(Keynote) From small (and not so small) animals to humans: Lessons learned from translational studies using photon and proton radiation beams at Penn
ABSTRACT. Radiotherapy (RT) is a major anti-cancer modality with approximately half of all patients receiving RT at some point during the course of their disease. Despite the dramatic improvements in conformality achieved by several radiation modalities (e.g., IMRT, SBRT, Proton and heavy ion beams, etc) and the development of combined modalities (e.g., radio-immunotherapy), normal tissue injury still remains a common side effect in many cancer patients, which constraints the efficacy of radiotherapy. Moreover, as survivorship rates increase, the quality of life of many patients who receive radiotherapy can be negatively impacted due to normal tissue toxicities. Over the past couple of decades, technologies to deliver radiation in experimental preclinical models such as mice and rats have been developed and advanced to the point where we can model complex treatment plans used in patients and test novel delivery, and combined modalities.
In my presentation, I will summarize data that our group at the University of Pennsylvania has accumulated since 2010, when we installed our first small animal irradiator. I will describe the infrastructure and "know-how" that we leveraged in order to perform cutting-edge studies in this area. I will outline work to develop and characterize more physiologically relevant mouse models of normal tissue injury with a focus on intestinal, cardiac, skin and head & neck tissues. Select cases which have culminated in findings being translated into novel clinical trials will also be presented. During the last decade, exciting new developments in the area of ultra-high dose rate (“FLASH”) radiotherapy have been made and this modality has reached the clinic. I will present some of our work in mouse models of intestinal damage, skin and salivary glands and trials in canine patients in partnership with the Penn Vet School in preparation of human clinical trials with FLASH we expect to open this year.
I will conclude by highlighting some collaborations with other groups Nationally and Internationally and by summarizing the key lessons we learned along the way.
ABSTRACT. Spatially fractionated radiation therapy (SFRT) challenges some of the classical dogmas in conventional radiotherapy by using a highly spatial dose modulation [1]. A significant increase in normal tissue dose tolerances was observed both in early clinical trials and in small animal experiments [1]. Tumour control effectiveness is maintained or even enhanced in some configurations as compared with conventional radiotherapy [1]. Among the different forms of SFRT, minibeam radiation therapy (MBRT) [2] uses very narrow (submillimetric) beams. This enables the alliance between the advantages of SFRT with the exploit of the dose volume effects. The alliance of charged particles and MBRT may allow to fully profit from the advantages of the spatial fractionation of the dose, while paving a direct avenue for optimal future patients’ treatments. Different approaches have been explored in the recent years. pMBRT represents the most evolved experimental form of SFRT with charged particles [3]. It allies the more localized dose distribution in the Bragg peak of protons, with the possibility of achieving simultaneously a homogeneous dose distribution in the tumor and a pronounced spatial fractionation of the dose in normal tissues. A series of radiobiological studies have already demonstrated the remarkable normal tissue reduction provided by pMBRT as compared to standard proton therapy [4-6]. In addition, the tumour control effectiveness of pMBRT equals or is better than that of standard RT in glioma-bearing rats, avoiding the extensive adverse effects observed in the latter and thus widening the therapeutic window [7].
The full picture on the biological mechanisms is still missing. Main potential players with some experimental evidence reported in the literature include ddifferential vascular effects, cell signaling effects (bystander-like effects) and abscopal effects, inflammation and immunomodulatory effects and stem cell migration. Currently, the relative weight and interrelation of those possible contributors and how these effects are translated when the spatial dose distribution is modified are still unclear. In this lecture a general overview about MBRT and its biological effects will be provided.
1. Prezado, Y., Divide and conquer: spatially fractionated radiation therapy. Expert Reviews in Molecular Medicine, 2022. 24: p. 12.
2. Dilmanian, F.A., et al., Interlaced x-ray microplanar beams: a radiosurgery approach with clinical potential. Proc Natl Acad Sci U S A, 2006. 103(25):
3. Prezado, Y. and G.R. Fois, Proton-minibeam radiation therapy: a proof of concept. Med Phys, 2013. 40(3): p. 031712.
4. Girst, S., et al., Proton Minibeam Radiation Therapy Reduces Side Effects in an In Vivo Mouse Ear Model. Int J Radiat Oncol Biol Phys, 2016. 95(1):
5. Prezado, Y., et al., Proton minibeam radiation therapy spares normal rat brain: Long-Term Clinical, Radiological and Histopathological Analysis. Sci Rep, 2017. 7(1): p. 14403.
6. Lamirault, C., et al., Short and long-term evaluation of the impact of proton minibeam radiation therapy on motor, emotional and cognitive functions. Sci Rep, 2020. 10(1): p. 13511.
7. Prezado, Y., et al., Tumor Control in RG2 Glioma-Bearing Rats: A Comparison Between Proton Minibeam Therapy and Standard Proton Therapy. Int J Radiat Oncol Biol Phys, 2019. 104(2): p. 266-271.
Long-term skin side effects of FLASH proton radiotherapy compared to conventional radiotherapy
ABSTRACT. Introduction:
FLASH proton radiotherapy delivers high-energy protons at ultra-high dose rates (>40 Gy/s) achieving local tumor control while minimizing the damages to surrounding healthy tissues. The FLASH effect has been observed across various tissue types in diverse animal models, effectively preserving normal tissue. While earlier experiments primarily studied short-term skin reactions, there is considerable interest in exploring its potential to reduce long-term skin toxicity [1].
Materials & Methods:
The irradiation setup was established at the experimental beamline of the University Proton Therapy Dresden. This setup enabled the irradiation of partial brain volumes in mice with a proton spread-out Bragg peak at ultra-high dose rate [2]. A pilot study with 14 C57BL/6 mice was conducted. Six mice were exposed to an ultra-high dose rate (240 Gy/s) of irradiation, while another six were subjected to a conventional dose rate (20 Gy/min), and two served as unirradiated controls. After a six months follow-up period, the facial skin was harvested and embedded in paraffin. Different histochemical stainings were used to analyse different aspects of the tissues, e.g. collagen deposition (Herovici and Masson`s trichrome) and standard morphology (H&E).
Results:
A preliminary analysis of skin reactions (hair loss and erythema) already demonstrates a clear tissue sparing effect in the FLASH group. Qualitative and quantitative analyses will be conducted on the stained samples using Fiji ImageJ to further prove the differences in skin reactions between FLASH and conventional irradiation.
