to be affected by hypoxia but may be uncorrelated to cellular proliferation [73, 74].

Initial studies with choline were performed with ¹¹C as the radiolabel [75]. Normal biodistribution of ¹¹C-choline shows relatively high accumulation in the pancreas, liver, kidneys, and salivary glands; variable uptake in the bowel; and little urinary excretion, with the last feature advantageous for the assessment of the prostate gland. Scher et al. [76] reported sensitivity of 87% and specificity of 62% for the detection of primary prostate cancer with histopathologic examination as the reference standard. Another similar investigation found a lower sensitivity of 66% and a higher specificity of 81% for localization of primary prostate cancer based on a sextant histopathologic analysis [77]. When ¹¹C-choline was compared with transrectal ultrasound, both PET and transrectal ultrasound tended to understage the disease [78]. To reduce the partial volume effect, Martorana et al. [79] assessed the diagnostic performance of ¹¹C-choline PET/ CT for prostate nodules 5 mm or larger in men with suspected prostate cancer before the 12-core standard transrectal ultrasound–guided prostate biopsy. Although PET had a relatively high sensitivity (83%) for detection of cancer, its sensitivity for assessment of extraprostatic extension was quite low (22%). Given the range of reported detection rates with ¹¹C-choline PET/CT, a rational query may be whether the detection rate is associated with relevant parameters, such as tumor grade, location, size, and PSA level. For example, it has been noted that tumor configuration may affect detectability (unifocal, 49%; multifocal, 21%; and rindlike, 11%) [80]. Giovacchini and colleagues [81] reported no significant correlation between lesion SUV_max, PSA level, and Gleason score on a patient-based analysis.

A study of the comparative diagnostic performance of ¹¹C-choline PET/CT, MRI, and MR spectroscopy for detection of primary prostate cancer, with histologic analysis as the reference standard, had sensitivity and specificity of 55% and 86%, respectively, for PET/CT, 54% and 75% for MRI, and 81% and 67% for MR spectroscopy [82]. A similar investigation reported sensitivity of 100% for ¹¹C-choline PET, 60% for MRI, and 65% for MR spectroscopy [83]. Eschmann et al. [84] compared ¹¹C-choline PET/CT with whole-body MRI for staging of prostate cancer with histologic analysis and follow-up for validation criteria. The sensitivity and specificity were 97% and 77%, respectively, for ¹¹C-choline PET and 79% and 94% for whole-body MRI. These results suggested that PET and MRI might provide complementary diagnostic information. With the recent development of sophisticated image fusion software, integrated PET/MRI whole-body imaging systems, high-field MRI, and multiparametric MRI techniques (e.g., diffusion-weighted imaging and dynamic contrast enhancement), there may be opportunities for synergism between PET and MRI for more accurate detection and staging of primary prostate cancer [85–87].

Considering the cumulative experience with ¹¹C- and ¹⁸F-choline PET, the overall reported sensitivity ranges between 38% and 98% for the detection of locally recurrent and metastatic disease in men with biochemical recurrence of prostate cancer [88–92]. The wide range of sensitivity may relate to the heterogeneity of patient population in terms of PSA level ranges, type of primary therapy, and the type and quality of validation criteria. For example, in one study, ¹¹C-choline was determined to localize recurrence in a higher percentage of men after primary radiation therapy than after radical prostatectomy (78% vs 38%) [93]. In another study, the sensitivity and specificity of ¹¹C-choline PET/CT for the detection of local recurrence after radical prostatectomy were 73% and 88%, respectively [94]. Scattoni et al. [95] reported sensitivity of 64% and specificity of 90% for detection of nodal metastases in men with PSA relapse after radical retropubic prostatectomy.

The detection rate of choline PET may be associated with the PSA level or the other PSA-derived parameters. Therefore, many investigations have focused on this potential relationship to identify a “trigger” PSA level to decipher when patients should be considered for imaging evaluation with choline PET. Unfortunately, many studies have not used the strict requirement for negative standard imaging results. Thus, the unique contribution of choline PET in this clinical setting has become blurred. Despite this limitation, the general notion has been that higher PSA levels, higher PSA velocity, and shorter PSA doubling times may be associated with higher detection rate on choline PET, probably reflecting the amount of tumor burden available for detection.

