Introduction

Paediatric gliomas are the most common central nervous system neoplasms in children [1]. Unlike their adult counterparts, paediatric-type diffuse high-grade gliomas (PDHGG) are rare, accounting for only 5–20% of all gliomas [2,3,4]. The prognosis of these forms is dismal, with median overall survival ranging between 9 and 16 months for PDHGG and less than 1 year for diffuse intrinsic pontine glioma (DIPG) [3, 5,6,7,8,9].

Complete surgical resection of these infiltrative tumours is virtually impossible, and surgery is typically not attempted in patients with DIPG, owing to the location of the tumour [10,11,12]. Concurrent adjuvant radiotherapy, in combination with chemotherapy, is the standard of care for newly diagnosed PDHGG, but still less than 5% of patients survive longer than 5 years post-diagnosis [13, 14]. Since no chemotherapeutic drugs have proven effective in the treatment of DIPG, radiation therapy is the current standard regimen [15,16,17,18].

Despite dramatic improvements in the genetic and epigenetic analyses of PDHGG [4], we are still in the early stages of developing gene-targeted therapies. An alternative approach is to focus on the tumour cell metabolism; in this context, molecular imaging could provide additional information on the biological behaviour of the tumour and possibly identify new targets for a tailored approach [19, 20]. Copper is a key element in cellular turnover, being a vital co-enzyme in a number of cell functions, including mitochondrial respiration [21, 22]. Unsurprisingly, this element plays a major role in tumorigenesis and in cancer metabolism; more specifically, it has been found that copper is linked with mitogen-activated kinase stimulation and, in particular, with BRAF and K-RAS [23,24,25]. Moreover, copper is involved in cancer mitochondria-dependent energy production [26], tumour invasiveness, and even chemotherapy resistance [27]. By contrast, the normal brain parenchyma has little need for this element. Imaging aggressive brain neoplasms with a copper radioisotope might provide a favourable target-to-background ratio.

In particular, 64Cu could play both diagnostic and therapeutic roles, as its decay scheme entails the production of positrons, high-energy beta particles and Auger electrons with high linear energy transfer (LET) [28,29,30]. Indeed, this high-LET radiation could play the most important role in the theranostic effect of [64Cu]CuCl2 [31]. This tracer [64Cu]CuCl2 PET/CT has been proposed as a promising procedure for identifying adult high-grade gliomas, which often display intense tracer uptake [32].

This pilot trial aimed to evaluate the potential diagnostic role of [64Cu]CuCl2 PET/CT in patients with PDHGG by comparing PET images with MRI. We aimed to estimate kinetics, tumour/background ratio (TBR) over time, the absorbed dose of [64Cu]CuCl2 in gliomas and in organs, as well as the effective dose.

Material and Methods

Radiopharmaceutical

Copper-64 dichloride ([64Cu]CuCl2, average specific activity, 3700 MBq/µg, radiochemical purity > 99%, radionuclide purity > 99%) was produced according to a procedure previously reported [32, 33]. Briefly, 64Cu was produced by bombarding an electroplated 64Ni target using a proton current of 18 µA and energy of 14.6 MeV. Following bombardment, 64Cu was purified from other contaminants by means of chromatography and an ion-exchange column (Biorad Laboratories). The radioisotope was eluted with concentrated HCl and sieved through a 0.2-µm filter (Merck Millipore). Radionuclide purity and 64Cu half-life were measured by means of an HPGe detector (Ortec), by identifying the characteristic 511, 1022, and 1345.8 keV photopeaks. A radionuclide purity ≥ 99.5% was considered acceptable. Radiochemical purity (RCP) was assessed by having [64Cu]CuCl2 react with the tetraazacyclotetradecane-N, N′, N″, N′′′-tetraacetic acid ligand; a purity ≥ 99% was deemed acceptable.

Patient Population and Diagnostic Protocol

The local ethics committee (Comitato Etico Regionale Liguria Registration Number: 076/2019) and the “Agenzia Italiana del Farmaco” (Italian Drug Agency) approved this study. All children’s legal guardians signed a written informed consent form. The trial was registered in the European Clinical Trial Database (EudraCT number 2018–004,667-30). The neuro-oncological departments of three Italian children’s hospitals (Istituto G. Gaslini in Genoa, Bambino Gesù Children’s Hospital in Rome and AORN Santobono-Pausilipon Hospital in Naples) prospectively enrolled patients with recurrent/progressive PDHGG according to the inclusion/exclusion criteria (Table 1).

Table 1 Inclusion and exclusion criteria of the population
Full size table

[64Cu]CuCl2 was injected into a cubital vein (median activity: 179 MBq, range 113–280 MBq, corresponding to a median of 3.9 MBq/kg, range 2.4–9.4 MBq/kg). Brain PET/CT acquisition was started 60 min thereafter and lasted 30 min. Subsequently, whole-body PET acquisition was started. Further acquisitions (i.e., brain and whole-body) were performed 24 and 72 h after the injection. In all patients, an MRI examination, including fluid attenuation inversion recovery (FLAIR), T2-weighted and pre- and post-contrast (0.1 mmol/kg, macrocyclic ionic agent) T1-weighted images, was performed. All patients were acquired on a Discovery ST PET/CT device (General Electric Healthcare Technologies, Milwaukee, WI, USA) and on a 1.5 Tesla MRI scanner (Intera Achieva, Philips, Best, the Netherlands).

Image Analysis: Registration and Voi Segmentation

PET/CT images were evaluated visually and semi-quantitatively in a patient-by-patient and lesion-by-lesion analysis after fusion of the images with MRI images (T1, T2 and FLAIR). Fusion was obtained automatically by means of a commercially available image-registration software tool (AW Server, General Electric Medical Systems). As previously described [31], volumes of interest (VOIs) were constructed on the L4-L5 vertebral bodies to assess the biokinetics and dosimetry of the active bone marrow. Subsequently, the active bone marrow in the whole body was calculated as a function of body weight [34]. VOIs were also delineated at the venous access site, to estimate the net activity administered.

In our analysis, the [64Cu]CuCl2 PET detection rate (DR) was defined as the ability to detect at least one pathological finding in each individual subject. The characteristics and extent of the tumour on conventional MRI were also correlated with [64Cu]CuCl2 PET uptake. Given the diffuse nature of the disease, tumour extent was delineated on the basis of T2/FLAIR MRI signal abnormalities. Any area of contrast enhancement within each lesion was also reviewed and correlated with [64Cu]CuCl2 PET uptake.

PET/CT was rated as positive if tumours identified on MRI exhibited tracer uptake above the level of the corresponding contralateral or remote normal brain tissue.

Kinetics of Tumours and Organs

The activity concentration of [64Cu]CuCl2 (as a percentage of injected activity/ml) in VOIs of tumours and organs for all PET datasets was recorded and fitted to a mono-exponential function or with the trapezoidal methods [35]. Hence, the time-integrated activity and its coefficient were determined. The mean activity concentration inside the tumour was corrected for the partial volume effect (PVE), as previously described [31]. The protocol is based on numerical recovery coefficients that are experimentally derived by using radioactive phantoms. Finally, the average standardised uptake values (SUV) in tumours over time was calculated and expressed as mean ± standard deviation.

Tumour/Background Ratio (TBR)

The tumour-to-background ratio (TBR) of all lesions was defined as the relation between the mean activity concentration of the lesion and the one of the background tissue; it was expressed as mean ± standard deviation. The background radioactivity concentration was obtained by calculating the mean value of VOIs drawn at 1-cm distance from each detectable lesion, at the four cardinal points.

Dosimetry of Tumours and Organs

Tumour and organ dosimetry was carried out by calculating the absorbed dose coefficient (absorbed dose per administered activity) according to the Medical Internal Radiation Dose (MIRD) system [36, 37]. The OLINDA/EXM dosimetry software was used to calculate mean organ absorbed doses based on reference anatomic models, including the reference newborn, 1-year-old, 5-year-old, 10-year-old, 15-year-old, and adult female and adult male [38]. Tumour dosimetry was calculated with the dose factors of sphere-shaped phantoms. Finally, the absorbed dose coefficients were obtained by multiplying the dose factors by the time-integrated activity coefficients. The effective dose was calculated on the basis of the tissue weighting factors of the organs (ICRP protocols 60 and 103) [39, 40].

Results

Patient Population

We prospectively enrolled ten patients (median age 9, range 6–16 years, 6 females), according to the inclusion criteria (Table 1). All patients had a Karnofsky performance status above 80%. Injection of the radiopharmaceutical was well tolerated, and no immediate or late side DIPG was established on the basis of clinical and MRI criteria, in accordance with RAPNO guidelines [41]. In two subjects with diffuse high-grade hemispheric gliomas, the WHO grade was determined according to histological features, owing to the lack of molecular information. Most patients (n = 8) were affected by diffuse midline gliomas (five H3K27-altered). Six of the ten patients had a positive [64Cu]CuCl2 PET/CT scan (all patients’ characteristics are reported in Table 2). In all patients, no uptake was detected in the normal parenchyma, while intense uptake was observed in the structures located outside the blood–brain barrier (e.g., choroid plexus and hypophysis) (Fig. 1).

Table 2 General characteristics of the patient population
Full size table
Fig. 1
figure 1

a, b [64Cu]CuCl2 PET/MR imaging fusion. Physiological uptake of [64Cu]CuCl2 in normal brain structures outside the blood–brain barrier (pituitary gland and choroid plexus).

Full size image

Diagnostic Evaluation and Lesion Analysis

On MR imaging, all gliomas showed increased T2/FLAIR signal intensity and variable (iso- to hypointense) appearance on T1-weighted imaging. After administration of contrast material, six tumours displayed contrast enhancement [eight contrast-enhancing areas, five of which exhibited signs of necrosis (ring-shaped enhancement)].

Regarding [64Cu]CuCl2 PET/MR imaging fusion, increased uptake was concordant with areas of contrast enhancement in all lesions, which, in some cases, was associated with ring enhancement. Infiltrative components without contrast enhancement did not show increased uptake (Table 2, Fig. 2). The [64Cu]CuCl2 uptake pattern was heterogeneous. Specifically, the highest avidity for [64Cu]CuCl2 was displayed in tumour areas with contrast enhancement along the margins of necrotic components (Fig. 3). Three non-necrotic contrast-enhancing areas displayed low uptake.

Fig. 2
figure 2

a FLAIR image. b Post-contrast T1-weighted image. c [64Cu]CuCl2 PET/MR imaging fusion. Subject 9. Absence of tracer uptake in a non-enhancing diffuse intrinsic pontine glioma (arrows, a–c).

Full size image
Fig. 3
figure 3

a Post-contrast T1-weighted image. b, c [64Cu]CuCl2 PET/MR imaging fusion. Subject 4. Necrotic component displaying ring enhancement (arrow, a) with intense tracer uptake (arrow, b) in a diffuse hemispheric glioma. On later imaging 24 h after injection, the tracer intensity within the lesion had increased (arrow, c).

Full size image

Five patients, with a total of seven areas with increased tracer uptake, underwent at least one further PET, 24 h after tracer injection; two of these five subjects (four lesions in total) underwent a third PET acquisition 72 h after the injection. In all these patients, the mean SUV increased over time (0.9 ± 0.5, 1.2 ± 0.7 and 1.8 ± 0.9 at the first, second and third time points, respectively). Likewise, mean TBR markedly increased over time (5.1 ± 3.5, 6.8 ± 3.7, and 11.3 ± 9.6). Subject 10’s lesion, which was located in the thoracic spinal cord, had a lower TBR than the others, although the uptake was about average, owing to the higher background uptake caused by the proximity to the liver. See Table 3 for details.