Conclusions:
The irradiation setup and the workflow will be presented together with the first biological data obtained from the skin samples. While the analyses of the samples are still in progress, the initial data already indicates a distinct advantage in preventing skin reaction with the use of FLASH irradiation. The study will continue, including immunohistochemical analysis of the brains harvested together with the facial skin samples.
References:
[1] Limoli, C. L., & Vozenin, M. (2023). Reinventing radiobiology in the light of FLASH radiotherapy. Annual Review of Cancer Biology, 7(1), 1-21. https://doi.org/10.1146/annurev-cancerbio-061421-022217
[2] Horst, F., Beyreuther E., Bodenstein, E., Gantz, S., Misseroni, D., Pugno, N., Schuy, C., Scifoni, E., Weber, U., & Pawelke, J. (2023). Passive SOBP generation from a static proton pencil beam using 3D/printed range modulators for FLASH experiments. Frontiers in Physics, 11. https://doi.org/10.3389/fphy.2023.1213779
Investigating the tumor control dose (TCD50) of microbeam radiation therapy (MRT) compared to conventional radiation therapy (CRT) of human tumor xenografts in nude mice
ABSTRACT. Introduction:
Microbeam radiation therapy (MRT) is an innovative preclinical concept in radiotherapy that collimates X-ray radiation in micrometer-wide, planar beams. Previous research has shown that MRT substantially spares normal tissue, while being equally effective in tumor ablation. The aim of this study was to determine the tumor control probability in an in vivo mouse xenograft model comparing MRT with conventional radiotherapy (CRT). The tumor control probability is a parameter that is used to describe the efficacy in tumor control of a certain radiation treatment.
Methods:
Human A549 lung cancer cell line was implanted subcutaneously in CD1-Foxn1nu mice. The mice cohort was randomly partitioned into two treatment groups: a uniform dose distribution (CRT) group with five dose subgroups and a MRT group with five dose subgroups. Each group consisted of five mice. All irradiations were performed at the small animal radiation research platform´s SARRP and XenX (XSTRAHL LTD, United Kingdom). The equivalent uniform dose (EUD) concept was used to calculate radiation doses for MRT. The spatially fractionated fields were created by using a specially designed collimator. Tumors were irradiated by 12-13 microbeams of 100 µm in width, with a CTC (center to center distance) of about 430 µm and a PVDR of 20. MRT was carried out with a valley dose rate of 0.185 ± 0.0005 Gy/min and an average peak dose rate of 4.6 ± 0.1 Gy/min. CRT irradiations were carried out with a dose rate of 4.6 ± 0.1 Gy/min. Tumor volumes after treatment were constantly measured with a caliper for a total follow-up time of 120 days. The volumes obtained were analyzed to determine tumor control or regrowth. The percentages of tumor control among all the dose groups for MRT and CRT were fitted with a logistic regression and the corresponding TCD50 values were obtained.
Results
Preliminary data demonstrated TCD50 value of 17.46 (15.68 – 19.31) Gy for MRT while the TCD50 value for CRT was significantly higher (p< 0.001) with 31.32 (27.24 – 32.75) Gy. Therefore, MRT is capable of controlling 50% of tumors while delivering a 44.2% reduced dose compared to CRT (p< 0.001). In addition, MRT is able to control 90% of the tumors by increasing the TCD50 dose by only 4.82 Gy while CRT is able to control 90% of the tumors by increasing the TCD50 dose by 9.28 Gy.
Conclusion
The results clearly demonstrate that MRT leads to a greater probability of tumor control by drastically reducing the delivered dose compared to CRT. This decrease in dose broadens the therapeutic window of MRT and therefore reduces normal tissue side effects. These findings clearly show the advantages of spatially fractionated radiotherapy.
FLASH and SFRT: Revisiting Spatial and Temporal factors in RT in the laboratory setting.
ABSTRACT. SFRT and FLASH-RT highlight spatial and temporal factors with potentials to impact therapeutic ratio in radiation treatment. The scope of SFRT is broad, comprising of 50 microns-range microbeam (MRT) and the mm-range minibeam (MBRT) in the laboratory setting to the cm-range GRID, or Lattice in clinical use. The mechanistic convergence of their spatial parameters, however, is unclear. For FLASH RT, the biological underpinning and the tolerance in dose and dose rate windows for tumors and normal tissues have also yet to be resolved. It is noteworthy that FLASH and SFRT are intrinsically intertwined in radiation treatment. Clinical translation of FLASH-RT would inherently involve spatial considerations such as the use of multiple-beam arrangement, or layer-by-layer proton pencil beam scanning (PBS). The combination of FLASH-SFRT in conjunction with other transformative modalities, such as immunotherapy, have also been suggested. However, the current reliance on advanced, and less accessible, irradiation systems presents a barrier for pre-clinical laboratory research for both FLASH and SFRT in the community. These obstacles compel our development of a new x-ray FLASH-SARRP system with novel pencil beam scanning capability by robotic positioning of the study animal for irradiation. A robotic motion stage capable of 7 cm/s is positioned in the 3 cm working space between the parallel-opposed 150 kVp fluoroscopy x-ray sources to translate a 2 cm study animal (mouse) across a stationary slit- or pencil-beam. Early results indicate that the step-and-shoot approach can deliver > 0.5 mm slit and > 1 mm diameter aperture beams at FLASH and conventional dose rates with high peak-to-valley dose ratios (PVDR) for SFRT investigations. Our investigation also shed light on the inadequacy of ionizatation-based dosimetry, Gy, to differentiate spatial (e.g LET) and temporal (e.g. FLASH) factors on radiation damage. The recognition led to our parallel investigation of molecular dosimetry. Plasmid DNA in water is chosen as a model molecule where strand breaks and lesions are assessed. The inanimate environment avoids implicating complex biological response. Early results indicate reduced damage at high 150 kVp x-ray fluence rate (> 90 Gy/s) and high dose (> 30 Gy), but not in the pristine bragg peak region of proton beams with increased LET. These new explorations into the spatial and temporal factors lead to revisiting the dogmas of uniformity in dose, fluence and target volume and perhaps new research frontiers to transform radiation treatment.
ABSTRACT. Introduction:
With the introduction of small animal image-guided radiotherapy platforms it became possible to irradiate orthotopic tumors in mice in a highly conformal way without co-irradiation of relevant normal tissue units. Nowadays these platforms are also used to irradiate tumors with spatially and temporally unconventional single high-dose and fractionated treatment regimens to identify new biological connections between the tumor(s) and normal tissue organs. Here we are investigating new radiotherapy regimens at the preclinical level in order to identify applicable solutions for challenging clinical situations.