Krause et al. [90] reported ¹¹C-choline detection rates of 36% for PSA levels less than 1 ng/mL, 43% for PSA levels 1–2 ng/mL, 62% for PSA levels 2–3 ng/mL, and 73% for PSA levels 3 ng/mL or higher. Castellucci et al. [96] evaluated the likelihood of lesion detection on ¹¹C-choline PET/CT in 190 men after radical prostatectomy who presented with PSA relapse in the 0.2–25.4 ng/mL range. The likelihood of lesion detection was increased with PSA levels higher than 2.4 ng/ mL and with PSA doubling times of less than 3.4 months or PSA velocities of higher than 1 ng/mL/y when the PSA levels were less than 2.4 ng/mL. In another investigation, ¹¹C-choline and FDG PET were both used in detecting disease in 73 men with PSA relapse [97]. Validation was by biopsy or increase in PSA without therapy and decrease in PSA after therapy. At all PSA levels, ¹¹C-choline had twice higher sensitivity of 60.6%, in comparison with 31% for FDG. In a recent report, Picchio et al. [98] compared ¹¹C-choline PET/CT and standard bone scintigraphy for the detection of bone metastases in 78 men with biochemical progression of disease (PSA level, 0.2–500 ng/mL). The findings on each scan were designated as positive, negative, or equivocal for malignancy, and then the diagnostic performance of the imaging studies were assessed twice, once with equivocal findings as true-positive and once as true-negative. The ranges of sensitivity and specificity were 89–89% and 98–100%, respectively, for ¹¹C-choline and 100–70% and 75–100% for bone scintigraphy. The authors concluded that ¹¹C-choline and bone scintigraphy are complementary, with the former offering higher specificity and the latter providing higher sensitivity. Also in that study, the diagnostic performance of ¹¹C-choline PET/CT was not affected by androgen deprivation therapy. Conversely, another group of researchers showed that androgen deprivation therapy significantly and concordantly reduces PSA levels and ¹¹C-choline uptake in men with castrate-sensitive prostate cancer [99]. The latter observation of an androgen effect on choline uptake in patients was in line with the recently reported in vitro cell line studies [32].

DeGrado et al. [100] reported the synthesis of an ¹⁸F-labeled formulation of choline. Over the past decade, ¹⁸F-fluorocholine has received much attention given the advantages of the longer half-life of ¹⁸F in comparison with that of ¹¹C [101]. Preclinical studies have found high ¹⁸F-fluorocholine uptake in prostate cancer with little effect from castration [102, 103].

Normal biodistribution of ¹⁸F-fluorocholine shows relatively high uptake in the pancreas, liver, spleen, and kidneys; variable uptake in the bowel; and excretion into urine. Similar to FDG and ¹¹C acetate, the uptake of ¹⁸F-fluorocholine overlaps among normal, benign, and malignant prostate tissues [104]. Also similar to the case with ¹¹C-choline, there are mixed results, with the potential utility of ¹⁸F-fluorocholine in the diagnosis and staging of primary prostate cancer [105]. Beheshti et al. [106] performed ¹⁸F-fluorocholine PET/ CT in men with clinically organ-confined tumor but who were at intermediate (PSA level, 10–20 ng/mL; Gleason score, 7) and high (PSA level, > 20 ng/mL; Gleason score, ≥ 8) risk for extracapsular extension before undergoing radical prostatectomy with extended pelvic lymph node dissection. The sensitivity, specificity, positive predictive value, and negative predictive value of ¹⁸F-fluorocholine for the detection of pelvic lymphadenopathy were 45%, 96%, 82%, and 83%, respectively, for all lymph node sizes and 66%, 96%, 82%, and 92% for lymph nodes 5 mm or larger (i.e., nodes larger than the PET spatial resolution limitation). In general, however, although there is no established role for ¹⁸F-fluorocholine PET in the initial diagnosis and staging of prostate cancer, the technique may be of some value in imaging-guided investigational treatments (e.g., focal therapy).

