Preclinical PET tracers for the evaluation of sarcomas: understanding tumor biology

Abstract

Sarcomas are rare tumors of mesenchymal origin. Sarcomas display significant histological heterogeneity, resulting in significant imaging heterogeneity. ¹⁸F-FDG PET has is increasingly used for the evaluation, staging and surveillance of patients with sarcomas. ¹⁸F-FDG PET maximum SUV has been shown to be correlated with sarcoma grade and overall survival. This has led to interest in alternative PET tracers to assess the biological characteristics of tumors and guide treatment decisions. Here we investigate novel PET/CT tracers used for the evaluation of sarcomas over the past 20 years and summarize what we have learned about sarcoma tumor biology from these studies.

Introduction

Bone and soft tissue sarcomas are a heterogeneous group of tumors thought to be derived from mesenchymal tissues [1]. Sarcomas are rare cancers which comprise approximately 21% of pediatric cancers and 1% of new cancer diagnoses in adults [1,2]. The prognosis for sarcoma patients can vary dramatically. Five-year survival after a new sarcoma diagnosis in patients without metastatic disease is 65% [1], but varies due to many factors including histologic grade, sarcoma subtype, and tumor size [2-4]. Tumor type, size and histologic grade are among the most important prognostic factors for patients with sarcoma. Tumor grade is related to many histopathologic factors including mitotic activity, intra-tumoral necrosis, and tumor cell differentiation [5].

Imaging is essential for the evaluation and treatment of sarcoma. Magnetic resonance imaging (MRI) has been shown to be the best modality to assess tumor size, extent and neurovascular involvement for surgical and radiation therapy planning [6]. Computerized tomography (CT) is the best imaging modality for the detection of pulmonary metastases and used for staging [6]. There has been progressive increasing interest in metabolic and functional imaging of sarcomas. ¹⁸F-Fluorodeoxyglucose (¹⁸F-FDG) is a glucose analog which is taken up rapidly by tumor cells. This allows for identification of tumor cells with high glucose metabolism. ¹⁸F-FDG positron emission tomography/computed tomography (PET/CT) is frequently utilized for sarcoma staging, surveillance, and to aid in biopsy guidance [1]. The maximum standardized uptake value (SUV) on ¹⁸F-FDG PET/CT studies has been correlated with sarcoma grade [7-9] and overall survival [10-12]. ¹⁸F-FDG PET/CT has also been shown to be clinically useful for differentiating benign from malignant peripheral nerve sheath tumors [13]. More recent reports suggest that ¹⁸F-FDG PET/CT may also clinically differentiate enchondromas from chondrosarcomas [14]. The significant clinical success of ¹⁸F-FDG PET/CT has resulted in significant interest in developing other PET tracers that can assist in the evaluation and treatment of patients with sarcoma.

Alternative PET/CT tracers are now being developed to take advantage of non-glucose metabolites to reveal information about sarcoma tumor biology that may have diagnostic or prognostic implications. Consequently, many PET/CT tracers are currently in development and have been studied in the context of bone and soft tissue sarcoma. The purpose of this review is to identify preclinical PET/CT tracers that may be used in future for sarcoma imaging, and to evaluate what we have learned from using these tracers.

Literature search

Articles were identified to be included this review by searching in PubMed for “(‘positron emission’ OR PET) AND (sarcoma) AND ([TRACER])”. Analogous searches were performed using Google Scholar, Ovid Medline, and Thomson’s Web of Science. Only articles written in English were assessed and only new PET tracers used between 01/01/2000 and 06/01/2018 were included. Studies with human subjects were included based on relevance to sarcoma clinical practice. Figure 1 shows PRISMA diagram of how studies were selected for the review. Table 1 shows novel PET tracer clinical characteristics.

Figure 1.

PRISMA diagram. Generated using the PRISMA Flow Diagram Generator by Toronto Health Economics and Technology Assessment Collaboration (http://prisma.thetacollaborative.ca/).

Table 1.