Materials and Methods:
Murine models carrying multiple tumor lesions or with metastatic disease were irradiated with single high dose and fractionated radiotherapy alone and in combination with different immune checkpoint inhibitors (ICI). Draining lymph nodes were either spared or co-irradiated in a neoadjuvant, concomitant or adjuvant way thereby creating multiple spatially and temporally distinct treatment regimens.
Results:
The propensity of radiotherapy to act as an in situ tumor vaccine resulted in the development of radiotherapy-ICI combinations in an attempt to overcome the treatment resistance. However, common practice of tumor draining lymph node irradiation (DLN IR) might be the culprit for successful treatment outcome. Here we demonstrate that delayed (adjuvant), but not neoadjuvant DLN IR overcomes the detrimental effect of concomitant DLN IR on the efficacy of combined radioimmunotherapy. Furthermore, we identify for the first time IR-induced disruption of the CCR7-CCL19/CCL21 homing axis as a potential key mechanism behind the deleterious effect of DLN IR and recognize a specific cellular DLN subpopulation as the major deregulated source of these cytokines in response to irradiation.
Conclusion:
Our study demonstrated novel spatially and temporally distinct tumor- and DLN-oriented radiotherapy regimens to maximize the efficacy of radioimmunotherapy across different tumor types and disease stages.
ABSTRACT. Radiotherapy is a commonly used treatment for many types of cancers, administrated either in conjunction with other therapies or as a standalone approach. Approximately 50% of cancer patients will undergo radiation therapy. Given the potentially harmful effects of radiation on both cancerous and healthy tissue, treatment is typically fractionated to minimize damage to healthy tissue. Some research areas explore the use of nanomaterials as radiosensitizers. In this context, a nanomaterial accumulated in the tumour “sensitizes” it, leading to increased tumour ablation compared to radiation alone. Conventionally, heavy metal-based nanomaterials have been considered ideal due to their high atomic numbers corresponding to higher X-ray attenuation coefficients. However, this study delves into the therapeutic potential of the organic porphysome nanoparticle for radiodynamic therapy, the mechanism of which remains largely unknown. Although some studies have suggested Cherenkov as a potential mechanism, it prevails at higher energies (MeV) and further investigation may reveal otherwise.
Previous literature has reported the use of the clinically approved photosensitizer Photofrin as a radiosensitizer and porphyrins in general for radiosensitization [1],[2]. One such study examined 15MeV photons and 5-ALA for treating lung cancer [3]. The porphysome is an organic lipid nanovesicle assembled from porphyrin -containing building blocks, featuring approximately ~80,000 photosensitizers per particle [4]. We have previously demonstrated its efficacy as a potent photodynamic therapeutic agent [5]. Currently, we are exploring the porphysome’s potential as a radiosensitizer, having observed a therapeutic effect in vivo.
For the animal model, C57BL/6J mice are injected with 5 × 10^5 MCA 205 cells to induce a syngeneic murine tumour on the hind flank. Around 7-9 days post-injection, the mice are then injected with porphysome. 24 hours later, the mice are irradiated at 16Gy using a clinical LINAC machine. Subsequently, the mice are monitored by measuring weight and tumour size with calipers three times a week and ultrasound once a week until the endpoint.
Preliminary results from the first two rounds of this study have shown a noticeable effect on the survival of the animals for the porphysome treated group. We have observed that intravenous (IV) injection shows more efficacy than intratumor injection, prompting us to continue with the IV method. Our next step involves investigating the potential role of Cherenkov as a mechanism by irradiating a group of mice at a 225keV energy to assess whether the therapeutic effects are diminished. Furthermore, we aim to explore the x-ray absorption characteristics of the porphysome to deepen the understanding of the functionality of this nanoparticle. Elucidating the mechanisms behind this radiosensitization will allow for more careful selection of agents for therapeutic work.
References
[1] M. Schaffer et al., Current Medicinal Chemistry. 12, 1209–1215 (2005).
[2] Z. Luksiene, Medicina (Kaunas, Lithuania). 40, 868–874 (2004).
[3] D.-M. Yang, et al., Biomedical Physics & Engineering Express. 8, 065031 (2022).
[4] J. F. Lovell et al., Nature Materials. 10, 324–332 (2011).
[5] K. Guidolin, L. Ding, J. Chen, B. C. Wilson, G. Zheng, Nanophotonics. 10, 3161–3168 (2021).
Developing a preclinical mouse model to study image-guided stereotactic ablative radiotherapy and systemic treatment in Liver Cancer
ABSTRACT. Introduction: Optimal treatment for Hepatocellular Carcinoma (HCC) is curative, either resection, liver transplantation or local ablation, however, only ~20% of patients are eligible for this. Many patients present with incurable disease, for whom transarterial chemoembolisation or systemic chemotherapy are palliative options. Unfortunately, both have low responses and high rates of recurrence and progression. Stereotactic Ablative Radiotherapy (SABR) is a newly approved treatment option for a subset of HCC patients who are not eligible for resection or other local treatments. Retrospective studies report that this treatment modality has both clinical efficacy and feasibility, but prospective trials are required to better understand the role of SABR in HCC. SABR has the potential for good tumour control, but with recurrent and disseminated disease, there is a significant opportunity for integrating SABR into multimodal combination therapy. With this, there is a significant research opportunity in the preclinical space for models of SABR in HCC. Our aim is to develop a clinically relevant model to study targeted radiotherapy and systemic therapies in the treatment of HCC.
Methods: We have optimised an orthotopic transplant model injecting a mouse-derived HCC cell line into the immunocompetent murine liver, with contrast-enhanced CT guided SABR therapy using 20Gy single fraction in combination with anti-PD1 immunotherapy. Treatment response and immune phenotyping was assessed using immunohistochemistry, flow cytometry and liver biochemistry.
Results: We have established a syngeneic orthotopic transplant model whereby transplantation of a murine-derived HCC cell line into the murine liver gives rise to an anatomically accurate mono-focal liver tumour in an immunocompetent setting. In keeping with clinical management and treatment planning, we have successfully implemented intravenous contrast enhanced imaging in our model to reliably detect and delineate liver tumours, enabling long-term longitudinal CT imaging follow-up. We demonstrate that 20Gy single fraction arc irradiation, induces significant DNA-damage in tumours compared to non-irradiated tumour controls, with minimal off-target dosing to the liver and surrounding organs at risk. In addition, we report reduced proliferation (BrdU) and increased senescence (p21) in irradiated tumours. Tumour infiltration of CD8+ cells occurs 7days after radiotherapy. Finally, we report that 2-week combination SABR and immunotherapy (aPD1) alters the ratio of FoxP3+ T regulatory cells in the tumour.
Conclusion: Our orthotopic transplant model is an exemplary platform to study stereotactic ablative radiotherapy in the treatment of HCC. Furthermore, it provides a unique preclinical opportunity to evaluate the combination of SABR with systemic therapies, including immunotherapy, which could identify treatment strategies that could improve patient response rates to SABR in the clinic.
Targeting VEGF-A in an immunocompetent orthotopic mouse model of mesenchymal glioblastoma improves anti-tumorigenicity and decreases proinflammatory response in normal brain tissue after fractionated radiotherapy
ABSTRACT. Background:
Glioblastoma is the most malignant primary brain tumor characterized by dismal prognosis and profound therapy resistance, which is most evident for tumors of the mesenchymal molecular subtype. VEGF-targeting therapy implementing the humanized antibody bevacizumab, despite failing to improve survival in randomized trials, yielded relevant clinical benefits in subgroups of patients demanding for further research on treatment regimens integrating VEGF-targeting therapy into standard therapy.
Materials & Methods:
A mesenchymal-subtype resembling glioblastoma mouse model was generated by intracranial implantation of GL261 cells into C57BL/6 mice and comparing microarray-generated transcriptome data of explanted tumors with RNA-seq-derived transcriptome data from human normal brain and glioblastoma tissues obtained from The Cancer Genome Atlas (TCGA) database. Mice were treated with computed tomography (CT)-based fractionated radiotherapy (2x 5x 2 Gy) and concurrent administration of the murine VEGF-A-specific antibody G6-31 (3x 0.1 mg). Tumor progression was monitored by contrast-enhanced CT scanning. Transcriptomic profiles were assigned to the molecular subtypes of glioblastoma as described by Verhaak et al. (2010) and subjected to Kyoto encyclopedia of genes and genomes (KEGG), gene set enrichment (GSEA), and leading edge (LEA) analyses. For monitoring of immune cell infiltration of tumors and normal brain tissue in response to radiation, human immune cell-related genes were translated to homologous mouse genes and compared to curated gene sets representing different myeloid cell populations.
Results:
Co-administration of G6-31 significantly increased the anti-tumorigenic potential of fractionated CT-based radiotherapy in this preclinical model of mesenchymal glioblastoma elongating the median survivals from 14 d for control mice to 21 d when administered as a monotherapy, and from 23 d to 32 d for the combined treatment (p = 0.022). Radiotherapy, as expected, increased the expression of mesenchymal subtype-associated gene signatures, and this was efficiently rectified by G6-31. Furthermore, G6-31 also decreased the expression of immune-cell-specific gene expression patterns, both in tumors and in normal brain tissues, upon irradiation implying reduced proinflammatory signaling and myeloid cell infiltration in each of these tissues.
Conclusion:
Our preclinical data suggest that VEGF-targeting therapy could enhance the efficacy of fractionated radiotherapy in an immunocompetent glioblastoma model resembling the treatment-refractory mesenchymal subtype by interfering with mesenchymal enrichment and myeloid cell infiltration in response to irradiation. Furthermore, the data show that a concurrent blockade of VEGF signaling decreases proinflammatory signaling in normal brain tissue which is line with several clinical studies reporting reduced radiation necrosis in glioblastoma patients upon co-treatment with bevacizumab.
Targeting radioresistance in different pancreatic tumor models
ABSTRACT. Introduction
High-precision irradiation devices for preclinical tumor models provide the ability to mimic the treatment of cancer patients and translate novel therapeutic approaches into the clinic. This study compares different irradiation regimes regarding toxicity, tumor response, and survival after high-precision radiotherapy (RT) in an orthotopic xenograft pancreatic tumor mouse model. A combined approach of inhibiting the unfolded protein response (UPR) pathway and RT is evaluated in a subcutaneous syngeneic pancreatic tumor model to target radioresistance.
Materials and Methods
The human pancreatic cancer cell line MiaPaCa-2 was used to perform an orthotopic xenograft pancreatic tumor model. Imaging and image-guided RT (IGRT) were performed with the irradiation device “Small Animal Radiation Research Platform” (SARRP, Xstrahl Ltd, UK). IGRT was either delivered as single-dose RT (SBRT) with 25 Gy or fractionated (FSBRT) with a biologically equivalent cumulative dose of 45.71 Gy in 5 fractions in arc technique (gantry rotation from 178° to -178°).
A combined treatment approach was validated in a subcutaneous syngeneic tumor mouse model after intraperitoneal injection of 25 mg/kg of the UPR-inhibitor Kira8 or vehicle with and without 10 Gy RT in AP/PA technique.
Tumor volume, physical condition, and body weight were monitored at least weekly. Survival was analyzed by the Kaplan-Meier-Method and Log-Rank-Test (Mantel-Cox).
Results
In the orthotopic xenograft pancreatic tumor model, the median survival was significantly improved by high-precision IGRT (p < 0.0001) compared to the control: 68 days (unirradiated tumor-bearing mice), 115 days (SBRT), and 140 days (FSBRT). The average mouse body weight decreased by up to 15% in 80% (SBRT) and 12.5% (FSBRT). In contrast, no changes in body weight were observed in the control group.
In the subcutaneous syngeneic pancreatic tumor model, injection of Kira8 did not affect the median survival compared to the vehicle control (both 48 days), but an improvement of the median survival was achieved with RT (75 days). Most importantly, the combined treatment with Kira8 and RT significantly improved the median survival (189 days, p≤0.0001) without side effects.
Conclusion
We demonstrated the feasibility of different concepts of high-precision IGRT and long-term longitudinal follow-up in a clinically relevant orthotopic xenograft pancreatic tumor mouse model. Improvement of survival with tolerable adverse effects was achieved in all irradiated orthotopic pancreatic tumors.
We established a novel approach using the Kira8 in combination with RT to overcome the radioresistance. We demonstrated a significant radiosensitizing effect in a syngeneic subcutaneous tumor model.
This preclinical platform reflects the clinical situation and allows the evaluation of novel IGRT concepts and combined treatment strategies for pancreatic cancer treatment.
