Despite extensive methodological development, most available techniques for the assessment of hypoxia have disadvantages. The use of polarographic oxygen electrodes is limited to accessible tumours and immunohistochemical techniques for identification of reduced nitromidazoles, such as pimonidazole, require biopsy. These two methods are invasive and consequently, the development and validation of non-invasive techniques, based on imaging is highly desirable (4). At present, positron emission tomography (PET) is the best validated imaging approach for imaging of hypoxia, but a number of Magnetic Resonance Imaging (MRI) based techniques have also been described (4, 40) (Tab. 3).
An optimal PET radiopharmaceutical to image hypoxia should selectively target clinically relevant hypoxic cells (pO2 = 0-10 mmHg), be lipophilic to guarantee fast intracellular transport and sufficiently hydrophilic to allow rapid clearance from normoxic cells in order to provide a high target-to-background ratio (TBR) within a short time from injection (4, 101). A wide range of PET tracers for imaging hypoxia has been described. These can be broadly divided into two main groups: the 2-nitroimidazole family and non-nitroimidazole compounds (mainly copper-complexed dithiosemicarbazone – [60,61,62,64CU]ATSM – derivatives) (102).
At the present time, the most widely used radiotracer, belonging to the 2-nitroimidazole-based, is [18F]fluoromisonidazole ([18F]FMISO), which passively diffuses into cells due to its lipophilic nature (102). Under hypoxic conditions, with pO2â€‰≤ 10 mmHg, the radiocompound undergoes electron reductions and forms reactive radicals. These bind covalently to intracellular macromolecules; so that the tracer is trapped inside hypoxic cells. In oxic cells, the tracer is reoxygenated and converted to the original form, passing out through the cell walls (103).
[18F]FMISO has been assessed and validated in a wide range of human tumours, showing high uptake in hypoxic cells (104). However, this tracer presents some drawbacks including slow clearance kinetics from non-hypoxic tissue resulting in high background in PET images (105). Early images with [18F]FMISO, within 5 minutes after injection, are usually acquired allow estimation of delivery, whereas delayed images, taken about 2-2.5 hours after injection, following washout of non-bound tracer, provide estimation of hypoxia. A semi-quantitative analysis is usually performed estimating tumour-blood and tumour-muscle ratios (7, 101, 104). It is important to select the appropriate time-windows for imaging in order to distinguish radiotracer retention from perfusion, which varies between types of tumours and individuals (4). Some authors found calculation of the tumour hypoxic volume (HV) (106) or of the fractional hypoxic volume (FHV) useful as an indicator of the severity and extent of tumour hypoxia (106, 107). The measurement of HV, which is the volume of the tumour that is hypoxic, requires the definition of an uptake threshold, while FHV is the percentage of pixels with values greater than a tumour-to-blood ratio (T/B) cut-off and requires in addition the estimation of the total tumour volume (4, 7). According to the available literature, the optimal cut-off of tumour-to-blood ratio (T/B) for distinguishing between normoxia and hypoxia ranges between 1.2 and 1.4 at 2-2.5 hours (108, 109). Some authors have claimed that quantifying hypoxia by simply using semi-quantitative parameters may not be reliable because severely hypoxic and necrotic areas show low uptake and can be disregarded by a SUV threshold identiï¬cation. In addition, the spatial disconnection between hypoxic regions and well perfused vessels may cause long diffusion time of the radiotracer (103, 110). Thus, it has been suggested that a bi-compartmental kinetic model may be necessary to compensate for hypoxic trapping of the tracer together with the effects of perfusion and of diffusion in interstitial space (110).
In the attempt to overcome the limitations of [18F]FMISO, new generations of radiotracers for the detection of tumour hypoxia have been developed. Among them, the most promising is probably [F-18]Fluoroazomycinarabinofuranoside ([18F]FAZA), which is another radiolabeled 2-nitroimidazole that is less lipophilic than [18F]FMISO: consequentially non-specific uptake is expected to be washed out more quickly (111, 112). Additionally, [18F]FAZA exhibits further excellent characteristics such as negligible cell-to-cell line variability and no binding in oxic cells (113). [18F]FAZA has also been compared to [18F]FMISO in biodistribution investigations performed by Piert et al. in three different tumour models (EMT6 tumour bearing BALB/c mice = mammary carcinoma, AR42J tumour-bearing mice = pancreatic carcinoma and A431 tumour-bearing Swiss nude mice = squamous carcinoma) (112). At 3 h after tracer injection(112) Piert and colleagues showed that [18F]FAZA present significantly lower dosimetric values in the body and a faster clearance from the blood than [18F]FMISO. On the other hand, tumour uptake of the tracers did not differ significantly in two of the three tumour models, but [18F]FAZA demonstrated superior tumour-to-background ratio (TBR) (112). Instead, Peeters and co-workers compared [18F]FAZA with [18F]FMISO and [18F]HX4 (another lipophilic PET radiotracer for the detection of hypoxia) in rhabdomyosarcoma R1-bearing WAG/Rij rats documenting a maximal and stabilized T/B at 3 hours for [18F]HX4 (7.2 ± 0.7) and at 2 hours for [18F]FAZA (4 ± 0.5). Noteworthy, already at 2 hours [18F]HX4 exhibited a T/B higher or equal to that of the other two radiotracers, whereas TBR for [18F]FMISO was still increasing at 6 hours (9 ± 0.8) and the tracer did not demonstrate plateau formation at late time-point (6 hours) (114).
