Emerging Tracers for Nuclear Cardiac PET Imaging

Abstract

Myocardial perfusion imaging using positron emission tomography (PET) has several advantages over single photon emission computed tomography (SPECT). The recent advances in SPECT technology have shown promise, but there is still a large need for PET in the clinical management of coronary artery disease (CAD). Especially, absolute quantification of myocardial blood flow (MBF) using PET is extremely important. In spite of considerable advances in the diagnosis of CAD, novel PET radiopharmaceuticals remain necessary for the diagnosis of CAD because clinical use of current cardiac radiotracers is limited by their physical characteristics, such as decay mode, emission energy, and half-life. Thus, the use of a radioisotope that has proper characteristics and a proper half-life to develop myocardial perfusion agents could overcome these limitations. In this review, the current state of cardiac PET and a general overview of novel ¹⁸F or ⁶⁸Ga-labeled radiotracers, including their radiosynthesis, in vivo characterization, and evaluation, are provided. The future perspectives are discussed in terms of their potential usefulness based on new image analysis methods and hybrid imaging.

Introduction

Single photon emission computed tomography (SPECT) has played an important role in assessing myocardial blood flow (MBF) and in diagnosing CAD since the 1980s [1, 2]. However, the technical limitations of SPECT imaging using ^99mTc-labeled radiotracers, such as the difficulty of photon attenuation correction and uptake of radiotracers in the liver, may limit the diagnostic accuracy of SPECT [3]. PET has a number of technical advantages over SPECT, such as higher spatial resolution, a standardized method for the correction of photon attenuation, and quantitative measurement of MBF [4]. However, the PET radioisotopes which are currently used in diagnosis of CAD have a short physical half-life ([¹³N]NH₃, 9.97 min; ⁸²Rb, 1.27 min; [¹⁵O]H₂O, 2.04 min). This means that these radioisotopes need an on-site cyclotron or generator for their widespread use in the clinic [5]. Thus, the use of radioisotopes that have proper physical characteristics and a proper half-life to label myocardial imaging agents (MIA) could overcome these limitations. Previously, radiosynthesis and in vivo characterization of ¹⁸F-labeled phosphonium cations were reported and discussed [6]. In this review, the current state of cardiac PET and a general overview of novel ¹⁸F (including phosphonium cations) or ⁶⁸Ga-labeled radiotracers are provided. ¹⁸F (half-life, 109.8 min; β⁺, 97%) or ⁶⁸Ga (half-life, 67.7 min; β⁺, 88%) are popular radioisotopes used to develop small molecule radiopharmaceuticals for PET due to favorable decay characteristics. Furthermore, clinical perspectives of cardiac PET are discussed.

The Current State of Cardiac PET

There is a large need for PET in the management of CAD due to its advantages over SPECT, such as superior image quality, less radiation dose, and absolute quantification of MBF. In a recent report, the excellent image quality with PET was actually associated with superior diagnostic accuracy in the detection of functionally significant CAD, as compared to that of not only SPECT but also coronary CT [7]. Both sensitivity and specificity of PET were well above 80%, in contrast to SPECT with suboptimal sensitivity and coronary CT with suboptimal specificity. Absolute quantification of MBF by myocardial perfusion PET has also shown promise in the clinical management of CAD. PET-measured absolute MBF parameters have shown superior diagnostic values compared to relative analysis of a perfusion defect [8, 9]. PET-measured MBF parameters also show prognostic values. Even in the subgroup of patients with abnormal perfusion (summed stress score, ≥ 4), coronary flow reserve (CFR, < 2.0) measured by [¹³N]NH₃ PET was still predictive of future cardiac events [10]. PET-measured CFR was predictive of cardiac events, being independent from angiographic severity of epicardial CAD. Bypass surgery, in comparison with percutaneous coronary intervention, was related to a better prognosis in patients with CFR < 2.0 [11]. PET-measured CFR is considered to reflect not only focal significant stenosis but also underlying diffuse atherosclerosis, and it may help to decide on the optimal treatment strategy in CAD patients [11, 12].

However, current myocardial perfusion PET radiotracers generally require an on-site cyclotron (except for ⁸²Rb) mainly due to their short half-lives. This results in limited distribution and clinical utilization of myocardial perfusion PET. Recent studies suggest that cadmium-zinc-telluride (CZT) SPECT may be a substitute for myocardial perfusion PET. It showed improved image quality as compared to conventional SPECT, improved diagnostic accuracy, less radiation exposure, and even quantification of MBF [13–17]. Although the measurement of MBF using CZT SPECT is in progress, the generally lower extraction fraction of ^99mTc-labeled agents may increase image noise and is suboptimal for accurate calculation of MBF [18]. It was recently shown that the resting MBF measured by CZT SPECT was comparable to that measured by [¹³N]NH₃ PET, but stress MBF and CFR were significantly underestimated [19]. The obtuse correlation curve between MBF values measured by CZT SPECT and [¹³N]NH₃ PET reflects the suboptimal non-linear extraction fraction of myocardial SPECT agents, which is highlighted during vasodilator stress. With respect to the fact that the mainstay of functional assessment of CAD is based on stress perfusion, the concerns related to the low extraction fraction of SPECT agents (^99mTc-labeled agents) raise questions about its clinical reliability.

Thus, an ¹⁸F-labeled myocardial perfusion agent was developed, and it was expected to provide comparable or even superior image quality compared to current PET agents, as well as wide availability due to its longer half-life [20–23]. This ¹⁸F-labeled myocardial perfusion PET agent, ¹⁸F-BMS-747158-02 ([¹⁸F]flurpiridaz), is being clinically tested. It recently entered a clinical phase III open-label trial, in which its diagnostic accuracy for stenosis ≥ 50% on invasive coronary angiography was compared with that of ^99mTc-methoxyisobutylisonitrile (MIBI). Contrary to our expectation, the specificity of [¹⁸F]flurpiridaz PET—one of the two co-primary endpoints of the phase III trial—failed to reach the threshold of non-inferiority to that of ^99mTc-MIBI SPECT. Although further analyses in subgroups of female [24] and obese [25] patients showed superior diagnostic accuracy and met the non-inferiority threshold for specificity, the application of ¹⁸F-flurpiridaz has not yet become a clinical reality.