Preclinical development of physical and biological strategies to reduce radiation-induced cardiac toxicity
ABSTRACT. Purpose/Objective: Despite technological advances in radiotherapy (RT), radiation-induced cardiotoxicity (RICT) remains a common complication in patients with lung, oesophageal and breast cancer. In a preclinical model, we have recently recapitulated clinical observations showing the heart base as a radiosensitive subregion that could preferentially be avoided to reduce the risk of adverse events following treatment. Also, we are exploring the potential to repurpose clinically approved drugs to protect against RICT. This study reports recent data supporting heart base avoidance and responses to combined treatment with neprilysin/angiotensin inhibition (Entresto, ENT).
Materials/Methods: Studies were conducted using 12-week-old C57BL/6J mice irradiated under CBCT image-guidance using the small animal radiotherapy research platform (SARRP, Xstrahl) targeting different regions of the heart. To assess cardioprotective function of ENT, mice were irradiated with a single fraction of 20 Gy to the superior 2/3 of the heart as a 90º arc field arrangement and ENT (100 mg/kg/day) was administered in the drinking water from one week before irradiation. Longitudinal transthoracic echocardiography (TTE) was performed at baseline and at 10-week intervals up to 50 weeks after irradiation. All animals were monitored by transthoracic echocardiography (TTE) and global longitudinal strain (GLS) was assessed using two-dimensional speckle tracking (2D-STE).
Results: Heart base irradiation leads to BED-dependent changes in systolic and diastolic function at 50 weeks post-irradiation. GLS showed significant decreases in a BED-dependent manner as early as 10 weeks after irradiation. BED-independent increases were observed in the left ventricle (LV) mass, volume and myocardial fibrosis. Treatment with ENT resulted in significant preservation of cardiac function for up to 30 weeks after treatment.
Conclusions: Our model of cardiac base irradiation accurately captures clinical observations of the heart base as a radiosensitive subvolume with loss of cardiac function dependent on BED. Importantly, we show that GLS can accurately detect radiation-induced changes in cardiac strain at 10 weeks after treatment that are indicative of late functional loss at 50 weeks. We show that treatment with ENT can act to prevent radiation-induced cardiac dysfunction and is a potential strategy for clinical exploration.
Investigation of TLD Fading at Body Temperature: Implications for Long-Term Radiation Dosimetry
ABSTRACT. Thermoluminescent Dosimeters (TLDs) play a pivotal role in small animal radiation work by providing dosimetric information for quality assurance, treatment planning, and in-vivo studies. Despite being widely employed, the phenomenon of TLD fading– marked by a reduction in the measured signal–has gone uncharacterized for TLDs held at 37℃. Without proper fading characterization, there is a risk of underestimating the radiation dose. Therefore, fading characterization is essential to ensure the validity of dosimetric measurements in cases involving prolonged exposure or when monitoring radiation doses within living animals. This study sought to determine the fading behavior of TLD-100 dosimeters at body temperature (37℃) with elapsed time and varying radiation dose.
In this study, the sorting and characterization of TLD-100 chips were carried out using the Cameron et al., 1968 annealing method. This process involves an initial anneal at 400°C for 1 hr followed by heating the TLDs in an 80°C oven for 24 hours, followed by gradual cooling to room temperature on an aluminum block. The TLDs are then acutely exposed to Cs-137 and left to rest for 24 hours before being read out using a Harshaw 5500 TLD reader after 24 hours of non-exposure. These steps are repeated three times to characterize the response of each dosimeter.
For fading characterization, the dosimeter responses were analyzed after subjecting them to ionizing radiation from a well-characterized Cs-137 source. During exposure, the TLDs were either in a room temperature or 37℃ water bath. Charged-particle equilibrium was achieved by interposing a 2mm-thick acrylic sheet between the radiation beam and the dosimeters. The TLDs received a daily air kerma dose of 12mGy. Every 24 hours for 9 days, 5 TLDs were extracted from each water bath. Subsequently, the TLD responses were measured with increasing exposure and length of time at temperature.
Within the scope of this study, analysis of the TLD-100 measurements indicated no statistically significant fading effects with an uncertainty of 2.8% at k=1. Therefore, fading rates were not found to vary with either the elapsed time since the beginning of exposure or the dose received.
This research contributes to the ongoing efforts to enhance the reliability and accuracy of TLD-based measurements when used in-vivo. The absence of statistically determined TLD fading within the time-temperature-dose scale of this research increases the confidence in the stability of TLD-100 dosimeters. These findings contribute valuable insights to the understanding of TLD behavior under varying environmental conditions, reinforcing the utility of TLD-100 dosimeters in diverse dosimetric applications.
Preclinical mini-beam radiotherapy of a glioblastoma model
ABSTRACT. Introduction
Radiotherapy (RT) is an essential tool for cancer treatment to deliver curative doses to the tumor, while sparing surrounding healthy organs. Although normal tissue doses have been reduced thanks to sophisticated beam delivery methods combined with accurate image-guidance systems, treatment-related adverse effects are still among the major limiting factors of RT. As an example, the delivery of high doses required to eradicate aggressive brain tumors like glioblastoma, without exceeding the tolerance doses of normal tissues is still a challenging issue.
Novel preclinical approaches have been developed to tackle these problems, in this work we investigated the use of mini-beam radiotherapy (MBRT), based on small (300-400 microns) parallel beams of radiation (x-ray or protons) with a peak/valley dose pattern.
This study aims to test the efficacy of MBRT, its toxicity compared to conventional RT, and the role of immune system response.
Materials and methods
In vivo MBRT was performed by injecting two million of GL261 cells subcutaneously into the left flank of 30 C57BL/6 mice (weight 17-23g).
The mice were treated as follows: 10 received 25 Gy, single-fraction conventional RT, 10 mice received 25 Gy mean dose, single-fraction MBRT, and 10 mice were used as the control unirradiated group. Both conventional and MBRT treatments were performed using one single beam to keep the same radiation geometry. To compare the two techniques, tumor volume was monitored using a caliper. Survival curves and skin damage were evaluated for up to 50 days after RT treatments. MBRT was performed using a custom collimator developed in our group and installed on a small animal irradiator (SmART, PXI) at the preclinical imaging facility.
Results
The tumor volume measurements showed a complete tumor control for the open-field (OF) group and a good tumor control up to 21 days after treatment for the MBRT group with respect to the control. Survival curves showed a similar (not statistically different) survivor percentage for the MBRT and open beam RT groups. No acute radiation damages were found for the MBRT group while the mice belonging to the conventional beam group showed severe (72%), moderate (14%), or mild (14%) damages.