Characteristics of PET tracers for sarcoma imaging

Tracer	Analogue	Half-life	Clinical role
11C-thymidine	Nucleic acid (thymidine)	20.4 minutes	Superior tumor-to-inflammation ratio
18F-FLT	Nucleic acid (thymidine)	110 minutes	Estimation of tumor proliferation, grade, mitotic index, and response to therapy
11C-methionine	Amino acid (methionine)	20.4 minutes	Superior tumor-to-inflammation ratio, estimation of response to therapy
18F-FPMET	Amino acid (methionine)	110 minutes	Superior tumor-to-inflammation ratio
18F-FPro	Amino acid (proline)	110 minutes	Accumulates in osteosarcoma in an animal model
11C-tyrosine	Amino acid (tyrosine)	20.4 minutes	Superior tumor-to-inflammation ratio, estimation of tumor PSR, grade, and mitotic index
18F-FPT	Amino acid (tyrosine)	110 minutes	Low tracer uptake in inflammation
18F-FMISO	Hypoxia (nitroimidazole)	110 minutes	Limited data evaluating correlation between avidity and tissue pO2
18F-EF5	Hypoxia (nitroimidazole)	110 minutes	Detection of tumor hypoxia
18F-FAZA	Hypoxia (nitroimidazole)	110 minutes	Detection of tumor hypoxia, estimation of radiotherapy response

Nucleoside analogues

¹¹C-thymidine

¹¹C-thymidine is a nucleoside analogue which is taken up and directly incorporated into the DNA of cancer cells [15]. ¹¹C-thymidine has a half-life of 20.4 minutes [16,17], and the metabolic product ¹¹CO₂ is excreted through exhalation [18]. ¹¹C-thymidine avidity is a direct measure of tumor cell DNA synthesis because it is incorporated into DNA [15]. This allows this PET tracer to assess how treatment including chemoradiation affects tumor DNA synthesis [15] and to differentiate tumor from non-replicative inflammatory cells [15]. ¹¹C-thymidine has been investigated as a component of multi-agent PET to estimate sarcoma grade. In a small pilot trial, 10 patients with soft tissue sarcoma (STS) were imaged with ¹¹C-thymidine to visualize cellular proliferation, and these results were compared with histological grade [19]. No correlation was found between ¹¹C-thymidine avidity and sarcoma grade; however, thymidine flux values ranged from 0.0001 to 0.147 in different sarcoma types [19]. This suggests variation in sarcoma cellular proliferation between different sarcomas, which may have implications in treatment [19]. To assess the potential of ¹¹C-thymidine to measure tumor response to therapy, Shields et al. imaged two patients with high grade sarcoma before and after the initiation of chemotherapy using ¹⁸F-FDG PET and ¹¹C-thymidine PET [15]. In one sarcoma patient, the steady state thymidine incorporation flux rate constant declined to 0%, and this corresponded with a 58% decrease in the ¹⁸F-FDG metabolic rate [15]. This patient experienced complete resolution of the primary tumor completely following chemotherapy. The second sarcoma patient experienced only a 3% reduction in thymidine incorporation, and the metabolic rate of ¹⁸F-FDG increased 85% after treatment; this patient rapidly died of progressive disease [15]. This proof of concept study indicated that ¹¹C-thymidine may be a target for future research into sarcoma response to therapy imaging.

¹⁸F-Fluoro-3’-deoxy-L-thymidine

¹⁸F-Fluoro-3’-deoxy-L-thymidine (¹⁸F-FLT) is another radiolabeled thymidine nucleoside analogue which can be used in place of ¹¹C-thymidine. ¹⁸F-FLT is taken up by sarcoma cells and incorporated into the thymidine salvage pathway [20]. Once ¹⁸F-FLT passively diffuses into the cell [21], it is phosphorylated by the thymidine kinase 1 enzyme (TK1) and trapped in the cell [20]. TK1 activity increases 10-fold when cells undergo DNA synthesis [22], resulting in differential ¹⁸F-FLT uptake between proliferating and non-proliferating cells. ¹⁸F-FLT-monophosphate can further be phosphorylated to ¹⁸F-FLT-triphosphate, but it is not incorporated into DNA due to the substitution of the 5’ OH in thymidine by fluorine [17,20]. Consequently, ¹⁸F-FLT PET/CT uptake is an indirect measure of tumor proliferation in cells which undergo DNA replication using the thymidine salvage pathway, although it is a poor measure of proliferation in cells undergoing de novo thymidine synthesis [20]. ¹⁸F-FLT PET/CT has advantages over ¹¹C-thymidine. ¹¹C-thymidine requires an onsite cyclotron, limiting its widespread use; however this is not necessary for ¹⁸F-FLT [22]. This is because ¹⁸F has a half-life of 110 minutes [22], whereas the radiolabeled carbon in ¹¹C-thymidine has a much shorter half-life of 20 minutes and is rapidly metabolized in vivo [22,23]. Consequently, ¹¹C-thymidine PET/CT produces lower quality images which impairs proliferation rate calculations when compared with ¹⁸F-FLT PET/CT [22].