PET imaging is able to provide quantitative maps of tumour hypoxia that accurately reflect the hypoxic cell density with a clear spatial link between the radiotracer (e.g. [18F]FAZA) uptake and an immunologically detectable hypoxia marker (e.g. pimonidazole) (115). Imaging hypoxia with PET may play a role in RT planning, supporting a higher delivered dose to tumours or regions displaying hypoxia (116, 117). Hypoxia imaging can also predict the success of radio-chemotherapy, as shown by Beck and colleagues using tumour-bearing mice and the hypoxia-activated chemotherapeutic agent tirapazamine. Pre-treatment [18F]FAZA PET may offer a way for the individualization of tumour treatment by focusing hypoxia-directed therapies or more aggressive treatments to tumours with high [18F]FAZA uptake (118).
Hirata first documented the utility of hypoxic tracers ([18F]FMISO in this study) in patients differentiating GBM from lower grade gliomas based on the level of tumour hypoxia (119). Hypoxia assessment by PET imaging seems to provide complementary information to MRI within the complex relationship existing between hypoxia and angiogenesis in GBM. This was confirmed in a study of Swanson et al, where the authors documented a strong correlation between the hypoxic burden, determined with [18F]FMISO, and altered vasculature, documented on Gd-enhanced T1-weighted MRI sequences (120). As with other tumours, the prognostic value of [18F]FMISO has been demonstrated in GBM, in a study evaluating the correlation between hypoxic volume, intensity of hypoxia and survival in 22 patients with GBM who underwent PET scan before biopsy or between resection and RT (121). In this study patients with higher hypoxic severity, assessed by maximum tumour to blood activity ratio (T/Bmax), and greater hypoxic volume (HV), resulted associated with shorter time to progression (TTP) and OS (121). [18F]Fluoroazomycinarabinofuranoside ([18F]FAZA) is another radiotracer tracer, which has showed promising results; unfortunately evaluation of hypoxia in GBM with this radiotracer has been reported only in one clinical study. The biggest study ever published, evaluating the utility of [18F]FAZA in 50 patients with different types of tumours, documented increased uptake of the tracer in all gliomas (n = 7), with a TBR range of 1.9-15.6, which is higher than usually reported with [18F]FMISO (12). According to Wiebe, one important point in favour of [18F]FAZA for the evaluation of hypoxia in brain tumours is the absence of uptake in normal brain tissue, compared to [18F]FMISO, which shows, although limited, non-specific cerebral uptake. The lack of uptake of [18F]FAZA in the normal brain derives from its hydrophilicity, which prevents its cross through the intact blood brain barrier (BBB), and its uptake in GBM, as reported in the study of Postema and co-workers, reflect the disruption of the BBB (12, 101). In the same time, its poor penetration through the BBB limits its use in situations, where BBB is not disrupted (122). However, as opposed to [18F]FMISO, literature on the use of [18F]FAZA in the brain is still limited and is based in a few preclinical studies and only a clinical one (123). Recently, Belloli and colleagues investigated the combined use of [18F]FAZA, [18F]FDG PET and MRI to follow the metabolic, anatomical and hypoxic modifications of specific lines of glioma cells (F98) inoculated (subcortically) in rats. The authors performed weekly PET animal studies and observed that [18F]FAZA and [18F]FDG were taken up respectively in the core and in external areas of the tumour. Interestingly the authors observed a partial overlap of the two tracers in the centre of the Gd-MR positive region and remodelling during disease progression (radial increase of [18F]FAZA distribution towards the periphery of the original injection site and increasing distribution of [18F]FDG towards the growth front of the tumour), suggesting that necrotic regions, detected on the [18F]FDG PET images, may contain hypoxic clusters of tumour cells, which may be revealed with [18F]FAZA PET imaging (124). It is not possible to fully compare the results of Belolli et al. e with other papers investigating the use of [18F]FMISO, because [18F]FAZA has been evaluated in different xenograft models with different biological characteristics. The only brain tumour models that have been investigated with both radiotracers are the 9L (125-128) and U251 glioma models (129, 130). Since these studies have used significantly different methodologies, an accurate comparison is not feasible. Taking into account the 9L model, Tran and co-workers, using [18F]FAZA, documented a mean TBR (at 3 h) of 2.16 ± 0.13 in 10 rats breathing room air (126). In contrast Valable et al. (128) in their study, involving two rat tumour models (C6 and 9L) imaged with dynamic [18F]FMISO PET scan, reported the highest mean TBR (2.5) in the C6 model at 2.5 h, but no uptake in the 9L model, justifying this finding to intrinsic differences between the two models. Indeed, C6 model is less vascularized, and has more severe hypoxia than 9L model (pO2 = 12-14 mmHg vs. 30-32 mmHg for 9L, which is close to the physiological pO2 in the normal brain = 40-45 mmHg) (129). However the findings of Valable and co-workers are not in keeping with those published in a previous study, which documented [18F]FMISO uptake in the 9L model, imaged with a static acquisition at 2 hours post injection (127). Interesting in vitro experiments in different cell lines suggest that anoxic/oxic binding ratios for FMISO can be influenced by other intrinsic factors including glucose level and cell growth state (131), which could have been partially different in the study of Dence and Valable.
Cite This Work
To export a reference to this article please select a referencing style below: