Prediction of tumour hypoxia status from the washout of positron emitters in beam-monitoring PET: carbon-ion irradiation to tumour rat models

Abstract

Background

The aim of this study was to identify hypoxic regions in tumours using non-invasive imaging without specific radioligands. Tumour hypoxia, which varies during the course of radiation therapy, is an important predictor of treatment effect, therefore, frequent monitoring is required. However, positron emission tomography (PET) using hypoxia-specific radioligands is impractical to perform in every fractionated irradiation. Alternatively, blood flow can be evaluated by measuring washout rates of irradiation-induced positron-emitters in range-verification or beam-monitoring PET. Based on this, we hypothesized that the blood flow supply and kinetics (washout) of the irradiation-produced positron-emitters in the tumour could reflect the tumour oxygen status. Beam-monitoring PET for hypoxic tumour rat models was compared with conventional hypoxia PET. Glioma C6 cancer cells were transplanted into nude rats. Tumours were irradiated with a ¹²C ion beam, and beam-monitoring PET was performed using a high sensitivity total-body small-animal PET system. A compartment model was applied to the kinetic data of the irradiation produced positron-emitters. Then, ¹⁸F-labeled fluoroazomycin arabinoside (¹⁸F-FAZA) PET imaging of the hypoxia area of the tumour of the same rat was performed.

Results

The parametric images of washout rate obtained with beam-monitoring PET were in good agreement with ¹⁸F-FAZA-PET images.

Conclusion

This study demonstrated the biological washout can be used as diagnostic tool for tumour hypoxia in routine beam-monitoring PET.

Supplementary Information

The online version contains supplementary material available at 10.1186/s13550-026-01381-4.

Introduction

The use of positron emission tomography (PET) has been studied for in-vivo beam monitoring and treatment verification in charged particle therapy [1, 2]. PET-based beam monitoring is based on the detection of positron (β⁺) emitters which are produced by nuclear fragmentation interactions of the irradiated treatment ion beam (such as a proton beam and a heavy ion beam) with the irradiated tissue. The irradiation produced β⁺-emitters diffuse from the irradiated tumour tissue mainly via blood flow, which is referred to as the biological washout effect [3, 4].

Recently, a clear correlation between the biological washout rate and tumour vascular status was reported in two small animal irradiation studies [5, 6]. In these studies, the biological washout rate in viable tumour models was faster than that of normal models due to vascular hyper-perfusion which characterizes angiogenesis [7]. In the newly formed blood vessel walls corresponding to angiogenesis, gaps exist between endothelial cells through which molecules can pass [8]. So, it can be reasoned that the irradiation produced β⁺-emitters can speedily wash out via blood vessels in viable tumour tissue. On the other hand, the biological washout rate in the hypoxic tumour model was slower than that of the normal tissue, due to poor vasculature [3, 4]. In a spontaneous tumour, blood vessels form heterogenous structures due to imbalance between the rate of tumour growth and vascularization [9]. Cancer cells around tumour blood vessels are supplied with sufficient oxygen and grow actively. Meanwhile, in regions some distance from blood vessels, hypoxic regions appear. And in regions further away, necrosis occurs with complete oxygen deprivation. Thus the degree of oxygen supply in tumours strongly depends on the distance from the blood vessels [9], and this was quantified by Wÿffels et al. (2000) [10].

In terms of radiation therapy, it is known that the hypoxic tumour vascular mechanism frequently leads to therapeutic resistance [11–14]. In particle therapy, Linear energy transfer (LET)-painting has been suggested as a method to overcome tumour hypoxia [15, 16]. LET-painting attempts to restrict high-LET radiation to compartments that are found to be hypoxic, while applying lower LET radiation to normoxic tissues. LET painting offers an improved therapeutic effect compared to conventional particle therapy, which relies solely on dose optimization without LET optimization [17]. It should also be noted that tumour hypoxia status and the location of hypoxic area is constantly changing with the state of the tumour vasculature, which also changes with the tumour growths and/or radiation expose [13, 18–22]. Since tumour hypoxia status is changing during a treatment period, it is impractical to carry out hypoxia PET in every fractionated irradiation. Instead, it is hypothesized here that tumour hypoxia information can be extracted from the biological washout rate of irradiation produced β⁺-emitters. However, it is not clear whether the blood flow supply and washout can reflect the tumour vascular oxygen supply. Therefore, the aim of this study is to obtain conclusive evidence in hypoxic tumour rat models by comparing beam-monitoring PET and hypoxia PET. Because the washout rate of the irradiation-produced β⁺-emitting nuclides reflects the tumour vascular status, the washout rate of these nuclides serves as a biomarker for the tumour oxygen status. This has been supported by our small-animal study, which demonstrated the washout-based tumour diagnosis by distinguishing hypoxic tumour model from viable tumour model [23]. The aim of the present study was to detect the distribution of washout rates within the tumour and verify its relationship with the oxygen status.

In this study, a ¹²C ion beam is used to irradiate hypoxic tumour rat models, and beam-monitoring PET is performed by an in-house high sensitivity total-body small-animal PET (TBS-PET) system [24]. The biological washout rate constant is obtained from dynamic beam-monitoring PET images. In addition, a hypoxia PET tracer, ¹⁸F-labeled fluoroazomycin arabinoside (¹⁸F-FAZA) is administrated to the same rats to evaluate the tumour hypoxic status and a custom-made high resolution PET for small animals, the Mirai PET, is employed. Finally, the usefulness of beam-monitoring PET in particle therapy for the purpose of evaluating tumour vascular status is discussed.

Materials and methods

Tumour rate model

Animal experiments were approved by the local ethical committee of specialists in the National Institutes for Quantum Science and Technology (QST) and were carried out in compliance with the institute’s animal experimentation guidelines. In this study, a tumour rat model was prepared to include hypoxic and necrotic regions in the tumour. The development of C6 glioma vascular states which were transplanted into the subcutaneous region in rat has been well described in Deane and Lantos (1981), Vajloczy et al. (1998), and Forster et al. (2017) [25–27]. Tumour growth rate, vascular status, and density of the cancer cell mass depend on the number of transplanted C6 glioma cells. Following Vajloczy et al.’s procedure (1998) [26], 4 × 10⁶ of C6 glioma cells were subcutaneously injected into the right thigh of seven nude rats (9 weeks old, F344/NJcl-rnu/rnu CLEA Japan, Inc.). Implanted tumours in each rat reached diameters of 6.7–12.7 mm in two weeks.

Experimental design

The experimental design of the entire study is shown in Fig. 1. The tumour rat model was prepared as stated in Sect. "Tumour rate model". After the tumour grew, a ¹²C ion beam was irradiated to the whole tumour of each rat. Then, the rat was placed into the in-house high sensitivity TBS-PET system (Experimental set up) which was installed outside of the irradiation room, and PET scanning was performed (hereafter, beam-monitoring PET experiment). The biological washout rate constant of the irradiation produced β⁺-emitters was analyzed from the dynamic data of the beam-monitoring PET experiment.

Fig. 1.

Overview of the experimental design of this study

Within a half day after the beam-monitoring PET experiment,¹⁸F-FAZA was intravenously administrated in the same rats (Sect."Data processing for analysis"). One hour after the administration, PET scanning was performed (hereafter, ¹⁸F-FAZA PET experiment) using the custom-made high resolution PET for small animals, commercialized as Mirai-PET (Mirai Imaging Inc., Fukushima, Japan) (Sect. "Data processing for analysis"). After both experiments, the tumour tissue section was obtained as a reference of tumour histopathological status (Sect. "Kinetic analysis").

Beam-monitoring PET experiment with ¹²C ion beam irradiation

Experimental set up

A ¹²C ion beam was generated by the Heavy Ion Medical Accelerator in Chiba (HIMAC) in QST. The centre of the tumour of each rat was placed at the isocentre (IC) of the beam and the entire tumour tissue was irradiated in the HIMAC biological experimental irradiation room. The ¹²C ion beam conditions were as follows: maximum energy 290 MeV; spread-out Bragg peak (SOBP) width 6 cm; and beam size diameter 10 cm. The SOBP width and beam size were adjusted according to each tumour size by using polymethylmethacrylate (PMMA) absorbers and brass collimators (x, y) (Fig. 2a). The irradiated dose of the ¹²C ion beam was 8 Gy. (The dosimeter was calibrated at 22° and 1013.3 hPa. The monitor output one pulse when it had a charge of 1 nC.) The irradiation time was approximately 1.5 min.

Fig. 2.

The experimental setup of (a) ¹²C ion beam irradiation and the beam-monitoring PET scan with the TBS-PET system. For irradiation, the centre of the rat tumour was positioned at the isocentre, and beam size was adjusted to tumour size using collimators (x, y). (b) The hypoxia PET imaging experiment with the Mirai-PET

To detect a tiny amount of irradiation produced β⁺-emitters, the in-house high sensitivity TBS-PET system [24] was applied. This system, which can cover the whole body of a rat in order to observe whole-body radioactivity (β⁺-ray) distribution alterations (Fig. 2a), was developed using a 4-layer depth of interaction (DOI) detector to achieve 2.2 mm radial spatial resolution and 20.2% peak sensitivity at the centre of field of view (FOV) with an energy window of 400–600 keV. After irradiation, each rat was placed into the TBS-PET outside the irradiation room. Then, beam-monitoring PET scanning was performed for 30 min. The time from the end of irradiation to the start of PET scanning was within 2.7 min for the experiments with six rats (Rat-A to Rat-F). Each rat was anaesthetized with isoflurane (2.5% for introduction and 1.5%–2.5% for maintenance) during irradiation and beam-monitoring PET scanning. A reflector lamp was used to maintain the body temperature of the rat at 37 ± 1 °C, and a monitoring system for animal vital signs (BSM-3592 Life Scope; Nihon Kohden, Japan) was used to monitor this temperature.

One of the rats (Rat‑G) had died prior to the experimental study; however, it was irradiated under the same beam conditions and used as a reference sample to demonstrate the absence of biological washout effects for comparison.

Data processing for analysis

The list mode data acquired from the 30 min beam-monitoring PET scanning were divided into 15 s duration time frames for the first 10 min after the start of data acquisition and into 30 s duration time frames for the remaining 20 min. The algorithm of the ordered subset expectation maximization method (OSEM) was applied for image reconstruction with 5 iterations and 8 subsets for a matrix size of 155 × 155 × 327 and voxel size of 1.0 mm³. The attenuation correction was not applied because of the small attenuation for rat scans.

Kinetic analysis

According to the pixel value (intensity to maximum value in irradiated tumour region), the tumour region in the reconstructed PET image was segmented into three areas using PMOD software (PMOD Technologies LLC, Switzerland): intensity greater than 20% to 60%; intensity greater than 60% to 80%; and intensity greater than 80%. These areas were contoured as the volumes of interest (VOIs) on the static (i.e., 30 min summed) PET image. Then these VOIs were applied for each frame and the time activity curves (TACs) were obtained. The activity intensity in each VOI was analysed as a function of time. 

To derive the biological washout rate constant, the compartment model with two types of tracers was applied here (Fig. 3). The applied model is briefly explained as follows. The kinetics of the biological systems in the rat body was assumed to act as a single tissue compartment. Two components were assumed, the main component which corresponded to the faster washout component in the radiopharmaceutical distribution phase and the other component which corresponded to the slower washout component in the elimination phase. The effective decay function, , for ¹²C ion irradiation was expressed in Eq. (1) under the assumption of the two dominant radioisotopes ¹⁵O and ¹¹C, using the kinetic depicted in Fig. 3. Other contributions of other possible isotopes, such as ¹⁰C, ¹⁴O, ³⁰P, ³⁹K and ¹³N, were negligible for fitting analysis due to their small production rates and short half-lives.

1

Fig. 3.

A single-compartment model with two types of tracers. Irradiated ¹²C ions are changed into two types of tracers and diffused from the tumour tissue to blood vessel. k_2,m and k_2,s are the transition rate parameters of the main component and the other slower component, respectively

Here and correspond to the activity at the end of the beam irradiation. M_m and M_s are the fractions of the respective main component and the other component and they correspond to the main ( and the other (biological washout decay constants. Then, the fitting function, , could be described as Eq. (2).

2

The fitting parameters , , , , were used to fit the TACs. Also, the fitting approach implemented by Mizuno et al. (2003) [4]was applied in which the slower component was fit first, and the faster component was fit afterwards.              

To see the correspondence between the distribution of activity intensity on the summed PET image and that of the biological washout rate in the tumour, the biological washout rate constant was derived in a voxel-wise manner.

¹⁸F-FAZA PET experiment

The tumour retention of ¹⁸F-FAZA for the same rat model was observed. ¹⁸F-FAZA is a frequently used hypoxia PET tracer in clinical investigations for adapting radiation therapy [28–31]. It has been used to delineate hypoxic regions in tumours before radiation therapy, and also to make post-treatment assessments a few months after radiation therapy [13, 30]. A short explanation of the ¹⁸F-FAZA accumulation mechanism of is as follows. Since the imidazole aromatic ring (C₃H₄N₂) in the structural formula has a high electron affinity, when ¹⁸F-FAZA reaches the inside of a tumour in a hypoxic state, the nitro group (-NO₂) in the imidazole aromatic ring is converted by a reduction reaction into an amino group (-NH₂) that binds to cellular components and accumulates inside the cells.

In this experiment, ¹⁸F-FAZA was synthesized in cyclotron-based hot laboratory of QST. Within a half day (about 10 h) after the beam-monitoring PET experiment, each rat was transferred to the animal experiment room where it is possible to handle unsealed radiation sources. Then, about 20 MBq of ¹⁸F-FAZA was intravenously administrated into each rat one-by-one, with an appropriate time-shift for the PET scan requiring 30 min for each. Hypoxia ¹⁸F-FAZA PET imaging was done by the Mirai-PET [32] which is equipped with a crosshair light sharing (CLS) detector which can measure both time-of-flight (TOF) and DOI information [33]. Also, the Mirai-PET has the unique feature of being able to customize the detector ring size to arrange a maximum of 128 detectors. The model which was applied in this study was designed to have a FOV of 16 cm in diameter and 15.4 cm in axial length. The scintillator pitch was 1.6 mm, and the spatial resolution was less than 1 mm. Figure 2b shows the setup of the hypoxia ¹⁸F-FAZA PET imaging experiment. ¹⁸F-FAZA was administered, and the rat was kept at rest for 60 min to ensure the ¹⁸F-FAZA circulated throughout the entire body. Next PET scanning was done for 30 min. A thermometer for continuous measurement of the core temperature of small animals (BWT-100 A, Bio Research Centre, Nagoya, Japan) was placed under the stomach. During the experiment, including administration, rest time, and PET imaging time, the rat was anesthetized with isoflurane (induction, 2.5%; maintenance, 1.5%). After hypoxia PET imaging, the OSEM image reconstruction method was performed with 8 iterations and 8 subsets for a matrix size of 200 × 200 × 312 and voxel size of 0.5 mm³.

Tumour tissue section

After both the beam-monitoring PET and the ¹⁸F-FAZA PET experiments, tumour of six rats (Rat-A to Rat-F) were removed and tumour tissue sections were prepared for each. A slice of the tumour tissue section was taken from each tumour centre and stained with Haematoxylin & Eosin (H&E). The H&E-stained tumour section was scanned by a digital scanner (NanoZoomer S360 Digital Slide Scanner, Hamamatsu Photonics, Japan), and pathological conditions of each tumour were observed by image viewing software (NDP.view2, Hamamatsu Photonics).

Results

Beam-monitoring PET experiment

Reconstructed PET images

The irradiated ¹²C ion beam which focused on the whole tumour was visualized in the six rat models (Rat-A to Rat-F). As an example, Fig. 4a shows the summed image (30 min) of the beam-monitoring PET experiment for Rat-A as a maximum intensity projection (MIP). The arrow in the Fig. 4a shows the irradiated direction. In the summed-MIP image, the irradiation produced β⁺-emitters that diffused throughout the whole body due to the biological washout effect were visualized, and the heterogeneous activity intensity distribution within the tumour was successfully observed. Supplemental Fig. 1 shows the VOIs when the tumour was segmented into three regions by pixel colour scale values on the enlarged tumour images (for Rat-A again). Supplemental Fig. 1a, b, and c show axial, sagittal, and coronal slices (not MIP images), respectively. Figure 5a show an enlarged axial slice image of the centre of the tumour for Rat-A; those for each rat (Rat-A to Rat-F) are presented in Supplemental Fig. 2a-1 to f-1.

Fig. 4.

Reconstructed summed (30 min) PET images for Rat-A. (a) The beam-monitoring PET image (MIP) acquired by TBS-PET after ¹²C irradiation. The arrow shows the beam irradiated direction. (b) The hypoxia ¹⁸F-FAZA PET image acquired by Mirai-PET

Fig. 5.

Heterogenous structure of the tumour for Rat-A. (a) Tumour image (30-min summation) and (b) the k_2,m-map, obtained from the beam-monitoring PET experiment. (c) Tumour image obtained with the hypoxia ¹⁸F-FAZA PET experiment. These are enlarged to the same scale. (d) Estimated tumour structure based on histological image shown in Supplemental Fig. 3. In the peripheral region of the tumour (corresponding to Supplemental Fig. 3 d), tumour is still growing, and washout rate was rapid due to vascular hyper-perfusion which characterizes tumour angiogenesis. Inner side, (corresponding to Supplemental Fig. 3c), a hypoxic region was observed in the vicinity of ¹⁸F-FAZA accumulation. In the core (corresponding to Supplemental Fig. 3b), a necrotic area with no blood flow was present with slight spillover from the surrounding slow washout region. This interpretation is further supported by pathological images shown in Supplemental Fig. 3

The biological washout rate

Figure 6a to f show the measured TACs of activity intensity in each VOI for Rat-A to Rat-F. The horizontal axes show the frame mid time (s). The vertical axes show relative intensity of each VOI in arbitrary units (A.U.), normalized by the first mid time point after the PET scan started. The TACs of the intensity segmented regions (VOIs shown in Supplemental Fig. 1) are plotted with solid circles of blue, red and black for intensities greater than 20% to 60%, greater than 60% to 80%, and greater than 80%, respectively. The TACs were separated into three groups, and the biological washout rate was apparently slower in the highest intensity area (> 80%) than that of the lowest intensity area (> 20% to 60%). All TACs were well fitted to the single tissue compartment model. The derived biological washout constants are summarized in Table 1. In this study, the biological washout rate constant of the main component, , was quantitatively determined, while that of the slower component, , was negligibly slow. Figure 5b show the -maps which were calculated voxel-by-voxel for Rat-A; those for each rat (Rat-A to Rat-F) are presented in Supplemental Fig. 2a-2 to f-2. It should be noted that Fig. 5b displays the calculated -values, whereas Fig. 5a shows the PET image summed over 30 min scan. Although spatial resolutions of these images are different each other, these are displayed in the same scale.

Fig. 6.

