Evaluation of Hypoxia with Cu-ATSM

Abstract

Imaging of hypoxia is important in many diseases states in oncology, cardiology and neurology. The radiopharmaceutical, copper labelled diacetyl-bis(N-methylthiosemicarbazone) (Cu-ATSM), has been used to assess hypoxia in many studies. In particular, Cu-ATSM has been used in oncologic settings to investigate tumor hypoxia and the role of this parameter in response to therapy and outcome. Other groups have conducted imaging studies assessing the role of hypoxia in cardiovascular disease and neurological disorders. Additionally, several groups have made significant progress into understanding the mechanism by which this compound accumulates in cells. Multiple preclinical and clinical studies have been conducted, shedding light on the important of careful image analysis when using this tracer. This review article focusses on the recent preclinical and clinical studies with this tracer.

Introduction

It has been recognized for several decades that hypoxia plays an important role in development of tumor aggressiveness and is a major cause of resistance to chemotherapy and radiation therapy. Hypoxia also is important in some of the cardiovascular and neurological disorders. Oxygen electrodes, considered to be the gold standard for detection of hypoxia, have been the initial tool for measurement of oxygen tension. (1–3) Early studies using these electrodes demonstrated a strong correlation between pretherapy oxygen measurements and outcome of radiation therapy in patients with cervical cancer and head and neck cancer. (4, 5) However, this technique is invasive, technically difficult to perform, evaluates only readily accessible lesions and is sample dependent and, thus, cannot readily address tumor heterogeneity, which is an inherent feature of solid tumors. Therefore, alternative noninvasive methods such as positron emission tomography (PET) imaging for measurement of hypoxia have been the focus of many investigations. Imaging not only is noninvasive and can be repeated during the course of therapy, but also evaluates the entire tumor and, thus, addresses tumor heterogeneity. Several PET hypoxic radiotracers have been developed such as nitroimidazoles derivatives, [¹⁸F]fluoromisonidazole ([¹⁸F]MISO), [¹⁸F]flortanidazole ([¹⁸F]HX4) and ([¹⁸F]fluoroazomycin-arabinoside ([¹⁸F]FAZA). One of the current leaders of PET radiopharmaceuticals for measurement of hypoxia is copper labelled diacetyl-bis(N-methylthiosemicarbazone) (Cu-ATSM). (6–8) This review article describes the applications of Cu-ATSM in oncologic and non-oncologic settings with a focus on work that has been reported since 2009.

Preclinical Oncology Studies

Cu-ATSM has mainly been used for imaging hypoxia by several groups in oncologic studies. (6–9) This work has largely continued in both vitro and in vivo settings. Cu-ATSM is a neutral lipophilic molecule with high cell membrane permeability and diffuses readily from the bloodstream to surrounding cells. Once within the cell, it undergoes reduction only in hypoxic cells and becomes trapped within those cells, but washes out rapidly from normoxic cells without any change. ATSM has be labeled using several radioisotopes of copper; [⁶⁰Cu] (23.4 min half-life, β+ = 81%), [⁶¹Cu] (3.4 h half-life, β+ = 62%), [⁶²Cu] (9.7 min half-life, β+ = 97.5%) and [⁶⁴Cu] (12.7 h half-life, β+ = 17%, β− = 40%). [⁶⁴Cu] has physical properties that can be used for both diagnostic imaging and therapy. New applications are also appearing, in which Cu-ATSM (or analogues thereof) are being used in adjuvant or combination therapy settings. For example, Damelin et al., (10) showed in esophageal squamous cell carcinoma that the novel combination of metformin with non-tracer Cu-ATSM could result in improved cytotoxicity compared to metformin alone, Weeks et al. (11) showed that [⁶⁴Cu]ATSM could enhance DNA damage and cytotoxicity in hypoxic cells.

Oncology studies with Cu-ATSM – Controversies

There is no such thing as a universal tracer – the same can be said for Cu-ATSM as a PET imaging agent. The overwhelming body of literature on Cu-ATSM supports its selectivity of hypoxia in tumors. This selectivity has been looked at via mechanistic studies, in vitro, and in vivo, but conflicting reports do exist. (6–9, 12, 13)