Cobben et al. assessed the use of ¹⁸F-FLT PET/CT for detection and grading of STS [24]. In a trial of 19 patients with 20 histologically-confirmed STS tumors, all 20 malignant lesions were detected with full body ¹⁸F-FLT PET/CT prior to treatment [24]. However the sensitivity of ¹⁸F-FDG PET/CT for detecting these sarcomas ranged from 88-92% [24]. Buck et al. imaged 22 patients with established or suspected soft tissue or bone tumors using both ¹⁸F-FLT and ¹⁸F-FDG PET for pretreatment planning [25]. Seventeen tumors were histologically confirmed to be malignant bone or soft tissue sarcomas, and five tumors were confirmed to be benign. Using a threshold cutoff value of 2.0 for mean SUV, all 17 malignant tumors were identified by ¹⁸F-FLT PET (100% sensitivity) and three out of four benign tumors were classified correctly (75% specificity) [25]. However, using this same threshold, one benign lesion was incorrectly classified as malignant by ¹⁸F-FLT PET.

¹⁸F-FLT PET also compares favorably with ¹⁸F-FDG PET as a noninvasive tool to predict sarcoma cell proliferation. In a study of eight RIF-1 sarcoma bearing mice treated with saline or with a single 5 mg/kg dose of cisplatin, normalized uptake of ¹⁸F-FLT uptake was linearly correlated with proliferating cell nuclear antigen labeling index (r=0.89, P=0.001), indicating that ¹⁸F-FLT PET avidity is a measure of cell proliferation in vivo [26]. Studies in human patients with sarcoma have also indicated that ¹⁸F-FLT PET uptake correlates with histopathologic grade. Buck et al. identified a significant correlation between mean ¹⁸F-FLT SUV and histopathologic grade (r=0.71, P=0.01) in 22 human patients with bone and soft tissue sarcoma [25]. Mean ¹⁸F-FLT SUV was significantly elevated in high grade sarcomas when compared to low grade sarcomas (mean=6.1, range=2.5-8.3 vs. mean=1.6, range=1.4-1.8; P=0.001) [25]. However, Buck et al. found no correlation between mean ¹⁸F-FDG SUV_max or SUV_mean and sarcoma histologic grade [25]. Another prospective study of 10 patients with locally advanced, nonresectable STS identified a significant correlation between sarcoma mitotic index and ¹⁸F-FLT SUV_max (r=0.87, P=0.001) [17]. Similarly, among 20 histology-confirmed STS in 19 patients, there were significant positive correlations between¹⁸F-FLT SUV_max and mitotic score (r=0.721, P<0.05), ¹⁸F-FLT SUV_max and histologic grade (r=0.724, P<0.05), and correlation between tumor-to-nontumor ratio and histologic grade (r=0.747, P<0.05) [24].

¹⁸F-FLT PET has also been used to assess sarcoma response to treatment. Benz et al. imaged 20 adult human patients with biopsy-proven, high grade, resectable STS using ¹⁸F-FLT PET/CT before and after neoadjuvant therapy [27]. ¹⁸F-FLT SUV_max decreased from a pretreatment mean of 7.1±3.7 g/mL to 2.7±1.6 g/mL after treatment (P<0.001), indicating a treatment response, and changes in peak FLT uptake were correlated with changes in tumor size (r=0.52, P=0.02) and tumor necrosis (r=0.68, P=0.001) [27]. However, there was no correlation between ¹⁸F-FLT uptake and TK1 expression (r=0.26, P=0.27) or Ki-67 activity (r=0.29, P=0.21) [27]. Figure 2 shows pretreatment and posttreatment ¹⁸F-FLT PET images from a histologic responder and a histologic non-responder. ¹⁸F-FLT PET can also be used as an early marker of sarcoma response to chemotherapy [26]. Treatment of 8 mice with radiation-induced fibrosarcoma with cisplatin resulted in significant decreases in tumor cell proliferation as measured by a decrease in proliferating cell nuclear antigen labeling index from 14.0%±2.0% to 6.2%±1.0% (P=0.001); this reduction in proliferation was detected by a simultaneous decrease in normalized ¹⁸F-FLT uptake in these animals from 0.76±0.08 to 0.51±0.08 (P=0.03) [26]. ¹⁸F-FLT PET has also been demonstrated to predict response to therapy in human patients with sarcoma. Van Ginkel et al. report that ¹⁸F-FLT SUV_max decreased from a mean of 3.5 to 1.7 (P=0.008) and that SUV_mean decreased from 1.9 to 0.8 (P=0.002) among 10 patients treated with locally advanced STS treated with hyperthermic isolated limb perfusion, TNF-α, and melphalan [17]. Furthermore, there was a correlation between pretreatment ¹⁸F-FLT SUV_mean and necrosis following treatment (r=0.642, P<0.05) [17].