Novel ¹⁸F-Labeled Myocardial Perfusion Agents

Mitochondria play an important role in cell life and cell death. Especially, mitochondria constitute 20 to 30% of the myocardial intracellular volume as powerhouse [23]. Thus, unique protein or characteristics of mitochondria are a major target for developing of PET myocardial imaging agents.

First target of mitochondria for developing of PET myocardial perfusion agents is mitochondrial complex I (MC-1). MC-1 is the first enzyme of four electron transport complexes in the inner mitochondrial membrane. Radioisotope-labeled myocardial imaging agents which target the MC-1 are reported from the early 2000s, such as classical MC-1 inhibitor rotenone derivatives ([¹²⁵I]iodorotenone or [¹⁸F]fluorodihydrorotenone) and pyridazinone derivative ([¹⁸F]flurpiridaz) [23, 26, 27]. Rotenone derivatives showed higher uptake than ^99mTc-MIBI in an isolated rabbit heart. However, the mechanism of rotenone derivatives in myocytes had not been fully cleared and a large amount of rotenone derivatives was retained nonspecifically in the mitochondrial membrane. [¹⁸F]flurpiridaz represented high quality of myocardium image and had ability to visualize and quantify cardiac defects in different animal models including humans. However, the phase III trial failed due to already mentioned diagnostic accuracy.

Second target is the mitochondrial membrane potential (MP) in cardiomyocytes. The MP of cardiomyocytes is higher than that of normal epithelial cells, and loss of MP is a signal for cell death caused by myocardial ischemia [28, 29]. Quaternary ammonium and phosphonium cations accumulate in mitochondria of cardiomyocytes because the lipophilic cations (positive charge) move into the mitochondrium due to its high-density and electrochemical transmembrane potential (negative charge) in cardiomyocytes [30]. The myocardial uptake mechanism of these cations is similar to that of ^99mTc-MIBI and ^99mTc-tetrofosmin for SPECT [31]. Therefore, these radiolabeled cation derivatives are potential candidates for developing MIA.

Quaternary ammonium cations as myocardial perfusion PET agent were reported from 2011 [32–34]. These cations were labeled with ¹¹C due to drawback of longer half-life of ¹⁸F. A waiting period of four to five half-lives is ideally needed between at rest and stress PET scan to avoid significant interference. However, ¹⁸F-based MIA prevent both scans during a single day or needed several hours like SPECT scan protocol using ^99mTc-labeled MIA. Six kinds of ¹¹C-labeled ammonium cations were developed and [¹¹C]dimethyl-diphenyl-ammonium trifluoromethane sulfonate ([¹¹C]DMDPA) showed the best performance among them in rat or pig models.

The reported ¹⁸F-labeled tetraphenylphosphonium salt (TPP) derivatives are summarized in Fig. 1. ¹⁸F-labeled TPP derivatives are classified into three types (type 1, short aliphatic or ether form linkers are adopted between the triphenylphosphonium and ¹⁸F; type 2, ¹⁸F-substituted phenyl or benzyl group is attached to triphenylphosphine; and type 3, ¹⁸F-labeled linkers are conjugated to triphenylphosphine derivatives by click chemistry) [30, 35–49]. Among these cations, 4-¹⁸F-fluorobenzyltriphenylphosphonium (¹⁸F-FBnTP) showed promise for the first time as a myocardial perfusion PET agent. ¹⁸F-FBnTP is metabolically stable and performs well in healthy and CAD models [30, 46, 48]. However, in vivo characteristics of ¹⁸F-FBnTP, such as higher uptake and longer retention in the heart and more rapid wash out from the liver, need to be improved. ¹⁸F-Labeled alkyl chain-conjugated triphenylphosphonium cations, [¹⁸F]fluoroalkyltriphenylphosphonium cations ([¹⁸F]FATPs), show improved characteristics through lipophilicity control [38, 41–44]. Lipophilicity of the radiotracer could control myocardial selectivity and hepatic clearance. Although little is known about the lipophilicity needed for high myocardial selectivity, the log P value range for cationic radiotracers is 0.5–1.3 to image organs with high mitochondrial density, as a result of their fast membrane-penetration kinetics [50]. The lipophilic interaction between the [¹⁸F]FATPs and the lipid layer is attractive because the ¹⁸F-labeled alkyl group increases the lipophilicity [51, 52]. Furthermore, ¹⁸F-labeled TPP derivatives that have different functional groups, such as methyl, methoxy, tert-butyl, dimethyamine-substituted phenyl, and triazole formation, were investigated to assess their performance as MIA for PET [35, 40, 42, 53].

Fig. 1.