Studies have been presented explaining the potential lack of hypoxia selectivity – for example the documented relationship between fatty acid synthase (FAS) expression in human prostate cancer cell lines (14) or that differences in the extent of trapping and retention of the Cu in cell types (12), may explain the lack of Cu-ATSM hypoxia selectivity in a subset of tumors. This phenomena was recently confirmed by the report by Valtorta et al., (15) where Cu-ATSM and F-FAZA distribution agreed in some tumors models, but did not in the PC-3 prostate cancer model. The authors accurately stated that all tracer validations need to be done in the correct and appropriate models prior to clinical application. It is unfortunate however, that some isolated publications still appear that present data in small studies — often in single mouse models — assert a generality that Cu-ATSM does not show selectivity. These studies are often presented without placing the data in context with previously published work and are reported in a comparison with nitromidazole based markers, which have a completely different trapping mechanism than Cu-ATSM. Further, these same works often question the validity of comparisons with a nitromidazole-based “gold standards” rather than direct measurements of tissue oxygenation. Carlin et al., compared four hypoxia tracers in a single mouse model (SQ20b) and stated that Cu-ATSM did not show the same selectivity as the F-18 tracers studied. (16) However, this report does not mention e.g., the FAS levels expressed in this tumor line. Notably, recent publications have asserted that Cu-ATSM provides complementary information compared with standard-of-care tracers, for example against FDG (17) and FDG/FLT (18) in canine solid tumors. Of particular interest is that these studies were in canine patients with spontaneous tumors rather than fast growing xenograft mouse models. In a third study in canine tumors, Hansen et al., (19, 20) compared FDG with Cu-ATSM, as well as with sectioned biopsies; Cu-ATSM scans at 3h postinjection showed a strong positive correlation to pimonidazole uptake in the most heterogeneous tumor regions.^17,18

Preclinical Cardiology Studies

Early studies by Fujibayashi et al. investigated [⁶⁴Cu]ATSM in an acute myocardial infarction model. (21) This work evaluated [⁶⁴Cu]ATSM distribution less than 30 minutes after occlusion to assure that the area studied exhibited ischemia but not infarction. The authors observed increased [⁶⁴Cu]ATSM uptake with decreased blood flow indicating that the agent was retained in tissue as oxygen was depleted. Importantly for image interpretation with this agent, these studies also illustrated the decrease in [⁶⁴Cu]ATSM uptake at very low flow rates (< 40% of normal). Preclinical studies with several radiocopper ATSM compounds in several canine models used a two compartment model to further characterize kinetic aspects of this tracer. (22) This study also included a metabolite analysis of the injected compounds and the authors reported flow independent [^*Cu]ATSM uptake. In all three models, the authors observed increased [^*Cu]ATSM uptake and retention in both hypoxic and ischemic tissue.

Building on this work, additional studies conducted by Takahashi et al demonstrated the use of [⁶²Cu]ATSM to image hypoxia in human subjects with myocardial ischemia. (23) Out of 7 patients imaged with coronary artery disease, the authors reported increased [⁶²Cu]ATSM in one patient with unstable angina after 20 minutes. Similar to previous work in preclinical models, the uptake of [⁶²Cu]ATSM correlated with areas with moderately reduced blood flow. Interestingly, abnormal [¹⁸F]FDG uptake was observed in 4 of the patients, 3 of which were clinically stable, indicating that [⁶²Cu]ATSM may be more sensitive for detecting at risk patients in this population. The authors concluded that this preliminary study suggests that these agents may be useful to assess the extent of hypoxia in acute ischemia but not chronic ischemia. (23)

A detailed study investigating the retention of [⁶⁴Cu]ATSM under normoxic and hypoxic conditions in an in vitro incubation system to allow for precise atmospheric control was recently reported by Handley et al (24). This study investigated the uptake of [⁶⁴Cu]ATSM under different oxygen conditions for various time intervals. The authors report that the monocytes tolerated all conditions during the intervals studies with no morphological changes. Monocytes cultured in 0% O₂ had statistically significant increased uptake of [⁶⁴Cu]ATSM over ones cultured in 95% O₂. (24) The authors report that after 30 min of exposure to hypoxia, there was nearly twice the level of tracer retention in hypoxic cells compared to controls. No difference was observed in cells incubated under hyperoxic conditions compared to controls. Under hypoxic conditions ⁶⁴Cu was found to localize mainly in the nucleus and cytosol of cells with 21% being retained in the mitochondrial fractions and no difference in the % of radioactivity in each compartment for the control conditions.