Fluorine-18-fluorocholine (similar to ¹¹C-choline) may have a role in the management of men with biochemical recurrence of prostate cancer with a diagnostic performance that appears to improve with increasing PSA level, although it is not recommended for routine use when the PSA level is less than 1 ng/mL [89] (Fig. 2). Pelosi et al. [107] reported a detection rate of 20% for PSA level 1 ng/mL or lower, 44% for PSA level 1–5 ng/ mL, and 82% for PSA level greater than 5 ng/mL. Another investigation showed a 41% true-positive rate in restaging patients with PSA levels less than 5 ng/mL [108]. A recent report determined the relationship between PSA kinetics and ¹⁸F-fluorocholine PET/CT detection rate in 82 men with biochemical recurrence of prostate cancer after total prostatectomy [109]. The median PSA level was significantly higher in PET-positive than PET-negative patients (4.3 vs 1.0 ng/mL; p < 0.01). A PSA level of 1.74 ng/ mL was determined to be the optimal PSA threshold for detection of recurrent prostate cancer, with a sensitivity of 82% and specificity of 74%. Moreover, the median PSA velocity was significantly higher for PET-positive than PET-negative patients (6.4 vs 1.1 ng/mL/y; p < 0.01) with an optimal threshold of 1.27 ng/mL/y. On a similar theme, Schillaci et al. [110] recommended that ¹⁸F-fluorocholine PET/CT may be considered in men with PSA level greater than 2 ng/mL, PSA doubling time of 6 months or less, and PSA velocity greater than 2 ng/mL/y.

Fig. 2.

Man with biochemical relapse of prostate cancer (prostate-specific antigen [PSA] level, 2.68 ng/mL). A–C, CT (A), PET (B), and fused PET/CT (C) images show pathologically increased ¹¹C-choline uptake in subcentimeter right external iliac lymph node. Pelvic lymph nodal area was irradiated with resultant decline in PSA to undetectable level at 3 months after completion of therapy regimen. (Reprinted with permission from [218])

Langsteger et al. [111] compared ¹⁸F-fluorocholine PET/CT and ¹⁸F-NaF PET/CT for the detection of bone metastases in 40 men with primary or recurrent prostate cancer. A lesion-based comparison showed no significant difference. A patient-based comparison found the same sensitivity of 91% for both ¹⁸F-fluorocholine and ¹⁸F-NaF but higher specificity of 89% for ¹⁸F-fluorocholine, in comparison with 83% for ¹⁸F-NaF. Although that study suggested that ¹⁸F-fluorocholine PET/CT may be able to replace ¹⁸F-NaF PET/CT, another report [112] concluded that combined imaging may be most useful for management decisions and accurate treatment response assessment. Beheshti et al. [113] correlated the uptake of ¹⁸F-fluorocholine in bone metastases to the morphologic changes on CT. Lytic lesions showed higher choline uptake than did blastic lesions (average SUV_max, 11 ± 3.2 for lytic lesions vs 7.8 ± 3.0 for blastic lesions). Hormonal therapy did not significantly affect the choline uptake in the osseous lesions. Three PET/CT patterns for bone metastases were identified: those with ¹⁸F-fluorocholine uptake only (i.e., bone marrow infiltration without morphologic changes on CT), those with both ¹⁸F-fluorocholine uptake and CT morphologic changes, and lesions with no ¹⁸F-fluorocholine uptake but high density on CT (nonviable tumor). Similar findings have been observed with FDG PET/CT [114].

Cellular Proliferation

Imaging cellular proliferation provides useful noninvasive diagnostic information about the rate of tumor growth and early assessment of treatment response [115–117]. PET in conjunction with radiotracers that track the thymidine salvage pathway of DNA synthesis has been studied relatively extensively for imaging tumor cellular proliferation [118]. Although ¹¹C-thymidine was an early candidate, its rapid catabolism complicated its kinetic model analysis and limited its practical utility [119–122]. In this section, I review the preclinical and pilot clinical experiences with two major ¹⁸F-labeled PET radiotracers in the imaging evaluation of cellular proliferation in prostate cancer.

[Fluorine-18]-3′-Deoxy-3′-Fluorothymidine (FLT)

FLT is phosphorylated by thymidine kinase (TK) 1, which is retained in proliferating cells without DNA incorporation and can be described by a three-compartment kinetic model [123–125]. Recently, Kukuk et al. [31] reported the pharmacokinetics of FLT, FDG, and ¹¹C-choline in two hormone-independent (PC-3 and DU145) and two hormone-dependent (CWR22 and PAC 120) prostate cancer xenograft mouse models using PET. Both FLT and FDG showed the highest uptake in PC-3 hormone-independent tumors. FDG uptake in hormone-dependent CWR22 tumors was also noted to be high but decreased significantly after androgen deprivation therapy. Although, in that study, FLT uptake was insufficient at baseline to provide reliable information on response to therapy, other studies have shown that FLT uptake is markedly reduced after castration or treatment with diethylstilbestrol [126]. A significant decline in FLT uptake has also been noted in the 22Rv1 hormone-refractory prostate tumors implanted in athymic mice in response to docetaxel treatment [127]. Despite these few early encouraging results, the exact role of FLT in the imaging evaluation of response to treatment in men with prostate cancer awaits more extensive studies. A complicating factor is also the physiologically high level of marrow FLT uptake that can hinder osseous lesion detection and assessment.