Conclusion
Considering that the survival curves of MBRT and open-field RT are similar we concluded that MBRT is a potentially novel treatment option with lower toxicity. At the same time, since no acute reaction was observed when using an average dose of 25 Gy, the MBRT mean dose could be increased to improve local tumor control. However, the underlying biological mechanism related to MBRT efficacy remains currently unclear and thus we aim to study the potential role of the immune system after treatment with radiotherapy.
Understanding myeloid and lymphoid cell heterogeneity in metastasis following radiotherapy
ABSTRACT. Introduction
Breast cancer (BC) is the most frequent malignancy in women, with advanced metastatic
BC currently considered incurable. Treatment of BC entails resection of the tumour, followed by radiotherapy (RT) and chemotherapy/immunotherapy. Despite this therapeutic protocol, most patients eventually experience metastasis. Adjuvant RT improves the risk of loco-regional recurrences, with several studies showing that its action is the result of both direct and indirect immune-related effects. However, the effects of local RT on distal metastatic sites are poorly understood. We hypothesize that these RT-driven systemic effects can change the composition and function of myeloid and/or lymphoid cells at metastatic sites, thus affecting the survival and growth of disseminated tumour cells (DTCs).
Materials &Methods
We are investigating our hypothesis in the context of triple-negative BC (TNBC), using orthotopic mouse models with induced lung metastasis and treated with a single, breast-localised dose of radiation delivered through the Small Animal Radiation Research Platform (SARRP), followed by downstream analysis of the results via flow cytometry, immunohistochemistry and scRNA sequencing. Concurrently, we are optimising a novel mouse model comprising (I) the orthotopic injection and spontaneous metastatisation of TNBC cells genetically engineered to isolate the metastatic niche and (II) the resection of the primary tumour to translate our results in a more clinically relevant scenario.
Results
Preliminary data from our lab show the ability of breast-localised RT to induce alterations within the immune landscape of the pre-metastatic lung, particularly suggesting a dual role of neutrophils, with both immunostimulatory and immunosuppressive capacities, especially within their interactions with the adaptive immune system. Moving to the metastatic setting, breast-localised RT on the primary tumour also induces a significant increase in the number of metastasis in the lung, along with numerical and functional differences in cellular populations from both lymphoid and myeloid lineages. Additionally, selective depletion of cellular populations modifies the total metastatic load of treated mice but does not revert their phenotype, suggesting a cooperative effect of different immune populations.
Conclusion
Future experiments will focus on dissecting the specific role of immune cells and the mechanisms influencing DTCs’ outgrowth in the lung, while also investigating these effects in a clinically relevant model including the resection of the primary tumour. This project aims to understand how current therapies influence immune cell behaviour in the metastatic TME to identify relevant targets for novel immunotherapy approaches against metastatic disease.
Precision of treatment couch, CB-CT and micro-CB-CT of a small animal radiotherapy device
ABSTRACT. Introduction
Precise small animal radiotherapy devices require a mechanical accuracy of better that 100 microns of all components. The goal of this study is to analyze the mechanical accuracy of treatment couch and cone-beam computed tomography systems (CB-CT) of a dedicated small animal radiotherapy device.
Materials & Methods
A “SmART+” (Precision X-Ray, Madison, CT, USA; installed in 2021) was analyzed in this study. It has two X-ray-tubes mounted to the gantry, rotating 360° around the computer-controlled carbon-fiber the treatment couch. A 225 kV X-ray tube (Comet AG Industrial X-ray Modules, Flamatt, Switzerland) is used for treatment and imaging in combination with an opposing flat panel imager. An additional 60 kV tube (called micro-CT option) can swing in the beam (hinge mechanism) to enhance the spatial resolution, according to manufacturer’s specifications, from 50 microns to 20 microns, respectively.
At first, we modified the treatment couch and micro-CT as described below. We attached a circular 1 mm collimator to the main x-ray tube to irradiate GafchromicTM EBT3 films, which were positioned on the treatment couch in two different planes. The couch was moved in 10 mm steps with the SmART system software. The resulting spot pattern was digitized with a scanner and analysed for geometric accuracy.
The image quality of both main X-ray and micro-CT tube was analysed with a commercial “Enhanced Resolution Performance Evaluation Micro-CT Phantom” (Shelley Medical Imaging Technologies, Ontario, Canada), as well as with custom 3D-printed test patterns.
Results
As the SmART standard carbon fiber treatment couch shows a weight-depending sack on the table’s load, we corrected for this using a hinge mechanism, which can be fine tuned with a screw, so that a correct horizontal position of the couch is possible for any regular weight on the treatment couch.
As the standard hinge mechanism of the micro-CT X-ray tube collides with the couch in treatment position, thus requiring a very high level of attention before use, we changed the hinge mechanism to a sliding one, so that to avoid mechanical collisions at any time.
The average deviation between set-up and measured positions of the treatment couch is 3.31 microns in the XZ-plane and 7.99 microns in the YZ plane, respectively, in between neighbouring positions 10 mm apart. When moving longer distances of 20 – 140 mm, it is 58.61 microns in x-direction, and 59.02 microns in z-direction, respectively.
The micro-CT option gives a visibly better spatial resolution than the standard X-ray tube, as demonstrated with both the Shelley phantom and custom printed objects. However, the volume covered with the micro-CT is much smaller (dimensions of a mouse head) than the volume covered with the standard X-ray tube (whole mouse). As the spatial resolution also strongly depends on the used scanning parameters, it is hard to say whether the specified resolutions of 50 microns or 20 microns can be reached under all circumstances.
Conclusion
The accuracy of treatment couch is sufficient for all experiments. The micro-CT option gives a visibly better spatial resolution than the standard X-ray tube.
Efforts among the UK professional community in support of the standardisation of dosimetry aspects of preclinical and radiobiological research.
ABSTRACT. Introduction
The necessity for standardisation in preclinical and radiobiological research was demonstrated during the 2011 NIAID/NCI/NIST Workshop on Radiation Dosimetry Standardisation for Radiobiology. At the meeting, collaborative groups and professional bodies emphasised the challenges associated with the harmonisation of the preclinical dosimetry chain. The adequateness of the established dosimetry codes of practice for reference dose measurements on various types of preclinical irradiation devices, the traceability of absorbed dose measurements in preclinical samples (small animals or cell cultures) and the lack of consensus when reporting dosimetric aspects of irradiations were all emphasised. More recently, in 2023, in a publication presenting a road map for preclinical radiotherapy, the potential for a lack of standardisation or inaccurate dosimetry to jeopardise the translation of preclinical research into new radiation therapies, was again highlighted [1]. We will update on efforts within the UK preclinical and radiobiological research community to address these issues.