In other recent work by Handley et al., studies illustrated progress into the next generation of ATSM analogues. (25) Using an isolated rat heart model, the group was able to assess the characteristics of hypoxia imaging agents after perfusion with a hypoxic buffer to investigate uptake and wash out of selected tracers. In this study hearts were perfused with normoxic buffer for 30 minutes and then hypoxic buffer for 45 minutes. The radiotracers were administered into the arterial line after the normoxic phase and again during the hypoxic phase. Following this, kinetics were assessed during the hypoxic phase. [⁶⁴Cu]ATSM was shown to have high hypoxia sensitivity with a hypoxic to normoxic ratio of 8:1. Tissue retention of [⁶⁴Cu]Cl₂ was found to be negligible regardless of the levels of oxygenation. Additionally, they report development of several promising compounds which may improve on the sensitivity and specificity of the parent [Cu]ATSM compound. (25)

Preclinical Neurology Studies

Hypoxia in the brain has mainly been studied in the context of stroke. A recent study by Williamson et al compared two ⁶⁴Cu bis(thiosemicarbazone) complexes ([⁶⁴Cu]ATSM and [⁶⁴Cu]ATSE) to ¹⁸F-fluoromisonidazole ([¹⁸F]FMISO) and ¹⁸F-fluoroazomycin arabinoside ([¹⁸F]FAZA) in a rodent model of stroke. (26) Surprisingly, the investigators did not observe significant uptake in the lesions with either of the copper radiopharmaceuticals.

Clinical Imaging of Hypoxia

Clinical evaluation of Cu-ATSM began in patients with solid tumors; however, more recently, patients with hypoxia from conditions other than tumors have been evaluated. While several different types of tumors have been evaluated with Cu-ATSM, its non-oncology evaluation has been limited to stroke and Parkinson’s disease.

Clinical Oncology

Tumor hypoxia has been shown to adversely impact treatment outcomes in several solid tumors, including non-small cell lung cancer (NSCLC). (27) In a study of patients with resectable NSCLC, intraoperative measurements of normal lung and tumor pO2 using Eppendorf polarographic electrodes has shown that tumors had generally lower pO2 than normal lung (median tumor-to-normal lung pO2 of 0.13). In addition, tumor hypoxia, elevated osteopontin levels, a hypoxic marker, and CD44 expression correlated with poor prognosis. As measurement of hypoxia is important both for predicting prognosis and in selection of the mode of therapy, several studies have evaluated patients with NSCLC. Takahashi et al. studied 6 patients with lung cancer and 4 normal subjects with [⁶²Cu]ATSM-PET. (28) They demonstrated rapid clearance of [⁶²Cu]ATSM from the blood, reaching a stable activity after a few minutes. [⁶²Cu]ATSM uptake was low in normal lung parenchyma (0.43 ± 0.09, tumor-to-arterial blood activity (T/B) ratio). However, increased activity was noted in all lung cancers (T/B: 3.00 ± 1.50). The investigators compared the [⁶²Cu]ATSM uptake with [¹⁵O]water and [¹⁸F]FDG uptake and found a negative correlation between blood flow, as measured by [¹⁵O]water, and flow-normalized [⁶²Cu]ATSM uptake in 3 of 4 patients studied with both tracers. In agreement with data from preclinical studies, this suggested an increase in retention of ⁶²Cu-ATSM in regions of low flow. The investigators found no correlation between [¹⁸F]FDG and [⁶²Cu]ATSM uptake, suggesting that these tracers measure different aspects of tumor characteristics. In a small study of patients with documented or suspected lung cancer, Zhang et al evaluated tumor perfusion with copper-pyruvaldehyde-bis(N4-methylthiosemicarbazone) ([⁶²Cu]PTSM) and compared it with [⁶²Cu]ATSM. The investigators demonstrated that the [⁶²Cu]ATSM-to-[⁶²Cu]PTSM activity ratio in the lung lesions was < 1 (range 0.45 to 0.85). (29) In addition, the authors found that when the [⁶²Cu]ATSM-to-[⁶²Cu]PTSM ratio increases, as a marker of increasing hypoxia in the tumors, plasma concentrations of epidermal growth factor decreases (r = −0.87; R2 = 0.76). The relationship of tumor hypoxia and response to therapy was evaluated in a group of patients with NSCLC. The authors found that [⁶⁰Cu]ATSM uptake was predictive of response to therapy in patients (n = 14) with NSCLC. (30) Tumor uptake of [⁶⁰Cu]ATSM, as measured by tumor-to-muscle (T/M) ratio, was significantly lower in responders than in nonresponders (1.5 ± 0.4 vs. 3.4 ± 0.8, p = 0.002). However, the mean SUV for [⁶⁰Cu]ATSM was not significantly different in responders and nonresponders (2.8 ± 1.1 vs. 3.5 ± 1.0, p = 0.2). The authors found that an arbitrarily selected T/M threshold of 3.0 discriminated patients who are likely to respond to therapy: all 8 responders had a T/M < 3.0 and all 6 nonresponders had a T/M ≥ 3.0. In addition, tumor [¹⁸F]FDG uptake, as measured by SUV, was not significantly different in responders and nonresponders (10.9 ± 4.1 vs. 12.7 ± 10.4, p = 0.7) and did not correlate with [⁶⁰Cu]ATSM uptake (r = 0.04; p = 0.9). Thus, in this patient population, the tumor uptake of [⁶⁰Cu]ATSM revealed clinically unique information predictive of tumor response to therapy. Recently Lohith et al. studied 8 patients (5 with squamous cell carcinoma (SCC) and 5 with adenocarcinoma) with [⁶²Cu]ATSM-PET and [¹⁸F]FDG PET on 2 separate days within a 1 week interval. (31) The [⁶²Cu]ATSM and [¹⁸F]FDG images were coregistered and multiple small regions of interest (10 mm in diameter) were drawn on tumors to obtain SUV values of the 2 tracers. The regression lines were determined between SUV values for [⁶²Cu]ATSM and [¹⁸F]FDG in each tumor. SUVs for [⁶²Cu]ATSM were lower than those for [¹⁸F]FDG in both SCC and adenocarcinoma. The authors demonstrated that SCC tumors showed high [⁶²Cu]ATSM and low [¹⁸F]FDG uptake in the periphery of the tumors but low [⁶²Cu]ATSM and high [¹⁸F]FDG uptake toward the center. The SUV values for the 2 tracers were negatively correlated with negative regression slopes, indicating that the spatial distributions of [⁶²Cu]ATSM and [¹⁸F]FDG were significantly different within the SCC tumors (spatial mismatching). However, tumors with adenocarcinoma histology had a spatially similar distribution of [⁶²Cu]ATSM and [¹⁸F]FDG with positive regression slopes, indicating that the spatial distributions of [⁶²Cu]ATSM and [¹⁸F]FDG were similar within the tumor (Figure 1A and B). The regression slopes for [⁶²Cu]ATSM and [¹⁸F]FDG differed significantly between SCC and adenocarcinoma (p < 0.001). Thus, intratumoral regions of high glucose metabolism and hypoxia could differ and be dependent on the histologic type of lung cancer. This information could play a significant role in the selection of the mode of therapy in patients with NSCLC.

Figure 1.

Left image: Transaxial PET images of [⁶²Cu]ATSM (A), [¹⁸F]FDG (B), fused (C) and corresponding CT image (D) of 76-y-old man with adenocarcinoma in left lower lung. Fused image is depicted for [⁶²Cu]ATSM-PET in color and for [¹⁸F]FDG-PET in gray scale. Both original tumor and lymph node metastasis showed similar uptake patterns between [¹⁸F]FDG and [⁶²Cu]ATSM.

Right image: Transaxial PET images of [⁶²Cu]ATSM (A), [¹⁸F]FDG (B), fused (C) and corresponding CT image (D) of 59-y-old man with SCC at right hilum. Fused image is depicted for [⁶²Cu]ATSM-PET in color and for [¹⁸F]FDG PET in gray scale.

Reprinted with permission from the Journal of Nuclear Medicine (31)

Tumor oxygenation has been recognized as an important predictive parameter for survival in cervical cancer. Höckel et al. demonstrated that pretreatment hypoxia, as measured with oxygen electrodes, in cervical cancer is associated with a shorter recurrence-free survival and overall survival. (32) Tumor uptake of [⁶⁰Cu]ATSM has been correlated with outcome in a small group of patients with advanced SCC of the cervix (FIGO stages IB1 to IIIB). Dehdashti et al., found that tumor uptake of [⁶⁰Cu]ATSM was inversely related to progression-free and overall survival (log-rank test p = 0.0005 and p = 0.015, respectively) (33). An arbitrarily selected T/M threshold of 3.5 discriminated patients who were likely to develop recurrence. In addition, the frequency of locoregional nodal metastasis was higher in hypoxic tumors (p = 0.03). These patients also underwent [¹⁸F]FDG-PET; however, there was no significant difference in tumor [¹⁸F]FDG uptake between patients with hypoxic tumors and those with normoxic tumors. Later, in a larger number of patients with advanced cervical cancer (FIGO stages 1B1 to IV) who were studied with [⁶⁰Cu]ATSM-PET and [¹⁸F]FDG-PET prior to radiotherapy and chemotherapy, (34), the authors confirmed their prior results that tumor [⁶⁰Cu]ATSM uptake was inversely related to progression-free survival and cause-specific survival (p = 0.006 and p = 0.04, respectively, log-rank test). The previously used T/M threshold of 3.5 was the best discriminator of patients likely to develop a recurrence from those unlikely to develop a recurrence. The 3-year progression-free survival of patients with normoxic tumors (as defined by T/M of ≤ 3.5) was 71%, and that of patients with hypoxic tumors (T/M of > 3.5) was 28% (p = 0.01). Similarly, tumor [¹⁸F]FDG uptake did not correlate with ⁶⁰Cu-ATSM uptake, and there was no significant difference in tumor [¹⁸F]FDG uptake between patients with hypoxic tumors and those with normoxic tumors (p = 0.9). Thus, the pretherapy [⁶⁰Cu]ATSM-PET provided clinically important information about tumor oxygenation that was predictive of outcome in this patient population. Examples of [⁶⁰Cu]ATSM uptake in hypoxic and normoxic tumors and differences in the pattern of [¹⁸F]FDG uptake and [⁶⁰Cu]ATSM uptake are shown in Figures 2 and 3. In a small group of patients with advanced cervical cancer, Grigsby et al. demonstrated that most hypoxic tumors, as measured by [⁶⁰Cu]ATSM, overexpress vascular endothelial growth factor (VEGF), epidermal growth factor receptor (EGFR), cyclo-oxygenase-2 (COX-2), and carbonic anyhdrase IX (CA-9) while overexpression of these markers was not noted in non-hypoxic tumors. (35) They also found that more than 30% of tumor cells were positive for apoptosis, as measured by TUNEL assay, in 5 of 6 patients with hypoxic tumors, while this level was present in only 1 of 9 non-hypoxic tumors (p = 0.005). In this study, hypoxic tumors, as measured by [⁶⁰Cu]ATSM, had shorter overall survival than those with normoxic tumors; the 4-year overall survival estimates were 75% for patients with non-hypoxic tumors and 33% for those with hypoxic tumors (Log rank Mantel-Cox; p = 0.04).

Figure 2.

Cervical Cancer: Transaxial CT (left), [¹⁸F]FDG-PET (top right) and [⁶⁰Cu]ATSM-PET (30–60 min summed images, bottom right) images of the pelvis demonstrate increased [¹⁸F]FDG uptake within the known primary tumor at the site of the cervical mass seen on CT (white arrow, SUVmax = 17.1) and also increased [⁶⁰Cu]ATSM uptake within the known primary cervical cancer (white arrowhead, tumor-to-muscle uptake ratio= 10.4), however, the uptake of [⁶⁰Cu]ATSM retained more avidly in the periphery of the tumor.

Figure 3.

Cervical Cancer: (A) Transaxial CT (top left) and fused [¹⁸F]FDG-PET/CT (top middle) and [¹⁸F]FDG-PET (top right) images of the pelvis show intense [¹⁸F]FDG uptake within the known cervical tumor at the site of the cervical mass seen on CT. (B). Transaxial CT (bottom left) and fused 30–60 min summed images of ⁶⁴Cu-ATSM-PET/CT (lower middle) and [⁶⁴Cu]ATSM-PET (lower right) of the pelvis at the same level demonstrate increased uptake within the primary cervical cancer, but more avidly in the periphery of the tumor.

The current standard of care treatment for locally advanced rectal cancer consists of neoadjuvant radiotherapy and chemotherapy followed by rectal resection. However, not all patients benefit from conventional chemoradiation and, thus, it is important to identify patients who may not benefit from such therapy prior to initiation of therapy and treat such patients with alternative regimens. The effect of tumor hypoxia on response to neoadjuvant therapy is relevant to the management of locally-advanced rectal cancer. Presence of hypoxia in rectal cancer has been documented with polarographic needle electrodes (36). In addition, immunohistochemical analysis of the tumor tissue for hypoxic markers demonstrated that overexpression of these markers predicts decreased overall survival in locally advanced rectal cancer treated by neoadjuvant chemoradiation and rectal resection. (37) In a group of patients with locally advanced rectal cancer, tumor [⁶⁰Cu]ATSM uptake was found to be greater in patients with worse prognosis compared with those with better prognosis; the overall and progression-free survivals were worse with hypoxic tumors than with normoxic tumors (both p < 0.05) (38) (Figure 4). In addition, ⁶⁰Cu-ATSM uptake within the rectal cancer was predictive of downstaging following neoadjuvant chemoradiation; the mean T/M ratio for [⁶⁰Cu]ATSM uptake was significantly lower in downstaged tumors than that of non downstaged tumors (2.2 vs. 3.3, respectively) (p = 0.03).

Figure 4.