Schematic structure of radioisotope-labeled tetraphenylphosphonium cation derivatives. a (4-[¹⁸F]fluorophenyl)triphenylphosphonium cation ([¹⁸F]TPP), b (3-[¹⁸F]fluoropropyl)triphenylphosphonium cation, c (5-[¹⁸F]fluoropentyl)triphenylphosphonium cation ([¹⁸F]FPTP), d (6-[¹⁸F]fluorohexyl)triphenylphosphonium cation ([¹⁸F]FHxTP), e (7-[¹⁸F]fluoroheptyl)triphenylphosphonium cation ([¹⁸F]FHtTP), f (8-[¹⁸F]fluorooctyl)triphenylphosphonium cation ([¹⁸F]FOTP), g (2-(2-[¹⁸F]fluoroethoxy)ethyl)triphenylphosphonium cation ([¹⁸F]FETP), h (2-(2-[¹⁸F]fluoroethoxy)ethyl)tris(4-methoxyphenyl)phosphonium cation ([¹⁸F]FETMP), i (4-[¹⁸F]fluorobenzyl)triphenylphosphonium cation ([¹⁸F]FBnTP), j (4-(2-(2-(2-[¹⁸F]fluoroethoxy)ethoxy)ethoxy)benzyl)triphenylphosphonium cation ([¹⁸F]FPEGBnTP), k tris(4-(dimethylamino)phenyl)(4-[¹⁸F]fluorobenzyl)phosphonium cation, l (4-([¹⁸F]fluoromethyl)benzyl)triphenylphosphonium cation ([¹⁸F]FMBTP), m (3-([¹⁸F]fluoromethyl)benzyl)triphenylphosphonium cation ([¹⁸F]mFMBTP), n (4-([¹⁸F]fluoromethyl)benzyl)tris(2-methoxyphenyl)phosphonium cation, o (4-([¹⁸F]fluoromethyl)benzyl)tris(3-methoxyphenyl)phosphonium cation, p (4-([¹⁸F]fluoromethyl)benzyl)tris(4-methoxyphenyl)phosphonium cation, q tris(4-tert-butylphenyl)(2-(1-(2-[¹⁸F]fluoroethyl)-1H-1,2,3-triazol-4-yl)ethyl)phosphonium cation (¹⁸F]MitoPhos_04) r (2-(1-(2-[¹⁸F]fluoroethyl)-1H-1,2,3-triazol-4-yl)ethyl)trip-tolylphosphonium cation (¹⁸F]MitoPhos_05), s tris(3,5-dimethylphenyl)(2-(1-(2-[¹⁸F]fluoroethyl)-1H-1,2,3-triazol-4-yl)ethyl)phosphonium cation ([¹⁸F]MitoPhos_07)

¹⁸F-Fluorination of Tetraphenylphosphonium Cation Derivatives

Synthesis methods of ¹⁸F-labeled TPP or TPP derivatives, such as 3-[¹⁸F]fluoropropyltriphenylphosphonium cation, 4-[¹⁸F]fluorobenzyltriphenylphosphonium cation ([¹⁸F]FBnTP), and 4-[¹⁸F]fluorobenzyltris-4-dimethylaminophenylphosphonium cation, have been summarized and reported previously [6].

¹⁸F-labeled fluoroalkylphosphonium cations ([¹⁸F]FATPs), including (5-[¹⁸F]fluoropentyl)triphenylphosphonium cation ([¹⁸F]FPTP), (6-[¹⁸F]fluorohexyl)triphenylphosphonium cation ([¹⁸F]FHxTP), (7-[¹⁸F]fluoroheptyl)triphenylphosphonium cation ([¹⁸F]FHtTP), (8-[¹⁸F]fluorooctyl)triphenylphosphonium cation ([¹⁸F]FOTP), (2-(2-[¹⁸F]fluoroethoxy)ethyl)triphenylphosphonium cation ([¹⁸F]FETP), and (2-(2-[¹⁸F]fluoroethoxy)ethyl)tris(4-methoxyphenyl)phosphonium cation ([¹⁸F]FETMP), were investigated to assess the appropriate range of lipophilicity for myocardial imaging using PET [37–39, 41–44]. Synthesis of [¹⁸F]FATPs has been reported previously. In brief, α,ω-di-tosyloxyalkane was used as a precursor for the synthesis of [¹⁸F]FATPs. After ¹⁸F-fluorination, triphenylphosphine or tris(4-methoxyphenyl)phosphine was added to the reaction vessel and heated with no separation step. Finally, ¹⁸F-labeled mixture was purified by semipreparative HPLC and the collected HPLC fraction was co-injected with its reference compound into an analytical HPLC system for identification. [¹⁸F]FATPs were synthesized within 60 min with 10–20% RCY. Radiochemical purity was > 98% and average specific radioactivity was > 5.9 Tbq/μmol.

Other ¹⁸F-labeled TPP derivatives, such as 4-([¹⁸F]fluoromethyl)benzyltriphenylphosphonium ([¹⁸F]FMBTP), 3-([¹⁸F]fluoromethyl)benzyltriphenylphosphonium ([¹⁸F]mFMBTP), mitophos series (MitoPhos_04, MitoPhos_05, MitoPhos_07), and ¹⁸F-labeled triethylene glycol derivative of BnTP ([¹⁸F]FPEGBnTP), have been reported since 2014 [36, 40, 54]. [¹⁸F]FMBTP and [¹⁸F]mFMBTP have the same formula; however, the position of fluoromethyl structure is different. As a precursor, 1,4-bis(bromomethyl)benzene (for [¹⁸F]FMBTP) or 1,3-bis-(bromomethyl)benzene (for [¹⁸F]mFMBTP) was used and radiotracers were synthesized by using a simple one-pot procedure. Their RCY values were approximately 52 ± 9.3% (n = 7, for [¹⁸F]FMBTP) and 50.6 ± 6.9% (n = 7, for [¹⁸F]mFMBTP). Both radiotracers showed excellent radiolabeling yields (> 50%) and radiochemical purities (> 99%). The specific activities of [¹⁸F]FMBTP and [¹⁸F]mFMBTP were approximately 30 and 38 GBq/μmol, respectively. Furthermore, the performance of the methoxy group-substituted [¹⁸F]FMBTP was reported [35].

The first mitophos compound ([¹⁸F]MitoPhos_01) was reported in 2013 [55]. To achieve a suitable log P value to provide favorable biodistribution, the functional groups of the phenyl rings present around the phosphorus center were modified and synthesis of three novel phosphonium cations ([¹⁸F]MitoPhos_04, [¹⁸F]MitoPhos_05, and [¹⁸F]MitoPhos_07) was reported, and they can be synthesized by using the common radiolabeled linker, [¹⁸F]fluoroethylazide ([¹⁸F]FEA), for click chemistry in the same reaction vial [40]. The conversion percentage of the [¹⁸F]FEA to the labeled phosphonium cations was in the range of 80%. The specific activity of [¹⁸F]MitoPhos_04 and [¹⁸F]MitoPhos_07 was 49.6 ± 6.5 GBq/μmol (n = 3) and 68 ± 18.8 GBq/μmol (n = 3), respectively. The total synthesis time was less than 1 h. Furthermore, initial studies of three-tracer one-pot synthesis have shown promising preliminary results.