[Fluorine-18]-2′-Fluoro-5-Methyl-1-β-D-Arabinofuranosyluracil (FMAU)

FMAU is a thymidine analog that is phosphorylated by TK and incorporated in DNA. FMAU is preferentially phosphorylated by the mitochondrial TK2 in comparison with the cytosolic TK1 [128]. Unlabeled FMAU was originally of clinical interest as an anticancer and an antiviral drug when used in pharmacologic doses [129]. In tracer doses, FMAU can be labeled with ¹¹C or ¹⁸F and has been noted to be useful for imaging tumor proliferation [130–134]. FMAU has also been used for imaging reporter gene expression using the herpes simplex virus type 1 TK1 system [135–138]. An automated GMP–compliant radiosynthesis of FMAU has been described recently [139].

Carbon-14-FMAU behaves very similarly to thymidine with respect to cellular uptake velocity, saturability of cellular incorporation, and intracellular metabolite pools and is reflective of tumor cell division [140]. A recent report showed that ¹¹C-FMAU uptake in a dog brain tumor model correlated with tumor growth rate and could be well described by a three-compartment kinetic model [141]. Other researchers have supported the adequacy of a three-compartment model for FMAU [133]. The D-isomer of the compound shows higher accumulation in tumors than those for the L-isomer and FLT [142].

FMAU is resistant to degradation and has very little accumulation in bone and in urinary bladder, which renders it a potentially ideal PET radiotracer for imaging in prostate cancer [143, 144]. There may be an association between androgen signaling and thymidine metabolism, possibly related to the androgen control of mitochondrial function, including TK2 enzymatic activity [145, 146].

Sun et al. [147] showed in a small pilot study that FMAU accumulated in locally recurrent (tumor-to-background pelvis activity ratio, 2.3–6.3) and osseous metastatic sites (tumor-to-background normal bone activity ratio, 2.4–3) of prostate cancer. Moreover, FMAU displayed rapid blood clearance (95% of blood activity cleared within 10 minutes) and stability (about 70% of urine activity as intact FMAU at 60 minutes).

Other substituted 2′-[¹⁸F]fluro-2′-deoxy- arabinofuranosyluracil derivates such as 2′-deoxy-2′-[¹⁸F]fluoro-5-bromo-1-β-D-arabinofuranosyluracil, 2′-deoxy-2′[¹⁸F] fluro-5-chloro-1-β-D-arabinofuranosyluracil, 2′-deoxy-2′-¹⁸F-fluoro-5-fluoro-1-β-D-arabinofuranosyluracil, and others have also been synthesized. However, their potential competitive advantage over FLT and FMAU will need further investigation [148, 149].

Receptors

Androgen Receptor (AR)

Androgens are essential for the development, growth, and maintenance of the prostate. The effects of androgens are exerted via the nuclear AR, which is a ligand-dependent (either testosterone or 5α-dihydrotestosterone) transcription activator involved in cellular proliferation and differentiation and is present in all histologic types of prostate tumors, in recurrent carcinoma, and in tumor metastases [150–152]. Almost all patients respond favorably to androgen ablation, but virtually all patients will eventually relapse to a castrate-resistant clinical state, which is believed to occur via bypassing or sensitizing the AR pathway. The factors involved may be AR mutation, such that the receptor is either activated promiscuously or is activated in a ligand-independent manner by amplification of coactivators, activation of oncogenes, as well as autocrine growth factors stimulation [152].

With the recent treatment trials including high-dose testosterone or the use of AR-targeted drugs (e.g., AR antagonist MDV3100), imaging biomarkers directed to the AR seems quite rational [153, 154]. Several ligands for the AR have been developed, including 16β-[¹⁸F]fluoro-5α-dihydrotestosterone (¹⁸F-FDHT) [155–157]. Larson et al. [158] assessed the in vivo targeting and biokinetics of ¹⁸F-FDHT in seven patients with clinically progressive metastatic prostate cancer. The metabolism of ¹⁸F-FDHT was rapid, with 80% conversion within 10 minutes to radiolabeled metabolites circulating bound to plasma proteins, with tumor uptake of tracer that was rapid and stable. In another study of 20 men with metastatic prostate cancer, ¹⁸F-FDHT PET had a sensitivity of 63%, with a noted decline in tumor uptake level 1 day after treatment with flutamide [159]. Moreover, positive PET studies were significantly associated with higher PSA levels. Pharmacokinetics assessment of ¹⁸F-FDHT showed that tumor uptake reached a plateau within 20 minutes and that radiolabeled metabolites were not bound to AR on the basis of in vitro studies with CWR22 cells [160]. Preliminary results from a study comparing ¹⁸F-FDHT and FDG has shown that there are AR-predominant, glycolysis-predominant, and AR-glycolysis–concordant states, which may have implications for treatment response and prognosis [161] (Fig. 3). Moreover, despite encouraging antitumor activity with MDV3100 in men with castrate-resistant prostate cancer (e.g., decline in PSA, documented response in soft-tissue disease, stabilization of bone disease, and conversion from unfavorable to favorable circulating tumor cell counts), the MDV3100-induced ¹⁸F-FDHT uptake changes in tumor did not necessarily parallel the changes in tumor FDG uptake, suggesting that ¹⁸F-FDHT may be a pharmacodynamic marker as opposed to a treatment response marker in this setting [47, 153].

Fig. 3.

Sagittal fused PET/CT and PET images of 16β-[¹⁸F]fluoro-5α-dihydrotestosterone (FDHT) (top) and FDG (bottom) uptake in two different patients with castrate-resistant metastatic prostate cancer.

A, Glycolysis–androgen receptor concordant phenotype.

B, Androgen receptor predominant phenotype. (Reprinted with permission from [47])

Radiation dosimetry of ¹⁸F-FDHT has shown absorbed doses from the lowest level in skin (0.00057 ± 0.000281 cGy/MBq) to the highest level in the urinary bladder wall (0.00868 ± 0.00481 cGy/MBq) with 1–2 hours voiding interval, and an effective dose equivalent of 0.00177 ± 0.000152 cSv/MBq [162]. The maximum recommended administered activity was noted to be 331 MBq (8.9 mCi) to keep the maximum normal-tissue absorbed dose below the recommended maximum permissible dose of 5 cGy per single administration.

Gastrin-Releasing Peptide Receptor (GRPR)

Bombesin is a 14–amino acid analog of the human gastrin-releasing peptide that binds to the GRPR. Because GRPR is overexpressed in prostate cancer, there has been an increasing interest in developing PET radiotracers that target the GRPR [163–173]. In a recent study, ⁶⁸Ga-labeled bombesin analog AMBA (⁶⁸Ga-DO-TA-CHCO-Gly-4-aminobenzyl-Gln-Trp-Ala-Val-Gly-His-Leu-Met-NH) was compared with ¹⁸F-fluorocholine in nude mice bearing human prostate cancer xenografts [174]. The tumor uptake of ⁶⁸Ga-AMBA was significantly higher than that with ¹⁸F-fluorocholine, suggesting that GRPR-targeted imaging may be superior to choline-based imaging. This and other encouraging results with radiolabeled bombesin analogs suggest a potentially important role in the imaging evaluation of prostate cancer that will deserve additional investigation.

Amino Acid

Carbon-11-Methionine

Tumor uptake of ¹¹C-methionine reflects enhanced amino acid transport and protein synthesis that may be present in malignancy. Relatively few data have been published on the potential utility of ¹¹C-methionine in prostate cancer. In a pilot study of 10 patients with progressive prostate cancer, FDG and ¹¹C-methionine were compared. Although both FDG and ¹¹C-methione were taken up by the selected index lesions, ¹¹C-methionine showed a higher tumor-to-blood ratio, more rapid tumor uptake after tracer administration, and a flatter uptake profile than those for FDG [175]. In a prospective study, 12 patients with PSA relapse and at least one site of new or increasing disease on conventional imaging (bone scintigraphy, CT, or MRI) underwent both FDG PET and ¹¹C-methione PET [176]. With conventional imaging modalities as the standard of reference, the sensitivities of FDG PET and ¹¹C-methionine were 48% and 72.1%, respectively. The authors also suggested that initial ¹¹C-methione uptake may indicate dormant sites of disease followed by an increase in FDG uptake as the disease progresses, therefore reflecting the time-dependent metabolic cascade of progressive metastatic prostate cancer.