Figure 2.

“Baseline and follow-up ¹⁸F-FLT positron emission tomography/computed tomography images are shown of 2 patients with angiosarcoma of the breast. The patient shown in (A) exhibited >95% tissue necrosis after neoadjuvant treatment and was classified as a histopathologic responder. The patient in (B) was a histopathologic nonresponder with <5% tissue necrosis after treatment. However, comparable decreases in ¹⁸F-FLT uptake of 85% and 71%, respectively, were seen. Therefore, the patient in (B) was misclassified as a metabolic responder by fluorothymidine positron emission tomography analysis. SUV_peak indicates peak standardized uptake value”. Reprinted, with permission, from reference [27].

¹⁸F-FLT PET also potentially has a role in predicting response to targeted therapies. The driving mutation in 85% of Ewing sarcomas is the EWS-FLI1 translocation, which promotes the expression of multiple genes including ENT1, ENT2, and TK1 [28]. ENT1 and ENT2 permit passive transport of ¹⁸F-FLT into tumor cells, and TK1 phosphorylates and traps the radiotracer in the cell [28]. Osgood et al. found that repression of EWS-FLI1 in Ewing sarcoma cell lines with small molecule inhibitors reduced mRNA expression of ENT1, ENT2, and TK1 in vitro; in contrast, chemotherapeutic agents like 5FU had no effect on TK1 expression [28]. Mice with Ewing sarcoma xenografts were treated with the small molecule inhibitors mithramycin (1 mg/kg), EC-8042 (24 mg/kg), or EC-8105 (1.5 mg/kg), and were subsequently imaged with ¹⁸F-FLT PET both before and after therapy [28]. All three agents resulted in significant suppression of ¹⁸F-FLT activity (P<0.05), but no suppression was evident after treatment with 5-FU. This suggests that ¹⁸F-FLT PET could potentially demonstrate susceptibility and response to targeted EWS-FLI1 therapy for Ewing sarcoma.

Amino acid analogues

¹¹C-methionine

¹¹C-methionine is a radiolabeled amino acid with a half-life of 20.4 minutes [16,17] and has been investigated as a PET tracer for sarcoma. Its half-life limits its current clinical use. Sarcoma tumor cells upregulate amino acid transport, transmethylation rate, and protein synthesis [29,30]. Radiolabeled amino acids such as ¹¹C-methionine are absorbed and incorporated into proteins and can serve as a marker of protein synthesis using PET imaging [30]. ¹¹C-methionine was superior to ¹⁸F-FDG for differentiating fibrosarcoma from inflammatory background in a mouse model [30]. S180 fibrosarcoma-bearing mice (n=4) were administered full-body PET/CT imaging with either N-(2-[¹⁸F]Fluoropropionyl)-L-methionine (FPMET), FDG, or ¹¹C-methionine [30]. The tumor-to-inflammation ratio of ¹¹C-methionine was 1.64; this was superior to the tumor-to-inflammation ratio of ¹⁸F-FDG at 1.14, although the sample size were too small for a robust statistical analysis [30].

¹¹C-methionine may be able to predict outcomes in patients after carbon ion radiotherapy (CIRT) [29]. Sixty-two patients with histologically-confirmed bone or soft tissue sarcomas which were inoperable or who had declined surgery, were imaged using ¹¹C-methionine PET/CT before and up to one month following treatment with carbon ion radiotherapy (CIRT) [29]. Post treatment, ¹¹C-methionine tumor-to-nontumor (T/N) ratio decreased from a mean of 4.58±2.57 to 3.11±2.04 (P=0.00029) [29]. Pretreatment ¹¹C-methionine uptake was predictive of outcome-patients with a pretreatment (T/N) ratio <6 demonstrated significantly higher 2-year survival when compared with patients with higher T/N ratios (69.4% vs. 32.3%, P=0.010). Posttreatment ¹¹C-methionine uptake also had prognostic implications. Patients with a posttreatment T/N ratio of less than 4.4 had better 2-year survival than patients with higher T/N ratios (63.7% vs. 41.3%, P=0.012). A reduction in ¹¹C-methionine T/N ratio of greater than 30% significantly improved 2-year survival (74.6% vs. 41.6%, P=0.049) [29]. The authors concluded that post-CIRT ¹¹C-methionine PET predicted response to therapy and 2-year survival, better than tumor type, size, or stage [29]. However, a separate study found that pretreatment ¹⁸F-FDG PET was superior to ¹¹C-methionine for prediction of STS response to neoadjunctive therapy [31]. Nine patients with STS were imaged with ¹⁸F-FDG and ¹¹C-methionine PET/CT before and after neoadjuvant chemoradiotherapy [31]. Using a cutoff of 45% SUV_max, ¹⁸F-FDG PET was able to distinguish between partial and complete responders to therapy (complete responders defined by the authors as 91-100% tumor necrosis on histology post-therapy) [31]. In contrast, it was not possible to differentiate partial and complete responders using change in SUV_max with ¹¹C-methionine [31].