[¹⁸F]FPEGBnTP was designed to develop a new BnTP derivative showing lower liver uptake and higher heart-to-liver ratio in comparison with [¹⁸F]FBnTP [36]. It was hypothesized that the lipophilicity of [¹⁸F]FBnTP would affect its accumulation in the liver and a lower lipophilic derivative would indicate faster clearance from the liver. Finally, ¹⁸F-pegylated BnTP was synthesized and its in vitro/in vivo performance was evaluated. Radiochemical yield of [¹⁸F]FPEGBnTP was 13–23% (n = 8) and radiochemical purity of the drug solution was > 95%.

In Vivo Characteristics of ¹⁸F-Labeled Tetraphenylphosphonium Cation Derivatives

Biodistribution results and microPET images of ¹⁸F-labeled TPP derivatives are shown in Table 1 and Fig. 2. The lipophilicity, functional group of the triphenylphosphonium core, and structure of ¹⁸F-labeled linker are summarized in Table 2. Directed ¹⁸F-labeled TPP showed intense cardiac uptake in biodistribution (1.64% ID/g at 5 min after injection; ID, injected dose) and microPET studies (2.3 nCi/cm³ at 1 min after injection) [45]. A biodistribution study revealed that blood activity changed significantly from 5 to 60 min, reducing from 0.15 to 0.02% ID/g. At 5, 30, and 60 min, lung activity was 0.69, 0.36, and 0.38% ID/g, respectively, whereas liver uptake was 0.34, 0.18, and 0.17% ID/g, respectively. Heart-to-lung ratios at 5, 30, and 60 min were 2, 5, and 4, respectively. The cardiac uptake of ¹⁸F-labeled TPP in rats showed a similar value to that of ^99mTc-tetrofosmin or ^99mTc-MIBI. The time-activity curve (TAC) generated from PET imaging of rats showed that ¹⁸F-labeled TPP uptake by the heart wall was approximately fourfold higher than blood uptake for 60 min.

Table 1.

Comparison of biodistribution at 60 min after i.v. injection of ¹⁸F-labeled TPP derivatives into mice

	[18F]TPPa	[18F]FBnTP	[18F]FPTP	[18F]FHxTP	[18F]FETP	[18F]FETMP
Blood	0.02 ± 0.003	0.05 ± 0.01	0.04 ± 0.01	0.19 ± 0.03	0.23 ± 0.02	0.73 ± 0.63
Heart	1.57 ± 0.18	12.16 ± 2.4	21.23 ± 3.60	23.20 ± 2.70	19.96 ± 2.08	29.88 ± 7.59
Lung	0.38 ± 0.11	1.82 ± 0.3	2.96 ± 0.63	3.70 ± 0.12	4.01 ± 0.74	6.38 ± 1.21
Liver	0.17 ± 0.03	8.09 ± 3.9	3.05 ± 0.92	1.30 ± 0.12	2.03 ± 0.17	11.22 ± 1.48
Kidney	1.75 ± 0.38	24.70 ± 4.1	11.68 ± 3.73	12.68 ± 3.13	16.86 ± 2.73	41.10 ± 9.94
Muscle	0.26 ± 0.11	1.27 ± 0.4	6.16 ± 1.78	4.53 ± 0.54	4.42 ± 1.15	5.41 ± 1.93
Heart-to-liver	8	1.50	7.25	17.83	9.87	2.74
Heart-to-lung	4	6.68	7.28	6.27	5.05	4.75
Heart-to-blood	75	243	523.34	126.76	86.78	61.07
	[18F]FMBTP	[18F]mFMBTP	[18F]FMBTP (para)	[18F]FMBTP (ortho)	[18F]FMBTP (meta)	[18F]FPEGBnTP
Blood	0.41 ± 0.01	0.61 ± 0.05	0.97 ± 0.04	0.53 ± 0.08	0.31 ± 0.04	1.08 ± 0.05
Heart	31.17 ± 1.5	28.30 ± 2.36	21.68 ± 0.69	20.06 ± 3.27	15.65 ± 3.16	3.3 ± 0.22
Lung	4.1 ± 0.69	5.08 ± 0.32	6.58 ± 1.42	3.20 ± 0.46	4.08 ± 1.16	0.81 ± 0.01
Liver	1.89 ± 0.32	2.39 ± 0.56	2.30 ± 0.19	1.30 ± 0.41	1.06 ± 0.14	0.83 ± 0.02
Kidney	17.1 ± 3.72	19.73 ± 1.69	62.62 ± 5.86	43.47 ± 3.24	29.43 ± 4.71	3.64 ± 0.18
Muscle	6.91 ± 1.04	4.29 ± 0.67	5.37 ± 0.82	4.06 ± 0.81	3.73 ± 0.63	1.28 ± 0.06
Heart-to-liver	16.43	11.84	9.43	15.43	14.76	4.00
Heart-to-lung	7.46	5.57	3.29	6.27	3.84	4.10
Heart-to-blood	74.74	46.39	22.35	37.85	50.48	3.13

^aBiodistribution in Sprague Dawley rats. Data are expressed as the percentage of injected dose per gram of tissue (% ID/g)

Fig. 2.

Coronal microPET images of normal rats (mouse) after intravenous injection of a [¹⁸F]FBnTP (mouse), b [¹⁸F]FPTP, c [¹⁸F]FETP, and d [¹⁸F]mFMBTP. H, heart; L (LIV), liver; RV, right ventricle; LV, left ventricle; SUV, standardized uptake value (Reprinted with permission [41, 44, 46, 54])

Table 2.

