EF5 is a lipophilic 2-nitroimidazole molecule developed for
detection of hypoxia in
(Lord et al., 1993;
Evans et al., 1995;
Koch et al., 1995).
Reduction rate of 2-nitroimidazoles is inversely dependent on oxygen partial pressure
(pO2), but also on levels of nitroreductase enzyme levels
(Koch & Evans, 2015).
-CF2CF3 terminus, which increases lipophilicity and
is chemically and biochemically very stable
(Koch & Evans, 2003).
In a pharmacokinetics study with EF5 in humans, plasma halflife was ∼12 h, and no
metabolites retaining the 2-nitroimidazole chromophore were found in plasma or urine
(Koch et al., 2001).
In mouse tumours EF5 binding increased substantially when pO2 was less than 5 mmHg
(Mahy et al., 2003).
Results by Koch et al (1995) suggest that
EF5 binding continuously changes with pO2 and is already several fold higher at 10 mmHg
than at more physiological oxygen levels.
Severe hypoxia is rare in human and animal tumours: only high-grade tumours tend to contain hypoxic
areas adjacent to necrotic regions where also EF5 binding can be observed
(Evans et al., 2004;
Koch & Evans, 2015).
Head-and-neck tumours are exceptional as hypoxia is present in all tumour grades.
Normal tissues generally are well oxygenated, too, except that skin
has pockets of severe hypoxia, and mild hypoxia can be present in the
(Koch & Evans, 2015).
The EF5 molecule contains five fluorine atoms to increase lipophilicity, and it has been labelled with 18F (Dolbier Jr et al., 2001; Chitneni et al., 2012; Eskola et al., 2012). In rat PET study [18F]EF5 with nonradioactive carrier has shown even distribution in all organs including brain, mainly urinary excretion with some uptake in GI tract (Ziemer et al., 2003). Mice study with high molar activity (low mass) of [18F]EF5 still did show uniform distribution, with lowest uptake in fat, testes, and brain; bone uptake remained low, indicating that no in vivo defluorination occurred (Eskola et al., 2012). In mice, polar radiolabelled metabolites were detected in plasma and muscle, and lipophilic metabolites in liver, tumour xenograft, and urine (Eskola et al., 2012). The fraction of unchanged [18F]EF5 2 h after administration was ∼37% in plasma, ∼38% in muscle, ∼3% in liver, ∼19% in tumour in mice; in a rat the fraction in plasma was 80% (Eskola et al., 2012). In mouse tumour xenografts, uptake of [18F]EF5 correlated with expression of HIF1α (Silén et al., 2014) and with tumour growth rate (Silvoniemi et al., 2014). Immunohistochemistry may require high concentrations of nonradioactive EF5 co-administered with [18F]EF5 (leading to low molar activity), and this may markedly affect the EF5 distribution in some tumour types (Chitneni et al., 2014).
In humans, [18F]EF5 was distributed evenly in soft tissues already minutes after administration and concentrations in blood and soft tissues were almost stable for 4 hours; radioactivity was eliminated via urine and bile, and no radioactive metabolites were detected in blood or urine (Koch et al., 2010). Also Eskola et al (2012) found virtually no 18-labelled metabolites in human plasma even after 4 hours, but in urine metabolite fraction was ∼20% in some individuals. Radiation dose from a [18F]EF5 PET study is clinically acceptable, with urinary bladder receiving the largest radiation-absorbed dose (Lin et al., 2012).
During first minutes the [18F]EF5 images resemble the images obtained using perfusion tracer [15O]H2O; the perfusion-dependent pattern disappeared at later times, also in tumours (Komar et al., 2008). The global head-and-neck tumour uptake did not show significant correlation between [18F]EF5, [15O]H2O, or [18F]FDG 3 h after tracer administrations (Komar et al., 2008). In cervical cancer, the hypoxic location detected with [18F]EF5 was confirmed with higher HIF1α and CA-IX expression (Narva et al., 2021). Biodistribution of [18F]EF5 is about tenfold faster than that of the most used hypoxia tracer [18F]FMISO (Koch, 2018). In head-and-neck cancer, [18F]EF5 had better prognostic value than [18F]FDG (Komar et al., 2014).
Intra-abdominal accumulation prevents using [18F]EF5 in ovarian cancer (Laasik et al., 2020).
Hypoxia-activated prodrugs share similar mechanisms as 2-nitroimidazoles. [18F]EF5 can be used to monitor early treatement response to these drugs (Chitneni et al., 2013).
While most [18F]EF5 studies have been aimed at detecting hypoxia in tumours, imaging hypoxia in other tissues is also possible. Increased [18F]EF5 uptake has been seen in large atherosclerotic plaques (Silvola et al., 2011).
Since the blood and tissue curves reach a fairly stable level 60 min after [18F]EF5 administration (Komar et al., 2008), and in vivo metabolism is minimal in humans, we can assume that the tissue uptake of [18F]EF5 is mainly reversible, and can well be analysed using tissue-to-blood or tissue-to-reference tissue ratios.
Static late scans with [18F]EF5 have been analysed using tissue-to-muscle ratios (Komar et al., 2008 and 2014; Silvoniemi et al., 2018; Narva et al., 2021). Qian et al., 2018" performed static PET scans 130 min after [18F]EF5 administration, and calculated regional SUVs and tumour-to-blood ratios.
Test-retest imaging ∼7 days apart has shown good repeatability of SUVmean, SUVmax, and tumour-to-muscle ratio at 3 h (Silvoniemi et al., 2018). Repeatability of [18F]EF5 PET in dose painting for radiotherapy was also good (Wright et al., 2021).
Dolbier WR Jr, Li AR, Koch CJ, Shiue CY, Kachur AV. [18F]-EF5, a marker for PET detection of hypoxia: synthesis of precursor and a new fluorination procedure. Appl Radiat Isot. 2001; 54(1): 73-80. doi: 10.1016/s0969-8043(00)00102-0.
Koch CJ, Evans SM. Non-invasive PET and SPECT imaging of tissue hypoxia using isotopically labeled 2-nitroimidazoles. Adv Exp Med Biol. 2003; 510: 285-292. doi: 10.1007/978-1-4615-0205-0_47.
Koch CJ, Evans SM. Optimizing hypoxia detection and treatment strategies. Semin Nucl Med. 2015; 45(2): 163-176. doi: 10.1053/j.semnuclmed.2014.10.004.
Updated at: 2023-01-29
Created at: 2023-01-28
Written by: Vesa Oikonen