Rectal Cancer: Transaxial CT (left), [¹⁸F]FDG-PET (top right) and [⁶⁰Cu]ATSM-PET (30–60 min summed images, bottom right) images of the pelvis demonstrate increased [¹⁸F]FDG uptake SUV_max = 5.7) within the known primary tumor at the site of the rectal mass seen on CT (white arrow), and also increased [⁶⁰Cu]ATSM uptake within the known primary rectal cancer (white arrowhead, tumor-to-muscle uptake ratio= 2.8).

Head and neck cancer is the sixth most common malignancy worldwide and the majority of these patients have locally advanced disease at diagnosis (39). Head and neck cancer is typically treated with a combination of treatment strategies including surgery, radiation and chemotherapy. However, despite multimodality treatment, 40% to 60% of the patients with locally advanced squamous cell cancer, which is the most common histology, will relapse. (40) Hypoxia has been recognized to be a poor prognostic factor in head and neck cancer and approximately 40% of these cancers have been shown to have hypoxia with a significant intertumor variation of oxygen tension. (2, 41, 42) Hypoxia protects tumor cells from radiation that is adequate to eliminate oxygenated tumor cells. One approach to overcome tumor resistance due to hypoxia is to escalate radiation dose to the hypoxic regions of the tumor using intensity-modulated radiation therapy approach. Chao et al. in a feasibility study demonstrated that hypoxia mapping of head and neck cancer is possible using [⁶⁰Cu]ATSM-PET and the images can be used to guide IMRT through coregistering [⁶⁰Cu]ATSM-PET images to the corresponding CT images. (43) Minagawa et el. in a pilot study demonstrated that [⁶²Cu]ATSM uptake before initiation of chemoradiation is predictive of response to therapy in patients with locally advanced head and neck cancers. (44) Tumor [⁶²Cu]ATSM uptake, as measured by SUV_max, before therapy was significantly higher in patients with residual/recurrent disease than those who were free of disease at 2 years following therapy (p < 0.05). Six of the 10 patients with tumors’ SUV_max > 5 had residual/recurrent tumors, whereas all of the 5 patients with tumors’ SUV_max < 5 were free of disease. They found no difference in [¹⁸F]FDG uptake between patients with or without residual/recurrent disease. Kositwattanarerk et al. found a negative correlation between regional [¹⁸F]FDG uptake and [⁶²Cu]ATSM uptake in SCC of head and neck; [⁶²Cu]ATSM accumulation was higher in the periphery than the center of the tumor while FDG uptake was high in the entire tumor. (45) However, the uptake of [¹⁸F]FDG and [⁶²Cu]ATSM was similar and homogenous in head and neck adenocarcinoma. These results were similar to that reported by Lohith et al. in NSCLC. (31) The exact mechanism for the difference in the pattern of [¹⁸F]FDG and [⁶²Cu]ATSM uptake in different tumor histology is not well understood, but it has been proposed that it may be related to the differences in tumor aggressiveness and growth or possibly because adenocarcinoma has diffuse microenvironmental changes in the entire mass that leads to a matched pattern of increased uptake for both tracers. Thus, the histology may be an important determinant of whether hypoxia imaging is necessary and, thus, such imaging can be reserved for the tumor histology that is expected to have a different pattern that is revealed by [¹⁸F]FDG-PET.

It is known that high grade gliomas are associated with microvascular proliferation and/or hypoxia and necrosis. (46, 47) Direct measurement with oxygen electrodes has shown a lower oxygen pressure in tumor than that in the surrounding cortical tissue (48). There is a very limited number of reports in using [Cu]ATSM for evaluation of tumor hypoxia in brain tumors. Tateishi et al. demonstrated that grade IV gliomas have high [⁶²Cu]ATSM uptake in comparison to grade III gliomas (p = 0.014) and no significant difference was seen between grades III and II (49). In addition, [⁶²Cu]ATSM uptake, as measured by tumor-to-normal brain (T/B) ratio showed a significant correlation with hypoxic marker, HIF-1a (p = 0.001); the mean T/B ratio cutoff threshold of 1.8 was highly predictive of HIF-1a expression, as assessed by immunohistochemical staining (sensitivity of 92.3% and specificity of 88.9%). Using a T/B ratio cutoff threshold of 1.8, regional uptake of [⁶²Cu]ATSM was seen in nearly 62% (13/21) of tumors within the contrast-enhanced region on MR images, which was significantly correlated with presence of a necrotic component (p = 0.002). Thus, [⁶²Cu]ATSM may be used to identify highly malignant brain tumors with greater degree of hypoxia. In another study, Hino-Shishikura et al. demonstrated that tumor hypoxia, assessed by [⁶²Cu]ATSM-PET/CT, correlates with microscopic diffusion capacity, assessed by diffusion weighted imaging (DWI), in primary brain tumors. (50) DWI is a special MRI sequence that has been successfully used for imaging acute brain ischemia and apparent diffusion coefficients (ADC) derived from DWI have been used to grade primary brain tumors. (51, 52) Patients with low-grade glioma, glioblastoma (GBM) and primary central nervous system lymphoma were studied; high [⁶²Cu]ATSM uptake and high DWI signal intensity were seen in GBM and lymphoma whereas low [⁶²Cu]ATSM uptake and low DWI signal intensity were seen in low-grade gliomas. In addition, an inverse correlation was noted between [⁶²Cu]ATSM uptake, as measured by SUV_max and T/B ratio, and ADC_min (p < 0.0001 and p < 0.0001, respectively). However, GBM and lymphoma could only be differentiated based on SUV_max, as the mean ADC_min was not significantly different in GBM and lymphoma, but the mean SUV_max was significantly greater in lymphoma than in GBM (p = 0.90 vs. p < 0.0001). It is unclear whether this is related to the differences in the amount of necrosis or vascularity found in these tumors.