Tóth et al. [177] studied 20 men with high PSA levels (3.49–28.6 ng/mL) and negative prostate biopsies. The true-positive detection rate of primary prostate cancer was 35% for the whole group and 46.7% in the PET-positive group. PET was negative in all five patients with negative repeat prostate biopsies. Therefore, it appears that nonstandard imaging with ¹¹C-methionine, if available, may be useful in this group of men with suspected prostate cancer but negative initial biopsy results. More recently, Shiiba et al. [35] performed a similar study in 20 men with suspected primary prostate cancer comparing ¹¹C-methionine and FDG in relation to the tumor Gleason sum score. The tracer uptake in the prostate gland were recorded by small 100-mm² round regions of interest placed at six standard locations in the peripheral zone and four standard locations in the apex of the transitional zone. The Gleason sum scores of tumor specimens were grouped as no grade (Gleason sum score of ≤ 5), low grade (Gleason sum score of 6 or 7), and high grade (Gleason sum score of 8–10). The difference in ¹¹C-methione uptake levels between the no-grade and low-grade groups was statistically significant. For FDG, the difference in uptake levels between the no-grade and high-grade groups was statistically significant. The cut-off SUV_max and the corresponding sensitivity and specificity for differentiating between no-grade and low-grade plus high-grade groups were 3.15, 78.7%, and 75.6%, respectively, for ¹¹C-methionene and 3.00, 62.8%, and 78.9% for FDG. For separating no-grade plus low-grade and high-grade groups, the cutoff SUV_max and the corresponding sensitivity and specificity were 3.76, 70.1%, and 89.7%, respectively, for ¹¹C-methionine and 3.47%, 62.7%, and 86.3% for FDG.

Anti-1-Amino-3-(¹⁸F)-Fluorocyclobuate-1-Carboxylic (Anti-¹⁸F-FACBC)

Anti[¹⁸F]-FACBC is a synthetic l-leucine analog that has been found to accumulate in prostate cancer with relatively little renal excretion [178]. The tracer accumulation in prostate cancer cells correlates with the expression level of the alanine-, serine-, and cysteine-preferring system–mediated amino acid transport, with the large neutral amino acid transporter as an important transport system in the typical intratumoral acidic microenvironment [179]. It has also been shown that the tracer does not get incorporated into proteins [180]. Schuster et al. [181] described their initial experience with anti-¹⁸F-FACBC PET in prostate cancer. Visual analysis detected malignancy in 40 of 48 prostate sextants, in seven of nine pelvic nodal basins, and in all four men with local recurrence (Fig. 4). In another study, investigators compared anti-¹⁸F-FACBC PET/ CT with ¹¹¹In-capromab pendetide SPECT/ CT in 50 men with suspected recurrent prostate cancer after definitive therapy for localized disease and negative standard bone scans [182]. Validation was by a combination of tissue histopathologic analysis, additional imaging, and laboratory and clinical data. The sensitivity and specificity for detection of prostate bed recurrence were 89% and 67%, respectively, for anti-¹⁸F-FACBC PET/ CT and 69% and 58% for ¹¹¹In-capromab pendetide SPECT/CT. For detection of extraprostatic involvement, the sensitivity and specificity were 100% and 100%, respectively, for anti-¹⁸F-FACBC PET/CT and 10% and 100% for ¹¹¹In-capromab pendetide SPECT/ CT. Therefore, anti-¹⁸F-FACBC PET/CT was clearly more accurate than ¹¹¹In-capromab pendetide SPECT/CT, especially for detection of occult extraprostatic sites of disease in men with biochemical failure. A case study of the potential utility of anti-¹⁸F-FACBC PET/CT in guiding prostate bed radiotherapy has also been reported [183].

Fig. 4.