N-(2-[¹⁸F]-fluoropropionyl)-L-methionine

¹⁸F-FPMET is an ¹⁸F-labeled radiolabeled amino acid with a half-life of 110 minutes [22] that has been used to measure tumor amino acid uptake. ¹⁸F-FPMET differs from ¹¹C-methionine because it not incorporated into protein in S180 mouse fibrosarcoma models [30]. Methionine uptake is upregulated in sarcomas, therefore ¹⁸F-FPMET should preferentially be taken into tumor cells via sodium-dependent transporters as a result [29,30]. Preliminary evidence indicates that ¹⁸F-FPMET PET has a comparable ability to distinguish tumor from inflammatory background as ¹¹C-methionine PET [30]. Among mice with fibrosarcoma imaged with PET/CT using either ¹⁸F-FPMET or ¹¹C-methionine, ¹⁸F-FPMET demonstrated a tumor-to-inflammation ratio of 1.64 compared to 1.62 for ¹¹C-methionine [30]. However, ¹⁸F-FPMET has yet to be studied in the clinical setting, possibly because it is a measure of amino acid uptake rather than protein incorporation.

Cis-4-[¹⁸F]-fluoro-L-proline

Cis-4-[¹⁸F]-Fluoro-L-proline (¹⁸F-FPro) is an amino acid analog. Sarcomas are mesenchymal tumors, so they have been hypothesized to highly express collagen. Proline is a major component of collagen; therefore sarcomas would be expected to demonstrate high levels of ¹⁸F-FPro uptake and protein incorporation [32]. In vivo mouse models of osteosarcoma indicate that cis-¹⁸F-FPro accumulates within osteosarcoma tumor cells [33]. Transplanted osteosarcoma tumors in mice demonstrated a maximum tumor cis-¹⁸F-FPro uptake of 11.9% injected dose per gram (ID/g) (n=7, sd=2.19) four hours after cis-¹⁸F-FPro infusion, whereas blood pool demonstrated an uptake of 2.86% ID/g (n=7, sd=0.29) [33]. High cis-¹⁸F-FPro uptake was also present in liver the (9.77% ID/g, sd=0.85), kidney (46.9% ID/g, sd=3.2), and pancreas (19.0% ID/g, sd=1.83) at the same time point [33]. Cis-¹⁸F-FPro uptake within sarcomas largely reflected amino acid transport rather than protein incorporation as only 33±7% of cis-¹⁸F-FPro activity was protein-bound after one hour [33]. Stoeffuls et al. imaged eight human patients with peripheral tumors, including one patient with Ewing sarcoma, using cis-¹⁸F-FPro and ¹⁸F-FDG PET. Cis-¹⁸F-FPro uptake in each tumor was significantly lower than the ¹⁸F-FDG SUV (1.7±0.6 vs. 5.7±3.0, P=0.01), and the Ewing sarcoma lesion exhibited a similar level of uptake as the liver and therefore was only faintly visible using cis-¹⁸F-FPro [32].