Materials & Methods
As part of their strategy to beat cancer, Cancer Research UK (CRUK) funded a network of radiation research centres known as RadNet. This network aims to support cancer radiation research by a multidisciplinary and collaborative approach. Preclinical and radiation biology research at a high standard is achieved through dedicated labs and their associated small animal irradiators or conventional cabinets. In 2021, as part of a working group in Molecular Imaging and Radiotherapy, the Standardisation and Dosimetry Subgroup was created. The group is co-chaired by scientists from NPL and the CRUK RadNet Manchester group. Since its formation it has attracted 20 members from across the RadNet network, other UK research institutions and international partners, with a balanced composition of those with medical physics or radiation biology background and at varied career stages.
During the first two years, the main objective was to disseminate activities within the group, linked to preclinical dosimetry and to increase awareness of the need for standardisation. The group also aims to establish beneficial national and international collaborations.
Results
A total of 16 meetings have taken place, with the delivery of 9 internal presentations on preclinical dosimetry topics. Furthermore, the group has supported two national surveys: RadNet imaging biomarkers MOC (mucoepidermoid carcinoma) tumour protocols and a preclinical dosimetry audit. It also provided quality assurance and dosimetry guidelines for a multi-centre project: “Radiotherapy in a standardised Head & Neck preclinical model”.
The dosimetry standardisation group also actively collaborates with the IPEM Preclinical Working Party, which is currently drafting an awareness paper to the wider research community on the need for standardisation of dosimetry for preclinical research. Active collaborations with the National Physical Laboratory (NPL) and more recently with the Compatibility of Irradiation Research Protocols Expert Roundtable (CIRPER) and the International Dosimetry Exchange Working Group have been developed.
Conclusions
Multidisciplinary and international collaborations contribute to increasing awareness around preclinical dosimetry, particularly its challenges and the need for standardisation. Future work by the group will concentrate on developing multi-institutional projects, focussing on implementing more reliable preclinical dosimetry protocols, improving the translational impact of preclinical studies.
Assessment and mitigation of geometric distortions in Magnetic Resonance images for pre-clinical irradiation at 15.2T
ABSTRACT. Introduction:
Ultra-high field (UHF) magnetic resonance (MR) systems are spreading to advance pre-clinical research, e.g. for non-invasive response assessment. However, system-dependent factors, such as magnetic field inhomogeneities (ΔB_0) and gradient non-linearities, induce geometric distortion.
Nevertheless, high-geometric accuracy of MR images is essential for small-animal irradiation with target volumes≤1cm³.
This work focuses on quantifying and mitigating system-dependent geometric distortion at 15.2T through prospective optimization of shimming strategies and registration with Computed Tomography (CT) images for voxel displacement correction, as part of UHF-MRI Quality Assurance (QA) for pre-clinical irradiation.
Materials&Methods:
A nylon cylindric grid phantom, adapted from a prototype (O’Callaghan et. al.10.1371/journal.pone.0096568), was manufactured to fit the 3.5cm inner-diameter coil used with a 15.2T MRI scanner (Bruker BioSpin, Germany). The phantom was initially measured empty with spiral cone-beam µCT (Molecubes,Belgium, 50kV) with 200µm³ isotropic resolution and then filled with Copper(II) sulphate solution (1mg/ml) for MR measurements.
The following coronal 3D MR images were acquired with readout encoding direction in ±Z-direction, 200µm³ isotropic resolution, Field of View=42x30x30mm³and receiver BandWidth=100KHz (G_z=56.1mT/m): 1) ΔB_0-maps based on double-gradient-echo (GRE) sequence (TR/TE= 11.2/2.32ms; FA= 10°; Averages=2), 2) GRE (TR/TE= 4/17ms; FA=10°; Averages=2).
Three shimming scenarios were established to optimize ΔB_0-related distortions. Three measurements were performed for each scenario at different time points without repositioning.
Scenario-1: automatic iterative global 1st-order shimming.
Scenario-2: scenario-1 plus automatic estimation up to the 3rd-order shimming coefficients based on pre-scanned ΔB_0-map acquired before GRE and SE acquisitions.
Scenario-3: scenario-1 with pre-defined optimized initial values up to z3.
The displacement (Δr_B0) from ΔB_0 was calculated by dividing ΔB_0 by the gradient strength G_z in ±Z-direction for all the scenarios in three volumes of interest (VOI) (5x5x10mm³, 5x5x15mm³ and 5x5x20mm³) around the magnet isocentre using an in-house developed Python script.
In a second moment, GRE images were rigidly registered to a CT image using an in-house developed Python script, which subsequentially calculated the displacement (Δr_tot)-incorporating all system-dependent distortions- between manually positioned landmarks along ±Z-direction on both images via Euclidean distance for all three VOIs.
The contribution of the gradient non-linearity distortion (Δr_GNL) was obtained by subtracting the calculated Δr_B0 from the Δr_tot estimation.
Results:
The Δr_B0 increased with expanding VOI sizes; the initial 0.48±0.02mm value in scenario-1 was reduced to less than 0.43±0.01mm in scenario-2. Scenario-2 showed the lowest displacement across all VOIs. Scenario-3 resulted in values within 0.51±0.02mm for the largest VOI, higher than the others, and 0.24±0.01mm was comparable to scenario-1 for the smallest VOI.
After rigid-registration, the best shimming optimized scenario-2 presented a Δr_tot in GRE images of 0.52±0.06mm to 1.25±0.15mm for the smallest and largest VOIs, respectively.
A Δr_GNL contribution of 0.82mm was reached for the largest VOI, which generally increased with increasing distance to the isocentre in ±Z-direction.
Conclusion:
Shimming optimization and rigid MR to CT registration were established to evaluate system-dependent distortions at 15.2T. A shimming optimization workflow is currently being introduced into regular QA for in-vivo pre-clinical protocols. Ongoing work focuses on non-rigid image registration to further reduce the contributions from gradient non-linearities to geometric distortions.
SAPPHIRE - Establishment of Small Animal Proton and PHoton Image-guided Radiation Experiments
ABSTRACT. Introduction:
Proton therapy is a popular treatment modality in radiooncology due to its high spatial dose conformity compared to conventional radiotherapy, allowing it to spare normal tissue. Although patient treatment is already established, there is still a need for preclinical in vivo experiments on proton RBE and new treatment modalities like FLASH.