The lipophilicity and the functional group of the triphenylphosphonium core and structure of ¹⁸F-labeled linker of ¹⁸F-labeled TPP derivatives

	Lipophilicity	Functional group of triphenylphosphonium core	Structure of 18F-labeled linker
[18F]FBnTP	− 0.38	N/A	Fluorobenzyl
[18F]FPTP	1.31 ± 0.02	N/A	Fluoropentyl
[18F]FHxTP	1.78 ± 0.05	N/A	Fluorohexyl
[18F]FHtTP	2.52 ± 0.01	N/A	Fluoroheptyl
[18F]FOTP	2.91 ± 0.02	N/A	Fluorooctyl
[18F]FETP	0.89 ± 0.02	N/A	(Fluoroethoxy)ethyl
[18F]FETMP	1.27 ± 0.01	Methoxy (para)	(Fluoroethoxy)ethyl
[18F]FMBTP	1.16 ± 0.003	N/A	4-(Fluoromethyl)benzyl
[18F]mFMBTP	1.05 ± 0.01	N/A	3-(Fluoromethyl)benzyl
[18F]FMBTP (para)	1.67 ± 0.01	Methoxy (para)	4-(Fluoromethyl)benzyl
[18F]FMBTP (ortho)	1.53 ± 0.02	Methoxy (ortho)	4-(Fluoromethyl)benzyl
[18F]FMBTP (meta)	0.54 ± 0.05	Methoxy (meta)	4-(Fluoromethyl)benzyl
[18F]MitoPhos_04	2.03 ± 0.19	tert-Butyl	Triazole
[18F]MitoPhos_05	1.31 ± 0.06	Methyl (para)	Triazole
[18F]MitoPhos_07	1.83 ± 0.04	3,5-Dimethyl	Triazole
[18F]FPEGBnTP	− 0.92	N/A	Triethylene glycol

N/A not applicable

Among 3-[¹⁸F]fluoropropyltriphenylphosphonium, [¹⁸F]FBnTP, and 4-[¹⁸F]fluorobenzyltris-4-dimethylaminophenylphosphonium, only the in vivo characteristics of [¹⁸F]FBnTP were reported [46]. [¹⁸F]FBnTP showed strongest uptake in the kidney, followed by the heart (12.16% ID/g) and liver (8.09% ID/g), and very low uptake in the blood (0.05% ID/g) at 60 min. Furthermore, [¹⁸F]FBnTP demonstrated excellent characteristics as a cardiac imaging agent in murine and normal canine models. The myocardium-to-liver uptake ratio reached 1 at approximately 25 min after intravenous injection of [¹⁸F]FBnTP in mice and the myocardium-to-liver uptake ratio was 1 in dog models for 90 min.

[¹⁸F]FATPs showed excellent heart-specific uptake and wash out from the liver than the other ¹⁸F-labeled TPP derivatives. The lipophilicity of [¹⁸F]FATPs was controlled by the structure of ¹⁸F-labeled linker and the functional group of the triphenylphosphonium core. The myocardial uptake of [¹⁸F]FATPs was more than 20% ID/g for 2 h after radiotracer injection. The heart-to-liver ratio and the heart-to-blood ratio were more than 4 and 30, respectively, at 10 min. However, clearance of [¹⁸F]FETMP from the liver occurred later than that of other [¹⁸F]FATPs due to the functional group of the triphenylphosphonium core (heart-to-liver ratio at 10 min, 2.33 ± 0.36). [¹⁸F]FATPs showed superior heart-background contrast on microPET imaging of murine models at each time point and better image quality than [¹³N]NH₃, which is the clinical gold standard myocardial imaging agent for PET [37, 41–44]. The [¹⁸F]FATPs images demonstrated good visualization of the heart, with excellent heart-to-liver and heart-to-lung ratios. The [¹⁸F]FATPs TACs demonstrated rapid accumulation in the heart (1–2 min), with stable retention for 60 min. The heart-to-liver ratios of [¹⁸F]FPTP, [¹⁸F]FHxTP, and [¹⁸F]FETP were 1.67 ± 0.21, 2.49 ± 0.08, and 2.6 ± 0.85, respectively, whereas the heart-to-lung ratios of [¹⁸F]FPTP, [¹⁸F]FHxTP, and [¹⁸F]FETP were 3.65 ± 0.2, 5.67 ± 0.52, and 6.31 ± 0.8, respectively, 1 min after intravenous injection. In contrast, [¹³N]NH₃ TAC revealed that the heart-to-liver ratio and the heart-to-lung ratio were 0.8 ± 0.14 and 0.93 ± 0.07, respectively, at 5 min after injection. Furthermore, [¹⁸F]FATPs were compared with [¹³N]NH₃ by using myocardial infarction (MI) models. Sharply defined myocardial defects were present with [¹⁸F]FATPs or [¹³N]NH₃ (10 min). [¹⁸F]FATPs showed similar values for defect size and contrast ratio; however, the [¹³N]NH₃ images significantly underestimated infarct size soon after tracer injection (0–10 min, P = 0.027), which might be due to the spillover of radioactivity into the left ventricular or right ventricular myocardium [37]. Furthermore, [¹⁸F]FATPs showed significantly higher first-pass extraction fraction value than those of [¹³N]NH₃ in isolated rat hearts perfused by the Langendorff method (flow rates 4, 8, or 16 mL/min, P < 0.05) [6]. The flow rate would be relevant because the MBF of normal rats under isoflurane anesthesia is 4.2 ± 0.9 mL/min [56].