Table 3 Mean SUVmean and target-to-background ratio of all 64CuCl2 positive lesions
Full size table

[ 64 Cu]CuCl 2 Distribution in Normal Organs and Tumours

Among the organs considered, the liver showed the highest uptake (mean percent injected activity: 38.9%). By contrast, all other organs displayed less marked tracer affinity, ranging from 0.17% (salivary glands) to 2.1% (kidneys). See Supplemental Fig. 1 for an outline. We also observed that the specific concentration depended on the patient’s body weight, as can be seen from the decreasing trend in uptake with increasing weight in all organs (Fig. 2).

In general, the tumours showed moderate copper avidity, which was lower than that of most of the normal organs, except for the active bone marrow and the cerebral parenchyma. Figure 4 depicts the specific concentration of [64Cu]CuCl2 in the organs of all patients and the tumour concentration.

Fig. 4
figure 4

[64Cu]CuCl2-specific uptake (% activity/mL) in organs and tumours.

Full size image

Dosimetry Estimates in Organs and Tumours

The liver displayed the highest normalised absorbed dose coefficient (6.39E-1 mGy/MBq), followed by the other abdominal organs (gallbladder, pancreas, kidneys and spleen). By contrast, the tumours received only 3.44E-2 mGy/MBq. However, the tumour dose coefficient was widely variable, ranging from 2.80E-5 (in a hemispheric DPHGG, H3-wildtype and IDH-wildtype) to 7.5E-2 mGy/MBq (in two pontine DPHGG, H3-wildtype and IDH-wildtype). Supplemental Fig. 2 and Table 4 depict these data.

Table 4 Absorbed dose coefficients of [64Cu]CuCl2. Mean values over the patients are also displayed, together with the standard deviation
Full size table

The mean effective dose, according to the estimates provided by the ICRP protocols 60 and 103, was 6.51E-2 mSv/MBq and 6.10E-2 mSv/MBq, respectively.

Discussion

The present data are the result of what is, to the best of our knowledge, the first attempt to evaluate the potential diagnostic role, the biodistribution and the dosimetry of [64Cu]CuCl2 PET/CT in paediatric patients with diffusely infiltrating gliomas. In PDHGG, including diffuse pontine lesion, this radiopharmaceutical shows a characteristic distribution within tumours. Increased uptake was observed in paediatric gliomas with contrast-enhancing foci on MRI, and, in some cases, correlated with radiological evidence of tumour necrosis on MR imaging. This pattern is in line with previous preclinical and clinical studies performed by Pérès et al. [42] and Tateishi K et al. [43]. However, the extent of MRI contrast-enhancing areas with concomitant [64Cu]CuCl2 increased uptake was limited in comparison with non-enhancing infiltrative components (T2/FLAIR tumour extension), and four out of ten subjects (40%) with entirely non-enhancing high-grade gliomas did not show increased tracer uptake. Indeed, this radiopharmaceutical underestimates the overall extent of PDHHG in comparison with amino-acid PET tracers such as 18F-DOPA [44,45,46]. However, the kinetic characteristics of [64Cu]CuCl2 make this element potentially suitable for late imaging, dosimetry and targeted therapy.

In positive lesions, the [64Cu]CuCl2 uptake pattern was mixed. This could be due to differences in tumour-specific biological variables, and might indicate a heterogeneous oxygen micro-environment [43[]. At the same time, since [64Cu]CuCl2 uptake was concordant with contrast enhancement on MRI, blood–brain barrier damage seems to be a potential prerequisite for the concentration of this tracer within lesions. Nevertheless, the significant increase in mean SUV over time in positive lesions suggests that a simple passive mechanism of uptake due to loss of integrity of the blood–brain barrier is unlikely. The contrast-enhancing MRI pattern seen in our population is in line with the findings of previous multicentre studies of both paediatric non-brainstem high-grade gliomas (HERBY study) [47] and DIPG [48], in which imaging showed little or no enhancement in 33% and 36% of lesions, respectively.