L-[1-¹¹C]-tyrosine

¹¹C-tyrosine is a radiolabeled amino acid that is incorporated into protein by tumor cells and has a half-life of 20 minutes [34]. ¹¹C-tyrosine has been investigated as a PET tracer to measure protein synthesis rate (PSR) in STS [5]. ¹¹C-tyrosine PET detects protein synthesis rather than glucose metabolism. ¹¹C-tyrosine PET in theory should distinguish local inflammation from viable sarcoma, because protein synthesis is relatively low in many inflammatory cells such as neutrophils when compared with sarcomas [5,33,35]. ¹¹C-tyrosine should be able to identify proliferative tumors in highly metabolically-active tissues (such as brain) that use glucose by detecting protein synthesis [34]. Protein production increases throughout the cell cycle, so elevated PSR in tumors may be indicative of tumor proliferation [5]. ¹¹C-tyrosine uptake is related to histopathologic measures of proliferation [5]. Twenty-one patients with untreated, histologically-diagnosed STS were imaged using ¹¹C-tyrosine to estimate PSR [5]. Sarcoma tissue was harvested by incisional biopsy of viable tumor within 8 weeks to assess histological grade, differentiation, mitotic index, and Ki-67 expression (see Figure 3) [5]. No correlation was found between sarcoma grade and the maximum protein synthesis rate (PSR_max) or average protein synthesis rate (PSR_ave) values derived from ¹¹C-tyrosine imaging, possibly because increased tumor necrosis, which increases histologic grade, lowers the PSR [5]. However, there was a significant correlation between PSR_ave and mitoses (R=0.67, P<0.001), PSR_ave and Ki-67 proliferative index (R=0.54, P<0.05), PSR_max and mitoses (R=0.64, P<0.01), and PSR_max and Ki-67 proliferative index (R=0.54, P<0.05) [5].

Figure 3.

(A) “¹¹C-tyrosine PET image of a sarcoma not otherwise specified with a relatively high PSR and (B) corresponding MIB-1 stained tissue section (400×) of the tumor showing a high number of proliferating cells. All dark nuclei express Ki-67 and have entered the cell cycle. (C) ¹¹C-tyrosine PET image of a malignant schwannoma with a relatively low PSR and (D) corresponding MIB-1 stained tissue section (400×) showing a low number of proliferating cells”. Reprinted, with permission, from reference [5].

¹¹C-tyrosine has also been assessed as a marker of tumor response to treatment. Van Ginkel et al. treated 12 patients with locally advanced, biopsy-proven STS with neoadjuvant rTNF-a and melphalan therapy using hyperthermic isolated limb perfusion techniques [35]. Patients were imaged with ¹¹C-tyrosine PET prior to limb perfusion, two weeks after completion of neoadjuvant treatment, and at 8 weeks post therapy and PSR was calculated [35]. There was no significant difference in pretreatment PSR in partial pathologic response patients (defined as viable tumor at resection) and complete response patients (defined as no viable tumor at resection) [35]. However, patients with complete responses demonstrated significant decreases in PSR pretreatment at week 2 (P<0.05) and at week 8 (P<0.05) posttreatment, suggesting that PSR estimated from ¹¹C-tyrosine PET can assess response to therapy [35].

O-(3-[¹⁸F]-fluoropropyl)-L-tyrosine

O-(3-[¹⁸F]-fluoropropyl-L-tyrosine (¹⁸F-FPT) is another ¹⁸F-labelled amino acid radiotracer which has been investigated for PET imaging of sarcoma. Tang et al. demonstrated that ¹⁸F-FPT PET is able to detect tumor and to differentiate tumor from inflammation in a mouse model of fibrosarcoma [36]. Fifteen mice were inoculated with S18 fibrosarcoma, and fifteen mice were inoculated with Staphylococcus aureus as a model of inflammation [36]. Animals were imaged using ¹⁸F-FPT PET and ¹⁸F-FDG PET 7-14 days after fibrosarcoma inoculation or 72 hours after S. aureus inoculation [36]. ¹⁸F-FPT uptake is not increased by inflammation; there was no significant difference in ¹⁸F-FPT uptake between the inflammatory abscess and the contralateral muscle in S. aureus-infected mice [36]. By contrast, the ratio of ¹⁸F-FDG uptake between inflammatory abscesses and contralateral muscle was 4.0:1 (P<0.05) [36]. Fibrosarcoma tumors also demonstrated significant ¹⁸F-FPT uptake; the differential uptake ratio (DUR) of fibrosarcoma tissue at 90 minutes was 0.98±0.19, whereas the mean DUR of contralateral muscle in these animals was 0.34±0.10 (P<0.05) [36].

Hypoxia pet radiotracers

Tumor hypoxia is an important factor in sarcoma prognosis because it is associated with poor response to radiotherapy [37,38], higher local recurrence rate, higher rate of developing pulmonary metastases and death [37,39]. Hypoxia within sarcomas leads to activation of hypoxia-inducible factors that upregulate the pathways responsible for tumor angiogenesis and metastasis [40]. Eppendorf electrodes placed directly into tissues to measure the partial pressure of oxygen (pO₂) are the gold standard for measuring tissue hypoxia; however, this technique is invasive and can only measure one region within a tumor [41]. Hypoxia imaging allows for non-invasive global measurement of sarcoma hypoxia [41].

¹⁸F-fluoromisonidazole (¹⁸F-FMISO) is a radiolabeled nitroimidazole, which preferentially accumulates in hypoxic cells [37]. In preclinical trials, ¹⁸F-FMISO PET demonstrated superior tumor-to-inflammation ratios than ¹⁸F-FDG PET in a KHT-sarcoma-bearing and inflammation-bearing mouse models [42]. Intratumor hypoxic volumes determined by ¹⁸F-FMISO PET are highly correlated with hypoxic volume demonstrated by pimonidazole-immunohistochemistry in rat models of rhabdomyosarcoma [43]. ¹⁸F-FMISO PET has also been used to identify tumor hypoxia in spontaneous canine bone and soft tissue sarcomas (confirmed by Eppendorf needle electrode measurement of pO₂) [44]. ¹⁸F-FMISO PET imaging in human patients with sarcoma has been less promising. A trial of 18 patients, including nine with STS, found no correlation between ¹⁸F-FMISO PET tumor-to-muscle ratios and Eppendorf pO₂ measurements [45]. A separate study involving six patients with histologically-confirmed STS also found no correlation between ¹⁸F-FMISO PET tumor-to-muscle ratio and Eppendorf measurements [46]. Although three tumors demonstrated extensive hypoxia in that study, only one demonstrated accumulation of ¹⁸F-FMISO [46]. A third study of 19 patients with STS found no correlations between pretreatment ¹⁸F-FMISO PET-derived hypoxic volume and tumor size, grade, ¹⁸F-FDG PET SUV_max, or VEGF expression (see Figure 4) [47].

Figure 4.

“¹⁸F-FDG and ¹⁸F-FMISO scans (with an attenuation scan for localization) of a grade 2 osteosarcoma of the right femur. The ¹⁸F-FMISO image shows pixels with a tissue to blood ratio ≥1.2 in color. The correlation of ¹⁸F-FDG vs ¹⁸F-FMISO using pixel-by-pixel analysis was r=0.48 for the whole tumor”. Reprinted, with permission, from reference [47].

2-(2-nitro-1H-imidazol-1-yl)-N-(2,2,3,3,3-¹⁸F-pentafluoropropyl)-acetamide (¹⁸F-EF5) is a radiolabeled hypoxia tracer with contains a 2-nitroimidazole moiety, permitting selective uptake in hypoxic cells [48]. Unlabeled EF5 has been well established as an immunohistochemical marker of hypoxia [48,49], and it has been associated with the presence of mitoses and higher histologic grade in human patients with STS [49]. ¹⁸F-EF5 has been tested in humans and was determined to be safe with an acceptable level of radiation exposure [50], and ¹⁸F-EF5 PET has been validated for the assessment of hypoxia in head and neck cancers [51]. This tracer has not yet been studied in human sarcomas, but ¹⁸F-EF5 PET detected hypoxia in one of two cats with histologically-confirmed fibrosarcomas [52].

¹⁸F-1-(5-fluoro-5-deoxy-α-D-arabinofuranosyl)-2-nitroimidazole (¹⁸F-FAZA) is another radiolabeled nitroimidazole derivative which may be used in place of ¹⁸F-FMISO. An advantage of ¹⁸F-FAZA over ¹⁸F-FMISO is faster washout from tissues [53,54]; however, the tradeoff is lower tracer uptake by tumors and a lower sensitivity of ¹⁸F-FAZA PET for the detection of hypoxic regions in rat carcinosarcoma models [54]. ¹⁸F-FAZA PET accurately reflected tumor hypoxia in rat models of rhabdomyosarcoma. ¹⁸F-FAZA PET tumor-to-blood ratios were strongly correlated with pO₂ as determined by Eppendorf oximetry (r=0.93) [55], and ¹⁸F-FAZA luminescent activity was also correlated with pimonidazole (a histological marker of hypoxia) fluorescence [53]. ¹⁸F-FAZA may be able to guide therapy decisions in the future. Mice with hypoxic sarcomas with higher tumor-to-blood ratios benefited from combined treatment of radiotherapy and nimorazole (a hypoxic radiosensitizer that generates free radicals to produce radiation damage), whereas this benefit was not seen in mice with less hypoxic sarcomas [56,57]. ¹⁸F-FAZA PET can also guide decisions about the use of carbogen breathing and dose escalation in radiotherapy in mouse models of rhabdomyosarcoma [41].