Recent advances in radiation therapy such as instance intensity modulated radiotherapy (IMRT), dose painting and image-guided radiotherapy have resulted in radiation dose delivery to achieve a complex dose distribution for cancer treatment. In particular, subvolume boosting to the radioresistant regions within the target volume have shown to be effective. (53, 54) As hypoxia is very important in response to radiation, significant attention has been focused on using hypoxic PET imaging for radiation planning of solid tumors. However, the methodology, in particular dose painting, is complicated. Investigators have proposed methods for creating voxel-based dose painting plans using the clinical TomoTherapy Hi-Art II treatment planning system and dose conformity using the RapidArc optimizer and beam delivery technique using [⁶¹Cu]ATSM-PET/CT. In the future, boosting to the hypoxic subvolume of the tumor may prove to be effective. (55),(56)

[⁶⁰Cu]ATSM versus [⁶⁴Cu]ATSM: Most clinical studies have used ATSM labeled with copper-60 that has a relatively short half-life of 0.395 h (92.5% β+-decay and 7.5% electron capture) (30, 33, 34, 38) and, thus, prevents its widespread use in centers without an in-house cyclotron. However, other copper positron emitting nuclides such as [⁶⁴Cu] (half-life of 12.7 h; 17.4% β+-decay, 38.5% β −, 43% electron capture), or [⁶⁰Cu] (half-life of 3.33 h; 62% β+-decay, 38% electron capture) with longer half-lives can be used, which allow for effective distribution to the facilities without in-house cyclotron. In order to compare PET image quality and tumor uptake with [⁶⁰Cu]ATSM and [⁶⁴Cu]ATSM, Lewis et al. studied 10 women with cancer of the uterine cervix. (57) They found that the image quality with [⁶⁴Cu]ATSM was better than that with [⁶⁰Cu]ATSM because of lower noise and also the pattern and magnitude of tumor uptake of [⁶⁰Cu]ATSM and [⁶⁴Cu]ATSM on studies separated by 1 to 9 days were similar. Thus [⁶⁴Cu]ATSM is an appropriate substitute for [⁶⁰Cu]ATSM and its longer half-life allows for commercialization and widespread delivery of this tracer. Currently, the American College of Radiology Imaging Network (protocol 6682) in a multicenter trial is evaluating the value of [⁶⁴Cu]ATSM in predicting response to therapy and patient outcome.