71-year-old man with biopsy-proven prostate bed recurrence. anti-¹⁸F-FACBC PET maximum-intensity-projection image at 20 minutes shows high uptake in prostate bed (arrow) and little bladder uptake (arrowhead). (Reprinted with permission from [181])

Bone Matrix

Fluorine-18–Labeled Sodium Fluoride (¹⁸F-NaF)

Bone scintigraphy with ^99mTc-based radiotracers grossly underestimates the true prevalence of bone metastasis and does not provide a quantifiable metric for monitoring treatment response [184, 185]. These shortcomings may be overcome with ¹⁸F-NaF, which is actually an old tracer approved for clinical use by the U.S. Food and Drug Administration in 1972 [186]. Despite the superior performance of ¹⁸F-NaF (i.e., high-quality images with rapid blood clearance and high bone-to-background ratio as well as shorter time from tracer administration to imaging), the tracer was not widely used in view of the technical limitations of gamma cameras with high-energy photon imaging and the scarcity of PET scanners. The situation has now changed, with more widespread availability of PET/CT scanners, commercial regional distribution of PET tracers, and recent shortages of ^99mTc-labeled tracers. The recent decision by the Centers for Medicare & Medicaid Services to reimburse for ¹⁸F-NaF PET/CT through the National Oncologic PET Registry to assess the effect of ¹⁸F-NaF PET/CT on referring physicians’ intended management of patients with known or suspected bone metastases has also contributed to its increased clinical use.

Fluorine-18-NaF is rapidly cleared from plasma in a biexponential manner with essentially all tracer retained by bone after a single pass. The bone uptake is related to chemisorption with exchange of ¹⁸F⁻ ion for OH⁻ ion-on the surface of the hydroxyapatite matrix of bone to form fluoroapatite and migration of ¹⁸F⁻ ion into the crystalline matrix of bone. There is minimal binding to serum protein and rapid renal clearance [187, 188]. The “super-scan” uptake pattern in widespread osseous metastatic disease is similar to that of standard bone scintigraphy with little renal tracer localization [189]. Procedure guidelines and dosimetry estimates for ¹⁸F-NaF have been published elsewhere [190].

Hsu et al. [191] reported in a preclinical animal model that ¹⁸F-NaF PET/CT can be useful in characterizing osseous lesions induced by human prostate cancer. Human studies have shown that, in particular, bone lesions with sclerotic or mixed changes tend to show high ¹⁸F-NaF uptake [192]. Even-Sapir and colleagues [193] compared ^99mTc-methylene diphosphonate planar bone scintigraphy, SPECT, ¹⁸F-NaF PET, and ¹⁸F-NaF PET/CT in 25 men with newly diagnosed prostate cancer with Gleason scores of 8 or higher or PSA levels of 20 ng/mL or higher or nonspecific sclerotic lesions on CT and 19 patients who were referred for evaluation of suspected recurrence or progression of disease. They followed rule-based criteria to categorize lesions as malignant, benign, or equivocal with a follow-up range of 6–15 months for establishing truth. In a patient-based analysis, the sensitivity and specificity were 70% and 57%, respectively, for planar bone scintigraphy, 92% and 82% for multiple-FOV SPECT, 100% and 62% for ¹⁸F-NaF PET, and 100% and 100% for ¹⁸F-NaF fluoride PET/CT. The high sensitivity and specificity of ¹⁸F-NaF PET/CT allows the detection of occult bone metastases that are missed on standard bone scintigraphy, with important implications on patient management. For example recently, my colleagues and I [43] reported a true-positive detection rate of 16.2% for occult osseous metastases in 37 men with biochemical recurrence of prostate cancer who underwent ¹⁸F-NaF PET/CT (Fig. 5).

Fig. 5.

54-year-old man with biochemical recurrence of prostate cancer and negative conventional bone scintigraphy. CT (left), ¹⁸F-NaF PET (middle), and fused ¹⁸F-NaF PET/CT (right) transaxial images show subtle sclerosis at right anterolateral aspect of thoracic vertebral body, which is clearly active on PET before androgen deprivation therapy (top) and remains active with increasing sclerosis after 4 months of treatment (bottom).

With regard to the use of ¹⁸F-NaF in monitoring treatment response, a very recent small pilot study of five patients with castrate-resistant metastatic prostate cancer showed that semiquantitative ¹⁸F-NaF PET was more accurate than qualitative comparison of scans in assessing response of bone metastases to 223Ra-chloride (Alpharadin, Algeta) therapy and in correlating better with PSA and alkaline phosphatase changes [194]. Despite potential advantages of ¹⁸F-NaF PET/CT over standard bone scintigraphy in quantification of treatment response, it should probably be recalled that ¹⁸F-NaF is not immune to the display of therapy-induced flare phenomenon [195].

The final adoption of ¹⁸F-NaF in replacing standard bone scintigraphy will likely depend on the cost and availability issues in a balancing act with the established increased diagnostic performance. The current emphasis on comparative effectiveness research may be able to provide valuable information for clinical decision making in individual patients and as a matter of health care policy in more general terms.