Time activity decay curves (TACs) of each VOI. The TAC of VOIs with intensities above 80%, greater than 60% to 80%, and greater than 20% to 60% of the maximum intensity are plotted as black, red, and blue filled circles, respectively. (a) to (f) correspond to Rat-A to Rat-F. Solid lines were obtained by fitting results with the single-compartment model. The grey dotted line in each graph represents the TAC when the tumour of dead reference rat, Rat-G, was irradiated with the ¹²C ion beam

Table 1.

The biological washout rate constants (k_2,m, k_2,s) of the irradiation produced β⁺-emitters in the high activity intensity and the other (lower activity intensity) areas

	Washout rate constants (min− 1)
	Int. > 80%		Int. >60% to 80%		Int. >20% to 60%
Rat-A	0.017	< 1.0E-05	0.074	< 1.0E-05	0.13	< 1.0E-05
Rat-B	0.029	< 1.0E-05	0.053	1.1E-03	0.22	< 1.0E-05
Rat-C	0.027	1.9E-04	0.070	1.2E-03	0.24	< 1.0E-05
Rat-D	0.043	< 1.0E-05	0.063	5.1E-04	0.26	< 1.0E-05
Rat-E	0.041	< 1.0E-05	0.059	< 1.0E-04	0.23	< 2.0E-05
Rat-F	0.036	2.1E-04	0.062	2.3E-04	0.23	< 1.0E-05
Rat-G(dead)	< 1.0E-10

¹⁸F-FAZA PET experiment

Figure 4b shows the summed (30 min) PET image from ¹⁸F-FAZA PET experiment for Rat-A. Intravenously administered ¹⁸F-FAZA was accumulated in the tumour site. (Accumulation in the bladder and kidney was also observed since ¹⁸F-FAZA has low lipid solubility and it is rapidly excreted into the urinary system.) The ¹⁸F-FAZA PET experiment also showed a heterogeneous activity intensity distribution within the tumour in each rat. Figure 5c show enlarged axial slice images of the centre of each tumour for Rat-A; those for each rat (Rat-A to Rat-F) are presented in Supplemental Fig. 2a-3 to f-3. It should be noted that Fig. 5c shows the static hypoxia PET image obtained with intravenous administration of ¹⁸F-FAZA and scanned by Mirai-PET [32], whereas Fig. 5a shows the beam-monitoring PET image acquired with TBS-PET [24]. Although these two images are enlarged to the same scale, their resolutions differ.

The summed PET image from the beam-monitoring PET experiment and that from ¹⁸F-FAZA PET experiment were fused by extracting body contours automatically using PMOD software. The intersection of the horizontal and vertical solid lines in Fig. 5 indicates approximately the same position on each image for Rat-A to Rat-F. The highest activity intensity area which was observed in the image of the beam-monitoring PET experiment roughly corresponded to the low accumulated area which was observed in the image of the ¹⁸F-FAZA PET experiment.

Figure 7 shows the relationship between the ¹⁸F-FAZA accumulation and value of the three segmented regions depending on intensity in the summed beam-monitoring PET image (for Rat-A to Rat-F). The regions of intensity > 80% (black), intensity > 60% to 80% (red), and intensity > 20% to 60% (blue) were clearly distinguished.

Fig. 7.

Plot of the biological washout constant (k_2,m) versus accumulation of ¹⁸F-FAZA in the tumours for Rat-A to Rat-F. Three regions with intensity > 80%, intensity > 60% to 80%, and intensity > 20% to 60% (of the maximum intensity in the summed beam-monitoring PET image) are plotted as black, red, and blue solid circles, respectively

Tumour histopathology

The slice though the centre of the tumour for Rat-A of the H&E-stained tumour tissue section was shown in Supplemental Fig. 2a. Heterogenous morphology was observed. Enlarged images of H&E-stained tumour tissue at point-b, point-c, and point-d (identified in Supplemental Fig. 2a) were shown in Supplemental Fig. 2b to d. At point-b, a dead cell-like distribution without any blood vessels was observed, indicating necrosis had occurred in this region. At point-c, a high density of tumour cells and only a few blood vessels suggesting poor blood flow were observed, indicating a hypoxic state. At point-d, abundant vascular structures (blood vessels which were filled with red blood cells) were observed, indicating the presence of a region of viable cells.