Clinical Non-Oncology

Cu-ATSM-PET studies in non-oncologic settings have been limited to the field of neurology. It has been shown that Cu-ATSM has the potential to assess oxidative stress in a variety of neurological disorders. (58–60) Ikawa et al. used [⁶²Cu]ATSM-PET and [¹⁸F]FDG-PET to assess the changes in the regional oxidative stress, blood flow and glucose metabolism in a patient with stroke-like episodes (MELAS) whose brain showed three different stages of stroke like lesions; acute, subacute and chronic. (58) They performed a 20-min dynamic [⁶²Cu]ATSM-PET. [⁶²Cu]ATSM accumulation in the initial 3 min reflected cerebral blood flow and [⁶²Cu]ATSM accumulation in the delayed phase (last 10 min of imaging) reflected over-reduction status, leading to oxidative stress. The acute lesion demonstrated mildly increased cerebral blood flow, as measured by increase in early [⁶²Cu]ATSM uptake, and increased glucose metabolism, as measured by increase in [¹⁸F]FDG, with a tendency toward an oxidative stress, as demonstrated by mild increase in delayed [⁶²Cu]ATSM uptake. In the subacute lesion, the oxidative stress was markedly enhanced (increase in delayed [⁶²Cu]ATSM uptake), but there was decreased cerebral blood flow, as measured by decrease in early [⁶²Cu]ATSM uptake, and decreased glucose metabolism, as measured by decrease in [¹⁸F]FDG uptake. In the chronic lesion, there was markedly decreased cerebral blood flow, as measured by decrease in early [⁶²Cu]ATSM uptake, and markedly decreased glucose metabolism, as measured by decrease in [¹⁸F]FDG uptake, and oxidative stress had disappeared (decrease in delayed [⁶²Cu]ATSM uptake), presumably reflecting neuronal death. Therefore, the combined [⁶²Cu]ATSM-PET and [¹⁸F]FDG-PET studies in this patient demonstrated that changes in the oxidative stress can be assessed noninvasively in MELAS stroke-like episodes. [⁶²Cu]ATSM-PET has been used to study cerebral occlusive disease. (59) Isozaki et al. evaluated 10 patients with major cerebral occlusive disease. (59) The investigators assessed cerebral blood flow, blood volume, metabolic rate of oxygen (CMRO₂) and oxygen extraction fraction (OEF) by PET imaging using [¹⁵O], and C[¹⁵O] and [¹⁵O]water and compared them with early-phase (first 3 min of imaging), delayed-phase (last 10 min of the 20-min dynamic imaging) and delayed-to-early ratio of [⁶²Cu]ATSM uptake. They found that the early-phase [⁶²Cu]ATSM-PET images corresponded well to cerebral blood flow images, and the delayed-to-early ratio [⁶²Cu]ATSM-PET images were similar to OEF images. The regional values obtained from delayed-to-early ratio [⁶²Cu]ATSM-PET images for all hemispheres were significantly correlated with regional OEF (r = 0.73, p < 0.0005). The asymmetry index of OEF and delayed-to-early ratio showed a significant correlation (r = 0.91, p < 0.0001). The cerebral blood flow, CMRO₂ and early-phase SUV of [⁶²Cu]ATSM were significantly lower (p < 0.01, paired t test), and the OEF and delayed-to-early ratios of [⁶²Cu]ATSM were significantly higher (p < 0.05, paired t test) in the affected hemisphere of the patients compared with the contralateral hemisphere. Three of their patients showed significant elevation of OEF and were diagnosed to have misery perfusion in the affected cerebral hemisphere. The sensitivity, specificity and accuracy for detection of misery perfusion were 100%, 71.4%, and 80%, respectively using OEF asymmetry index and 100%, 85.7%, and 90%, respectively using [⁶²Cu]ATSM asymmetry index. Thus, this study demonstrated the potential of dynamic [⁶²Cu]ATSM-PET for detection of misery perfusion with a single radiotracer injection and only a 20-min imaging.

[⁶²Cu]ATSM-PET also has been used in Parkinson’s disease to assess the oxidative stress due to mitochondrial dysfunction in the striata of patients with Parkinson’s disease as oxidative stress is assumed to contribute to the degeneration of dopaminergic neurons in the nigrostriatal system during Parkinson’s disease. (60) Ikawa et al. studied 15 patients with Parkinson’s disease and 6 healthy controls with [⁶²Cu]ATSM-PET. (60) They used delayed phase (last 10 min of the 20-min dynamic imaging) for analysis and the striatum-to-cerebellum SUV ratio (S/C ratio) was calculated. The mean S/C ratio of the bilateral striata of the patients was significantly increased compared with that of the controls (p < 0.05). In addition in patients with Parkinson’s disease, they found a positive correlation between the S/C ratio of the bilateral striata and the Unified Parkinson’s Disease Rating Scale (UPDRS) rating (r = 0.52, p < 0.05) as well as between the S/C ratio of the striatum contralateral to the initially affected body side and the UPDRS rating (r = 0.62, p < 0.05). This indicates that [⁶²Cu]ATSM-PET has the ability to show the increase in striatal oxidative stress in Parkinson’s disease patients compared with the controls. Moreover, the uptake increased with the progression of disease severity. Thus, [⁶²Cu]ATSM-PET may prove to be useful in monitoring disease activity during therapy.

Summary

Cu-ATSM has continued to demonstrate value for the delineation of hypoxia in numerous diseases. As with all tracers its benefit and application is not universal but it is clear that when applied in the correct setting its adds tremendous value to the clinical findings. Continued work both in the preclinical and clinical settings is fully warranted.