Prostate-Specific Membrane Antigen (PSMA)

PSMA is a cell surface transmembrane glycoprotein that is overexpressed on prostate tumor cells and thus provides a rational target not only for diagnosis and direct therapy but also for monitoring of PSMA expression changes with non–PSMA-based therapy [196]. For example, it has been reported that PSMA expression may be suppressed by androgen treatment, whereas antiandrogens (e.g., MDV3100) may up-regulate its expression [197].

Despite the relevance of PSMA in prostate cancer, the currently available ¹¹¹In-labeled capromab pendetide, which targets the internal moiety of PSMA in apoptotic or necrotic tissue, has shown limited predictive value in imaging the prostate fossa and has low sensitivity for detecting osseous metastases [198]. The PET radiotracer 89Zr-desferrioxamine B (DFO)-7E11 (half-life, 78.4 hours), which also targets the intracellular epitope of PSMA in dead or dying cells, has been recently evaluated in murine xenograft models of human prostate cancer and showed potential utility for monitoring and quantifying tumor response to irradiation [199].

There have also been significant efforts in designing other tracers that target PSMA [200–203]. J591 is a monoclonal antibody against the extracellular domain of PSMA that has been radiolabeled with 89Zr-desferrioxamine B (DFO)-7E11 for immunoPET of PSMA expression in live (as opposed to dying or dead) prostate cancer cells [204]. Other platforms include the use of PSMA inhibitors, such as ¹⁸F-labeled N-[N-[(S)-1,3-Dicarboxypropyl] carbamoyl]-4-[¹⁸F]fluorobenzyl-l-cysteine and ¹⁸F-labeled phosphoramidate peptidomimetic, which have been shown in preclinical animal model studies to localize to PSMA-positive–expressing tumors with high specificity [205, 206].

Gene-Mediated Imaging

Interesting research is being performed in gene-mediated theranostics to diagnose and treat castration-resistant prostate cancer [207–210]. These methods include the potential utility of a fused enhancer derived from the PSA and the PSMA gene regulatory region with activity augmentation by the two-step transcriptional amplification system [211–213]. The augmented prostate-specific two-step transcriptional amplification method has also been shown to be AR dependent, which can potentially reflect the functional status of AR [214]. In one experiment, recombinant human adenoviral vectors were used to directly image nodal metastases of a prostate cancer model through prostate-restricted tumor expression of optical and PET reporter genes [215]. Such a platform may allow direct noninvasive visualization and characterization of the sentinel lymph node, such as the reporter gene herpes simplex virus type 1 TK and reporter probe 9-(4-(¹⁸F) Fluoro-3-[hydroxymethyl]butyl)guanine, instead of the current indirect method with sentinel lymph node localization followed by tissue sampling. Such a platform will also allow the possibility of tumor-specific suicide gene therapy with prodrugs such as ganciclovir [216, 217]. These encouraging results will be complemented by additional studies on cancer specificity versus tissue specificity, efficiency of reporter gene delivery, transduction, and host immunoreactivity.

Summary

There has been much recent activity in the research and development of PET radiotracers for potential use in prostate cancer. Many of these radiotracers are in the early phases of evaluation in preclinical studies, whereas some are being tested in pilot clinical projects. The most common PET radiotracer, FDG, may have its major role in monitoring of treatment response in advanced metastatic prostate cancer. Although both ¹¹C-acetate and ¹¹C-choline appear to be somewhat equally useful in imaging prostate cancer in individual patients, most of the recent focus has been on ¹⁸F-fluorocholine for detection of locally recurrent or metastatic disease in men with PSA relapse. Fluorine-18-NaF may soon replace standard ^99mTc-based bone scintigraphy in view of its superior diagnostic performance. Other nonspecific radiotracers for prostate cancer, such as ¹⁸F-FMAU, ¹⁸F-anti-FACBC, and ⁶⁸Ga-AMBA, may provide new insights in prostate cancer biology that will also be useful in specific clinical circumstances. More specific radiotracers, such as ¹⁸F-FDHT, PSMA, and gene-based bio-markers, may offer new ways to understand the biologic and functional heterogeneity of prostate cancer and to assess responses to the emerging targeted therapies. On the basis of the current evidence, the future routine clinical use of PET in the imaging evaluation of prostate cancer is ensured.