[¹⁸F]FMBTP and [¹⁸F]mFMBTP were reported in 2014 [54]. Both compounds were electropositive, and their log P values at pH 7.4 were 1.16 ± 0.003 (n = 3) and 1.05 ± 0.01 (n = 3), which were comparable to those of [¹⁸F]FETMP (1.27 ± 0.01) and [¹⁸F]FHxTP (1.78 ± 0.05). [¹⁸F]FMBTP showed high heart uptake (25.24 ± 2.97% ID/g at 5 min after intravenous injection) and the activity was retained for 2 h (28.99 ± 3.54% ID/g at 2 h). However, the liver uptake was decreased dramatically from 14.93 ± 1.46% ID/g (at 5 min) to 0.89 ± 0.29% ID/g (at 120 min). The heart-to-liver, heart-to-lung, and heart-to-blood ratios of [¹⁸F]FMBTP were 5.04, 5.80, and 32.9, respectively, at 30 min postinjection, and they increased to 32.5, 7.61, and 151.9, respectively, at 120 min postinjection. Furthermore, [¹⁸F]mFMBTP showed higher heart uptake (31.02 ± 0.33% ID/g at 5 min postinjection) and better target-to-non target ratios at earlier time points. Both these new radiotracers showed moderate muscle uptake at all time points after injection (6.91 ± 1.04% ID/g for [¹⁸F]FMBTP and 4.29 ± 0.67% ID/g for [¹⁸F]mFMBTP at 60 min), which was similar to that of the previously reported lipophilic cations of [¹⁸F]FPTP (6.16 ± 1.78% ID/g at 60 min p.i.)10 and [¹⁸F]FHxTP (4.53 ± 0.54% ID/g at 60 min p.i.), respectively. Finally, in vivo PET studies were performed using [¹⁸F]mFMBTP in rats and Beagle dogs. [¹⁸F]mFMBTP showed an excellent image quality in rats for 2 h after intravenous injection. The highest myocardial uptake was reached rapidly at a very early time point after injection and it remained high across all time points to 2 h. With respect to liver uptake, it peaked at 5 min postinjection and rapid clearance was observed. It was even lower than lung uptake after 30 min postinjection. The heart-to-liver ratio increased obviously due to quick clearance from the liver (1.37, 6.62, 12.06, and 18.64 at 5, 30, 60, and 120 min postinjection, respectively). In the case of Beagle dogs, the heart could be clearly seen at 10 min after injection, and heart uptake did not obviously decrease till 2 h. The values for heart-to-liver and heart-to-lung uptake ratios were calculated as 1.54 and 6.00 at 10 min after injection, 2.83 and 15.19 at 30 min after injection, 3.98 and 38 at 60 min after injection, and 7.76 and 35.28 at 120 min after injection, respectively. The results of PET imaging in dogs showed a trend similar to biodistribution results in rats and mice. In addition to this, the influence of an additional methoxy group and its different positions on myocardium uptake and pharmacokinetic properties of [¹⁸F]FMBTP were examined and reported in 2017 [35]. However, [¹⁸F]FMBTP showed the highest heart uptake value.

Radiosynthesis and evaluation of ¹⁸F-labeled triethylene glycol derivative of BnTP, [¹⁸F]FPEGBnTP, were reported in 2016 [36]. However, in vitro cellular uptake and in vivo biodistribution of [¹⁸F]FPEGBnTP were preliminarily evaluated to elucidate the possibility of its use as an MPI agent. Lipophilicity of [¹⁸F]FPEGBnTP was compared with that of [¹⁸F]FBnTP. Log P values of [¹⁸F]FBnTP and [¹⁸F]FPEGBnTP were − 0.38 and − 0.92 respectively, indicating that [¹⁸F]FPEGBnTP is relatively less lipophilic than [¹⁸F]FBnTP. [¹⁸F]FPEGBnTP showed heart uptake of 3.33 ± 0.22% ID/g at 60 min after intravenous injection. In contrast, liver and lung exhibited lower uptakes (0.83 ± 0.02 and 0.81 ± 0.01% ID/g at 60 min, respectively). However, radioactivity uptake in the blood (1.08 ± 0.05% ID/g) was relatively higher than that in the liver and lung

⁶⁸Ga-Labeled Myocardial Perfusion Agents

¹⁸F-labeled MPI agents based on the lipophilic cation have been mentioned above; however, these agents depend on the presence of nearby cyclotrons. Recently, ⁶⁸Ga (t_1/2 = 68 min) eluted from germanium/gallium (Ge/Ga) generators (parent–daughter combination system) showed high quality with excellent emission properties for clinical PET imaging [57]. The parent nuclide (⁶⁸Ge) onto a titanium column and the daughter nuclide (⁶⁸Ga) are simply eluted as ⁶⁸GaCl₃ with hydrochloric acid. This generator can be used for more than a year due to the long half-life of the parent nuclide ⁶⁸Ge (t_1/2 = days). Furthermore, the short half-life of ⁶⁸Ga can be eluted two or three times per day from the generator. Thus, the ⁶⁸Ge/⁶⁸Ga generator is appropriate for experimental or clinical application [58]. Many kinds of synthetic complexes of ⁶⁸Ga(III) have been reported as candidate PET tracers for cardiac nuclear imaging [58–71]. In this review, we have categorized them according to their characteristics, as perfusion imaging agents, MI imaging agents and cardiac metabolic imaging agents, and we have summarized them (Table 3).

Table 3.

⁶⁸Ga-labeled cardiac imaging agents

Category	Radiotracer	Target	Reference
Perfusion imaging	68Ga[(4,6-MeO2sal)2BAPEN]+	Mitochondria	[58, 59]
	68Ga[(3-MeOsal)2Me4BAPEN]+	Mitochondria	[63]
	68Ga[Ga(3-EtOsal)2Me4BAPEN]+
	68Ga[(sal)2BAPDMEN]+	Mitochondria	[63, 72]
	68Ga[(4-MeOsal)2BAPDMEN]+	Mitochondria	[72]
	68Ga[(4,6-MeOsal)2BAPDMEN]+
	68Ga[(3-MeOsal)2BAPDMEN]+	Mitochondria	[63, 72]
	68Ga[(6-MeOsal)2BAPDMEN]+	Mitochondria	[72]
	68Ga[3-isopropoxy-ENBDMPI]+	Mitochondria	[64]
Myocardial infarction imaging	68Ga-NOTA-NGR	Aminopeptidase N (CD13)	[65]
	68Ga-pentixafor	Chemokine receptor 4 (CXCR4)	[67]
	68Ga-NOTA-RGD	αvβ3 integrin	[66]
	68Ga-NODAGA-RGD		[68]
	68Ga-DOTA-peptide (phage display)	MMP-2, MMP-9	[69]
Cardiac metabolic imaging	68Ga-NOTA-undecanoic acid	Fatty acid oxidation	[70]
	68Ga-NOTA-dodecanoic acid
	68Ga-NODAGA-undecanoic acid	Fatty acid oxidation	[71]
	68Ga-DTPA-undecanoic acid