Discussion

In this work, the biological washout rate distribution in the hypoxic tumour models was obtained by beam-monitoring PET after ¹²C ion beam irradiation. In summed PET image of each rat, activity intensity varies from the tumour centre to the periphery (Fig. 5a). And as Fig. 6; Table 1 showed, the higher the activity intensity in the summed PET image, the slower the biological washout rate.

It is known that gliomas induce a distinct vasculature that differs systematically from the tumour centre to the periphery. The glioma vasculature was described by Deane and Lantos (1981), and Vajloczy et al. (1998) [25, 26] reported similar findings for gliomas growing subcutaneously in a rat. As the tumour grew, the core developed avascular necrosis, while there were capillary buds corresponding to angiogenesis at the tumour periphery. In larger tumours (> 4 mm in diameter), there was a hypoxic area in between the central avascular necrotic region and the peripheral angiogenetic area [25].

In this study, Fig. 5 illustrated the typical solid glioma tumours. In the core, the irradiation produced β⁺-emitters were accumulated (Fig. 5a) with a slow biological washout rate (Fig. 5b), and ¹⁸F-FAZA might not have been able to penetrate (Fig. 5c) due to avascular necrosis. Surrounding the core of the tumour, there was a hypoxic area as indicated by ¹⁸F-FAZA accumulation (Fig. 5c). In the periphery, tumour angiogenesis was observed (Supplemental Fig. 3-d). Newly blood vessels associated with angiogenesis the biological washout rate was faster (Fig. 5b) due to vascular hyper-perfusion corresponding to angiogenesis [8]. This interpretation was further supported by the H&E-stained tumour section as shown in Supplemental Fig. 3, as well as by the good correlation between the value and ¹⁸F-FAZA accumulation demonstrated in Fig. 7.

Based on these results, it was reasonable to interpretate the areas with intensity > 80% (black), intensity > 60% to 80% (red), and intensity > 20% to 60% (blue) as corresponding to necrosis, hypoxia, and angiogenesis regions, respectively (Supplemental Fig. 2).

Figure 7 showed that the biological washout rate could distinguish between the viable, hypoxic, and necrotic areas more clearly than the summed (30 min) ¹⁸F-FAZA PET image which could not distinguish the viable and necrotic area clearly. As a follow-up experiment, the authors will conduct dynamic ¹⁸F-FAZA PET scans to confirm whether the low accumulated area was a necrotic area where PET tracer could not penetrate due to a lack of blood vessels, or it was an area where PET tracer did not accumulate after penetration.

The relevance of tumour vascular status determinants for radiation therapy outcome and prognosis in patients was described more than one hundred years ago by Schwartz (1909) [34], and later in numerous clinical reports [35–42]. The locations of hypoxia and necrosis are known to move quickly due to changes in tumour vascular status with tumour growth and response to radiation therapy [13, 18]. Those locations can be detected in beam-monitoring PET, which will help to delineate an intensive irradiation target [15]. This study demonstrated the use of beam-monitoring PET for on-site evaluation of tumour vascular status to provide additional information for adaptive therapy and early assessment of progression.