The established proton and photon irradiation setups at the University Proton Therapy Dresden (UPTD) enabled successful experiments. However, the challenge was to create a unified framework accommodating both types of irradiation for comparable studies. Additionally, introducing onboard imaging minimizes uncertainties and movement between imaging and irradiation, ensuring reproducible investigations.
This led to the establishment of SAPPHIRE, a setup using the small animal imaging and irradiation device SmART+ IB for preclinical image-guided proton and photon therapy experiments, bridging preclinical and clinical applications.
Material & Methods:
The existing setup for mouse brain irradiation with protons (https://doi.org/10.3389/fonc.2022.982417) is integrated in the SmART+ IB, whereby scattered dose contributions were minimized through optimized component positioning via Monte Carlo simulations.
Four CBCT imaging protocols were evaluated for image quality, examining homogeneity within phantoms, contrast-to-noise ratio, and mouse skull structure visibility, while also quantifying the imaging dose per mouse.
A custom-made QA phantom was employed to ensure the accurate alignment of the proton beam, collimator, and mouse target before each experiment.
To predict the proton radiation range accurately during treatment planning, device-specific empirical and stoichiometric Hounsfield Look-Up Tables (HLUT) were created using tissue-equivalent materials for the best imaging protocol.
To characterize the proton beam, lateral and depth dose profiles were measured at 90 MeV energy, employing EBT3 film stacks, a microDiamond detector and a Giraffe multilayer ionization chamber. Photon irradiation’s depth dose profiles and lateral profiles were obtained using EBT3 films for seven collimators at the SmART+ IB’s isocenter operating at 200 kV, 20 mA, with a 0.5 mm copper filter.
Results:
The analysis of the CBCT homogeneity reveals a negligible range uncertainty of approximately 0.1 mm. The 60 kV, 0.5 mA imaging protocol is chosen for future experiments due to its balance between low imaging dose (<26 mGy) and high image quality. The QA procedure provides accurate alignment with a deviation of <0.35 mm, requiring approximately 45 minutes before each beamtime.
Acceptable range uncertainties of ∆r80 ≤ 0.51 mm are determined for the stoichiometric and empirical HLUT.
A comprehensive 3D analysis of proton and photon beams yield expected FWHM and penumbra values. Using various collimators, the dose rates range from 0.78-2.75 Gy/min at a 10 mm depth for photons and reach 17.72 Gy/min at the Bragg peak for protons. Both models are implemented in the treatment planning software µRayStation and will be validated soon.
Conclusion:
The technical prerequisites of the SAPPHIRE project are fulfilled utilizing the SMART+ IB, designed for preclinical image-guided proton and photon comparison experiments at the UPTD. Both photon and proton treatment models were implemented in µRayStation and suitable setups and QA protocols for mouse imaging and irradiation were established, paving the way for clinic-like preclinical animal experiments in the future.
Preclinical Oxygen Imager for Oxygen-guided Radiation Therapy in Murine Tumor Models
ABSTRACT. Introduction: Hypoxia (partial oxygen pressure below 10 mmHg) is a common feature of many solid tumors and one of the important factors negatively affecting the response to radiation, chemo, and immuno-therapy. Therefore, by taking oxygen into account, we can improve the response to cancer treatment in preclinical models and develop a basis for moving promising therapies to clinics.
Materials and Methods: Pulse electron paramagnetic resonance oxygen imaging (pEPROI) is a non-invasive magnetic resonance imaging method focusing on imaging partial oxygen pressure (pO2) in tissues. pEPROI measures the relaxation maps of a water-soluble oxygen-reporting trityl molecule that distributes in a body upon injection and converts them into oxygen images. Over the past decade, EPROI has advanced at a rapid pace to deliver 1-10 minutes long oxygen images in live mice and rabbit limbs with oxygen resolution and absolute accuracy up to 1 torr 1,2. Oxygen-reporting molecules (OX063 and OX071 trityl radicals) are stable radicals with very low toxicity and can be administered systemically through intravenous (IV) or intraperitoneal (IP) injections or can be directly administered into the tumor.
Results: We have developed a 25 mT preclinical EPROI instrument, JIVA-25, that has recently been used for many studies in cancer, leading to the demonstration of oxygen-guided radiation therapy in three preclinical tumor models and assessment of an FDA-approved radiosensitizer drug 1,3-5. We also demonstrated the potential of pEPRI to assess the blood-brain barrier (BBB) leakage and brain pO2 in a mouse model of neuroinflammation in a recent study 6. OX071 penetrated the damaged BBB while the intact BBB stopped ~95% of the trityl. This study enables oxygen-guided therapy in brain tumor models with compromised BBB.
Conclusion: The availability of the preclinical oxygen imager paves the way for precision radiotherapy taking tumor oxygenation into account in preclinical tumor models. This technology will allow the translation of promising therapies to the clinics.
References:
1 Gertsenshteyn, I. et al. Absolute oxygen-guided radiation therapy improves tumor control in three preclinical tumor models. Frontiers in Medicine 10, doi:10.3389/fmed.2023.1269689 (2023).
2 Epel, B., Viswakarma, N., Sundramoorthy, S. V., Pawar, N. J. & Kotecha, M. Oxygen Imaging of a Rabbit Tumor Using a Human-Sized Pulse Electron Paramagnetic Resonance Imager. Molecular Imaging and Biology, doi:10.1007/s11307-023-01852-3 (2023).
3 Rickard, A. G. et al. Evaluating Tumor Hypoxia Radiosensitization Via Electron Paramagnetic Resonance Oxygen Imaging (EPROI). Molecular Imaging and Biology, doi:10.1007/s11307-023-01855-0 (2023).
4 Martin, R. M. et al. Toward a Nanoencapsulated EPR Imaging Agent for Clinical Use. Molecular Imaging and Biology, doi:10.1007/s11307-023-01863-0 (2023).
5 Shaw, M. A. et al. SOX71, a biocompatible succinyl derivative of the triarylmethyl radical OX071 for in vivo quantitative oxygen mapping using Electron Paramagnetic Resonance. Molecular Imaging and Biology (2023, in press).
6 Epel, B. et al. Assessment of Blood-Brain-Barrier Leakage and Brain Oxygenation in Connexin-32 Knockout Mice with Systemic Neuroinflammation Using EPR Oxygen Imaging. Magnetic Resonance in Medicine (2024, in press).