Negative [64Cu]CuCl2 PET/CT lesions were all located in the midline. According to the 2016 and 2021 WHO classifications, these lesions are categorised as WHO grade IV, based on their molecular profile, including histologically low-grade H3K27-altered diffuse astrocytomas. In this context, accelerated cell proliferation could be variably present [49, 50]. Therefore, midline tumours with mild-to-moderate proliferation indices may not exhibit copper avidity, even in the presence of a diffusely infiltrating indicating a dire prognosis. On the other hand, the highest uptake was shown by a hemispheric PDHGG with a histological diagnosis of glioblastoma, which normally exhibits characteristics linked with accelerated proliferation [51].

Our study adds further insight into copper kinetics, in that the standardised uptake values increased over time, up to 3 days after administration. This pattern suggests that once incorporated, [64Cu]CuCl2 does not leave the tumour cells; such a prolonged accumulation is probably typical of gliomas since our previous experience with prostate cancer documented rapid clearance in the first hour after the maximum uptake [31].

Overall, on the basis of this pilot study, we suggest that [64Cu]CuCl2 PET/CT might be selectively used in paediatric gliomas showing contrast enhancement in the whole tumour volume; in such lesions, the prolonged accumulation might constitute the rationale for potential theranostic use. However, further larger studies are needed to test this concept. Specifically, in paediatric patients with BRAF V600E-mutant gliomas—a distinctive clinical-biological group of paediatric gliomas which typically present a contrast-enhancing MRI pattern—the role of [64Cu]CuCl2 in patients who do not respond to current treatment regimens [52] could be explored, given that copper is required for oncogenic BRAF signalling and tumorigenesis [23]. As demonstrated in a previous multicentre study of BRAF V600E-mutant paediatric high-grade gliomas, new strategies beyond BRAF-inhibitor monotherapy are needed [52].

From the dosimetry point of view, should a theranostic approach be considered, the liver is the limiting organ. Indeed, our analysis showed that, on average, the liver received about twenty times more energy than the tumour per MBq employed. However, previous evidence suggests that the 64Cu-mediated cell damage is dependent on the emission of Auger electrons, with this specific component delivering up to a 25-fold higher dose than beta radiation [31, 53, 54]. Given the low range of Auger electrons, the dose actually absorbed by the cell depends on whether the radiopharmaceutical is absorbed into the cell nucleus or not [55]. Should the tracer remain in the hepatic cell cytoplasm, then the dose absorbed by the liver cell would be determined by the beta components mainly, which could be less limiting in the therapeutic use of the radiopharmaceutical, especially if we also consider the relative resilience of the liver [56]. An ideal scenario for the theranostic application of [64Cu]CuCl2 would be one of intra-nuclear absorption in the tumour cells and a cytoplasmatic distribution in non-target tissue. Such an evaluation, however, requires the set-up of targeted experiments.

Finally, the dosimetry information provided by this study could be carried over to other disease settings. Particularly, the potential of this tracer could be evaluated in neuroblastoma, which is the most frequent extracranial neoplasm and whose cells express a copper transporter, playing a role in chemotherapy sensitivity [57,58,59].

Some limitations of this study must be borne in mind. Being a pilot, project aimed mainly at assessing the feasibility, tolerability and dosimetry of [64Cu]CuCl2 in children, it involved only a small number of patients. Moreover, as many of these patients had a midline glioma, the amount of data available on hemispheric glial tumours was limited. Finally, we did not have information on the late phase of all lesions in all patients, since ethics considerations prevented us from carrying out repeated imaging in cases without visible tracer uptake on the first examination.

Conclusions

[64Cu]CuCl2 is a safe tracer in the imaging of paediatric gliomas. It is selectively taken up by MRI contrast-enhancing/necrotic tumours, suggesting that blood–brain barrier damage is a prerequisite for tracer uptake. Moreover, this radiopharmaceutical shows excellent target-to-background contrast and an accumulating pattern over time. The possibility of employing this radiopharmaceutical for therapeutic applications could represent the subject of further research.

Supplementary Information.