⁶⁸Ga-labeled lipophilic cations have been reported as myocardial perfusion agents [59, 60, 62, 73, 74]. Among them, ⁶⁸Ga[(4,6-MeO₂sal)₂BAPEN]⁺ showed heart-specific accumulation. The myocardial uptake mechanism of ⁶⁸Ga-BAPEN is presumably similar to that of ^99mTc-MIBI or [¹⁸F]FATPs because radiotracers are lipophilic and have monovalent ions with + 1 charge. However, their labeling procedures are complicated and they need time-consuming operations. Thus, a practical kit was developed to overcome these limitations and to increase usage convenience [58]. In a biodistribution study with balb/c mice, ⁶⁸Ga-BAPEN showed strongest uptake in the kidney (52.45 ± 6.25% ID/g), followed by the intestine (18.90 ± 4.05% ID/g), liver (17.73 ± 2.53% ID/g), and heart (11.77 ± 0.83% ID/g) and low uptake in the muscle (1.88 ± 0.43% ID/g) and blood (0.39 ± 0.04% ID/g) at 1 h postinjection. Furthermore, ⁶⁸Ga-labeled hexadentate bis(salicylaldimine) chelates (BAPDMEN) derived from BAPEN were reported and they showed promising preclinical results [72]. In rats, the myocardial uptake of the ⁶⁸Ga-labeled tracers exceeded that of ^99mTc-MIBI after 1 min postinjection and it showed prolonged retention in the myocardium for 2 h. Tarkia et al. investigated and reported the myocardial uptake and plasma protein binding properties of BAPEN and BAPDMEN derivatives using a pig model [63]. A novel lipophilic monocationic Ga(III)-complex, ^67/68Ga-[3-isopropoxy-ENBDMPI]⁺, was reported in 2014 [64]. ⁶⁷Ga-[3-isopropoxy-ENBDMPI]⁺ showed specific heart uptake (11.98 ± 0.74% ID/g) and low liver uptake (7.44 ± 0.44% ID/g) at 60 min postinjection. They explained that low liver uptake was caused by expression of P-glycoprotein (Pgp) in hepatocytes. The uptake mechanism of lipophilic cation is opposed by the action of ATP-binding-cassette (ABC) membrane transporters, such as the multidrug resistance (MDR) Pgp (ABCB1) and MRP1 (ABCC1), which transport lipophilic cationic metal complexes out of the cells. Thus, ⁶⁷Ga-[3-isopropoxy-ENBDMPI]⁺ is retained for a prolonged period in cardiomyocytes which are rich in mitochondria and lack Pgp, while hepatocytes which express Pgp along their canalicular surface rapidly excrete ⁶⁷Ga-[3-isopropoxy-ENBDMPI]⁺ into the bile and intestines. They determined whether Pgp affected the accumulation of ⁶⁷Ga-[3-isopropoxy-ENBDMPI]⁺ in in vitro and in vivo conditions. Finally, they performed a microPET study with Sprague Dawley rats and images showed sustained retention of ⁶⁸Ga-[3-isopropoxy-ENBDMPI]⁺ in the heart and rapid clearance from the liver (Fig. 3).

Fig. 3.

Representative microPET/CT images of ⁶⁸Ga-[3-isopropoxy-ENBDMPI]⁺. Sprague Dawley rats were injected intravenously with ⁶⁸Ga-[3-isopropoxy-ENBDMPI]⁺ and images were acquired 60 min postinjection (Reprinted with permission [64])

⁶⁸Ga-Labeled Agents for MI or Cardiac Metabolic Imaging

MI is a dynamic process that begins with coronary occlusion and causes heart muscle damage by decreasing or stopping blood flow to the heart. Because of this, MI is visualized to target biomarkers related to cell damage or healing process, such as inflammation, angiogenesis, and matrix metalloproteinases (MMPs) [65–69].

Peptide sequences, asparagine-glycine-arginine (NGR) motif binds to aminopeptidase N (CD13), which is expressed on inflammatory cells [65]. In the myocardium, CD13 is expressed in pathophysiological conditions such as acute MI, where it facilitates inflammatory cell infiltration. Tillmanns J et al. synthesized ⁶⁸Ga-NOTA-NGR and compared it with [¹⁸F]FDG or ⁶⁸Ga-NOTA-arginine-glycine-aspartic acid (RGD) motif binds to integrin α_vβ₃ receptors in myocardial ischemia and reperfusion (MI/R) rat models. The [¹⁸F]FDG identified infarcted areas, where increased uptake of ⁶⁸Ga-NOTA-NGR was apparent. However, ⁶⁸Ga-NOTA-RGD showed higher liver uptake than ⁶⁸Ga-NOTA-NGR. Another agent related to the inflammatory process is ⁶⁸Ga-pentixafor [67]. Chemokine receptor 4 (CXCR4) is expressed on inflammatory cells after acute MI. CXCR4 belongs to the family of G-protein-coupled receptors and is involved in many kinds of biological processes, such as the entry of HIV-1, the development of metastasis, and several inflammatory conditions. Therefore, the CXCR4 receptor is an attractive biomarker to identify the activated inflammatory cells located in the MI area. ⁶⁸Ga-pentixafor is a ligand with high affinity and selectivity to hCXCR4-receptors (IC₅₀ = 5 nM) and it has been tried in human MI patients. As a result, ⁶⁸Ga-pentixafor PET images reveal markedly increased CXCR4 expression in the MI area, indicating an active inflammatory process.