One of the limitations in this study was the accuracy of alignment of different images from the beam-monitoring PET experiment, the ¹⁸F-FAZA PET experiment, and the H&E-stained tumour section observation. Due to regulatory issues in Japan regarding unsealed radioactivity handling areas, the beam-monitoring PET and the ¹⁸F-FAZA PET were performed in separate rooms. After the beam-monitoring PET experiment using the TBS-PET system in HIMAC, the rats were moved to another building and placed in the Mirai-PET system for the ¹⁸F-FAZA PET experiment. Also, the tumour tissue section might have been distorted during excision, fixation in alcohol, and freezing for slicing. Furthermore, the specific hot chemistry of the isotopes formed in a living body is still unknown. It should be noted that the irradiation produced β⁺-emitting nuclides may form a variety of molecules in living tissue, such as ¹¹CO, ¹¹CO₂ and H₂¹⁵O [41, 42, 44], and that half-lives of these are different from that of the ¹⁸F-FAZA (C₈H₁₀FN₃O₅), hence their kinetics are clearly different. As the next step, beam-monitoring PET imaging and hypoxia PET imaging with precisely aligned coordinates should be carried out by developing an experimental plat form for small animal experiments.

In this study, the single tissue compartment model with two washout rate components was applied to derive the biological washout rate constants. And the obtained TACs fit with the model well. However, only the main component, , was derived; the other slower component, , was too slow to be determined quantitatively (Table 1) from the acquired data duration of 30 min. The other slower component might be caused by trapping of the nuclides by stable molecules and/or formation of high molecular weight which cannot easily penetrate the vessel wall. When the ¹²C ion beam is irradiated to the tissue with therapeutic energy, the main produced β⁺-emitting nuclide is ¹¹C from a projectile reaction [43]. Although no direct identification of the chemical forms of the produced β⁺-emitters in living tissue has been made, ¹¹C ions may form a variety of molecules, such as ¹¹CO and ¹¹CO₂ [44–47] which can pass through blood vessel walls and wash out. The ¹¹C ion beam irradiation study in rat brain depicted the blood profile like an image of the whole rabbit body [46]. This study suggested that ¹¹C ions formed ¹¹CO which strongly bonded with haemoglobin in the blood vessel. Radioisotope beams, such as ¹¹C and ¹⁵O ion enable more accurate analysis of the biological washout effect than conventional stable beam irradiations do [48]. Further investigation is necessary to identify the chemical forms of the ¹¹C-labeled molecules and to get accurate modelling of the biological washout effect.

Conclusion

The biological washout rate of the irradiation-produced β⁺-emitting nuclides reflects distinct tumour vascular status, necrotic, hypoxic and viable. Clear evidence supporting the predicting tumour hypoxia status from the biological washout in in-beam PET monitoring was shown in tumour rat models irradiated with carbon-ion beam, by comparison with ¹⁸F-FAZA PET.

Abbreviations

PET

Positron emission tomography

¹⁸F-FAZA

¹⁸F-labeled fluoroazomycin arabinoside

TBS-PET

Total body small-animal PET

H&E

Haematoxylin & Eosin

MIP

Maximum intensity projection

TAC

Time activity curve

DOI

Depth of interaction

Author contributions

All authors contributed to the study conception and design. CT analysed and interpreted the dynamic data regarding the tumour rat model. CS provided guidance for compartment model analysis. HT performed the data processing and reconstruction of the beam-monitoring PET experiment. HW was responsible for anaesthesia, administration of radiopharmaceutical and management of animal. HK developed high sensitive PET which was used for the beam-monitoring experiment. TY designed the PET devices and supervised its development. GA operated the detector and device of high resolution PET for hypoxia PET experiment. TI performed data processing and reconstruction of the hypoxia PET experiment. CS. HS was responsible for culturing the cancer cells, preparing the tumour tissue sections, and performing the pathological diagnosis. IK provided guidance on pathological diagnosis. The first draft of the manuscript was written by CT and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.

Funding

This work was supported by JSPS KAKENHI Grant Numbers 24K10826 and 23KK0206. This work was partially supported by the QST President Grants and Nakatani Foundation, Japan.

Data availability

The datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

Animal experiments were approved by the local ethical committee of specialists in the National Institutes for Quantum Science and Technology (QST) and were carried out in compliance with the institute’s animal experimentation guidelines.

Consent for publication

Not applicable.

Competing Interests

The authors have no relevant financial or non-financial interests to disclose.