Discussion

Nucleoside analogs, amino acid analogs, and hypoxia PET tracers are currently undergoing evaluation in preclinical studies as potential complements to traditional ¹⁸F-FDG sarcoma imaging.

¹⁸F-labelled compounds are probably the most likely to be used in clinical practice in future because of the relatively long half-life of ¹⁸F. ¹⁸F-FLT is likely the best radiotracer to evaluate DNA synthesis in sarcoma. ¹⁸F-FPT is probably the best radiotracer to study protein synthesis in sarcoma given the preclinical performance and relatively long half-life. All three hypoxia tracers have various merits. ¹⁸F-EF5 has the additional benefit of being able to be directly correlated with uptake from cold EF5. However, ¹⁸F-EF5 synthesis is more complicated than that for ¹⁸F-FMISO. An advantage of ¹⁸F-FAZA over ¹⁸F-FMISO is faster washout from tissues however, this results in lower tracer uptake by tumors and a lower sensitivity of ¹⁸F-FAZA PET for the detection of hypoxic regions.

What we have learned is that there is significant heterogeneity within and between sarcomas. DNA synthesis, measured using ¹¹C-thymidine or ¹⁸F-FLT, showed significant differences between sarcomas; and the change in DNA synthesis after treatment could predict change in tumor size, necrosis and patient clinical course. DNA synthesis was also correlated with sarcoma grade and mitotic index. However, DNA synthesis was not correlated with cellular proliferation measured by Ki-67 activity.

In contrast, protein synthesis was not correlated with sarcoma grade. Protein synthesis was correlated with increased mitotic figures, and cellular proliferation measured by Ki-67 activity. Change in protein synthesis was also correlated with response to therapy and predicted overall survival better than tumor size, grade or tumor stage. The previously published reports suggest that not all DNA synthesis results in protein synthesis; however, both DNA and protein synthesis were predictors of outcome in patients with sarcoma. It is unclear how strongly protein synthesis is correlated with collagen synthesis. Collagen synthesis measured by cis-¹⁸F-FPro uptake varied between sarcoma subtype with significant uptake in osteosarcoma and uptake similar to background in Ewing’s sarcoma.

Like protein synthesis, hypoxia was not correlated with tumor grade, and was not correlated with tumor size, glycolysis or VEGF. This suggests that hypoxia imaging provides information about the tumor biology beyond that captured by ¹⁸F-FDG PET/CT. This also suggests that DNA synthesis seems independent of sarcoma hypoxia. Hypoxia imaging has been limited, however, by poor correlation between ¹⁸F-FMISO avidity and Eppendorf pO₂ measurements. ¹⁸F-FAZA and ¹⁸F-EF5 are promising new hypoxia radiotracers, and preclinical evidence has demonstrated correlation between EF5 and FAZA deposition with histologic and Eppendorf measurements of hypoxia respectively. The presence and extent of hypoxia imaging may allow clinicians to predict response to radiotherapy and thereby guide clinical decision making.

Radiolabeled nanoparticles may provide a new direction for investigation in sarcoma imaging. Nanoparticles are self-assembling inorganic compounds which can accumulate in tumors by passing through aberrant, more permeable blood vessels within tumors or by directly binding tumor targets [58]. Radiolabeled nanoparticle PET has not been assessed in sarcoma; however, this technology has been investigated for tumor detection and for assessment of angiogenesis [59] and may provide clinically useful information about tumor perfusion.

The ability to quantitatively assess tumor DNA synthesis, tumor protein synthesis, and tumor hypoxia in vivo will help clinicians better understand underlying sarcoma biology. Several of these tracers have shown already shown promise in the preclinical setting to predict sarcoma prognosis and response to therapy. These targeted PET tracers have the potential to be increasingly used in the era of precision and personalized medicine. The use and development of targeted therapies for sarcoma has been limited due to the rarity and heterogeneity of these tumors. These tumors are very heterogeneous and sometimes subject to sampling error from biopsies. Imaging tracers have the potential to transform clinical management, because imaging can provide global characterization of sarcomas and are not subject to sampling error. The additional information from these novel tracers will better inform patient treatment and in the future ultimately improve patient outcomes.

Disclosure of conflict of interest

None.