The α_vβ₃ integrin is expressed on angiogenic vessels and is a potential biomarker for pathological processes. Its expression is upregulated in cancer lesions and metastases as well as in acute MI as part of the infarct healing process [66]. RGD is the principal integrin-binding domain present within extracellular matrix (ECM) proteins such as fibronectin, vitronectin, and fibrinogen. Thus, RGD or synthetic RGD derivatives provide several advantages for targeting integrin for molecular imaging applications [75]. ⁶⁸Ga-labeled RGD derivatives, such as ⁶⁸Ga-NOTA-RGD and ⁶⁸Ga-NODAGA-RGD were evaluated in MI rat and pig models [66, 68]. In PET studies, increased uptake is represented by both radiotracers in the MI area. It means that localized α_vβ₃ integrin expression in MI may be used as a biomarker to image MI.

MMPs are proteolytic enzymes that play a central role in the degradation of ECM proteins in MI [76, 77]. Especially, the expression of gelatinases MMP-2 and MMP-9 is elevated after myocardial injury. Thus, radioisotope-labeled MMP-2/9 target compound could be used as a PET agent for MI. ⁶⁸Ga-labeled specific MMP-2/9 binding peptide which was identified from the phage display library was evaluated in a MI rat model and it was reported in 2016 [69]. This radiotracer showed accumulation in the damaged rat myocardium after injury; however, its instability and inadequate in vivo characteristics make it unsuitable for further evaluation.

Fatty acids are the major energy source in the myocardium and a change in their metabolism reflects abnormality in the myocardium. Thus, radioisotope-labeled long-chain fatty acids have been investigated as imaging agents since the 1990s [78, 79]. Finally, several kinds of ⁶⁸Ga-labeled long-chain fatty acids were synthesized and evaluated [70, 71]. The first reported ⁶⁸Ga-labeled fatty acids were ⁶⁸Ga-NOTA-undecanoic acid and ⁶⁸Ga-NOTA-dodecanoic acid [70]. These radiotracers showed significant initial uptake in the myocardium; however, significant retention of the activity was found to be associated with blood. Thus, synthetic modification is needed to increase the potential of these radiotracers for myocardial imaging. Another work using ⁶⁸Ga-labeled fatty acids aimed to investigate the influence of different chelators of undecanoic acid on cardiac uptake and pharmacokinetics [71]. Jain K et al. synthesized ⁶⁸Ga-NODAGA-undecanoic acid and ⁶⁸Ga-DTPA-undecanoic acid and evaluated their in vivo characteristics. Biodistribution studies showed significant retention of ⁶⁸Ga-NODAGA-undecanoic acid in the myocardium. Moreover, rapid washout of ⁶⁸Ga-DTPA-undecanoic acid from the liver was observed. However, they suggested amalgamation of the structural features of the two radiotracers for the final design of the ⁶⁸Ga-labeled fatty acid complex that might help to achieve the desired pharmacokinetics for cardiac metabolic imaging.

Conclusion

The development of new radiotracers and/or novel utilization of existing radiotracers are essential steps in cardiac molecular imaging with PET. As described above, many radiotracers are awaiting implementation in the clinical workflow for various cardiac diseases. In parallel with these active research works, clinical efforts should be made to design PET more relevant for patient care.

For example, recent progress in PET in cardiac amyloidosis can provide a good opportunity to promote cardiac PET imaging. PET radiotracers including ¹⁸F sodium fluoride ([¹⁸F]NaF), [¹¹C]Pittsburgh compound B ([¹¹C]PIB), [¹⁸F]florbetapir, and [¹⁸F]florbetaben showed subtype-specific, promising results in cardiac amyloidosis. [¹⁸F]NaF showed distinguished uptake in transthyretin (TTR) cardiac amyloidosis [80], while the others were highly specific for immunoglobulin light chain (AL) cardiac amyloidosis [81–83]. However, indications for PET imaging for a specific subtype have not been established; the interpretation of PET images can be variable due to diverse analytical methods suggested in recent studies. It is mandatory to establish/create comprehensive guidelines on PET imaging in the context of clinical diagnostic work-up of cardiac amyloidosis. The American Society of Nuclear Cardiology published a practical document of practice points for ^99mTc pyrophosphate imaging for TTR cardiac amyloidosis [84]. Patient selection, imaging procedure, interpretation, and quantification methods were comprehensively included, making ^99mTc PYP readily accessible. The same can be done for PET imaging of AL cardiac amyloidosis. It would complete the workflow of molecular imaging, covering the two major subtypes of cardiac amyloidosis.

Cardiac sarcoidosis is also a good disease candidate to promote cardiac PET imaging. ⁶⁸Ga-DOTANOC, a neuroendocrine tumor imaging agent, was recently introduced as it showed superior diagnostic accuracy as compared to [¹⁸F]FDG [85]. Taking advantage of the accumulation in active inflammation, several ⁶⁸Ga-labeled somatostatin analogues are now being investigated in clinical trials. There are ongoing clinical efforts, although they are limited to [¹⁸F]FDG PET. The revised major guidelines include specific [¹⁸F]FDG PET findings as one of the diagnostic criteria, and they recommend PET for diagnosis and evaluation of the inflammatory activity in cardiac sarcoidosis [86, 87]. Volumetric parameters are now being applied to diagnosis and prognostic stratification [88, 89], in addition to the traditional maximal uptake value.

It is also encouraging that a relevant multimodality imaging approach is under investigation for these two disease entities, especially PET-magnetic resonance (MR) [80, 90, 91]. Excellent morphological evidence and tissue characterization are expected to have a synergistic effect with PET. Hybrid PET-MR can be one-stop-shop imaging capable of not only making the diagnosis but also evaluating TTR/AL subtyping or inflammation activity in amyloidosis and sarcoidosis, respectively.

Preclinical and clinical progress can trigger and promote/stimulate the development of each other. There are still challenges and competitors for nuclear cardiology imaging. But cardiac PET has a unique strength based on the infinite potential offered by the development of new radiotracers.

Funding Information

This research was supported by the Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education (2017R1D1A1B03029055 and 2016R1D1A3B01006631).

Compliance with Ethical Standards

Conflict of Interest

Dong-Yeon Kim, Sang-Geon Cho and Hee Seung Bom declare that there is no conflict of interest.