Background and Aims

Methods

Results

Conclusions

Background and Aims

Reduced skin toxicity is now shown pre-clinically after >40Gy/s FLASH radiotherapy (RT) as compared to ~0.03Gy/s conventional RT. While this area of study has been advanced significantly, the biological basis of the FLASH sparing effect on skin is yet to be discovered, and diagnostic tools to assess the response biological mechanisms and tools that can be used in the translation of FLASH study in humans are needed. Here we report on direct mechanistic in situ measurements of skin damage, with the aim to quantify and compare microstructural and microvascular changes in skin after RT, while varying dose and supporting the findings with ex vivo histological observations.

Methods

In IACUC approved animal study, right thighs of 54 nude mice (n=6 per dose group) were treated with 0-40Gy single doses of 300Gy/s FLASH and conventional RT. Skin was imaged with functional optical coherence tomography - a non-invasive microscopic 3D imaging technique with light penetration at 1-3 millimeter depths in biological tissues.

Results

Quantification was completed for skin pigmentation, epidermal thickness, remodeling of collagen fibers, alteration of blood and lymphatic networks. These demonstrated inter-connected “passenger versus driver” temporal behaviors of skin tissue components in both RT-treatment types. Response to FLASH RT was characterized by reduced damage to collagen bundles and blood/lymphatic networks together with less desquamation of skin surface (less damage to epidermis) and higher pigmentation.

Conclusions

This study presents first of its kind in-vivo functional longitudinal observations and quantification of the comparative FLASH effects in a mouse model of skin early damage, healing and recovery.

Background and Aims

Using glioblastoma as model, the aim of the present study was to investigate the role of G2/M arrest in tumor response to FLASH-RT and to characterize Damage Associated Molecular Patterns (DAMPs) that might amplify anti-tumor immunogenic response.

Methods

In vitro, GL261, H454 and PDGC2159 GBM and HaCat normal cells were synchronized (or not) in G2/M phase using a CDK1 (9uM) or PLK1 (25nM) inhibitor 24hours before 20Gy FLASH-RT (2.10³Gy/s, 2 pulses of 10Gy, 100Hz) or CONV-RT (~0.1-0.2Gy/s, 10Hz) with eRT6 (Jorge, 2019). Calreticulin, HSPA5, ATP, HMGB1, DNA release, micronuclei formation and cGAS-STING-type I IFN response were investigated. In vivo, murine GL261 and PDGC2159 GBM cells were orthotopically grafted to C57Bl6 and Swiss nude mice. Mice were treated with a single dose of 10Gy delivered Whole-Brain either with FLASH (≥10⁷Gy/s, 1 pulse) or CONV-RT (~0.1-0.2Gy/s). Tumor control, normal brain toxicity, immune response and in situ vaccination were evaluated.

Results

In vitro, the level of micronuclei positive cells was similar after FLASH and CONV (40% vs 0% in non-RT) and HMGB1 mRNA level was enhanced (+1.8fold) in FLASH vs CONV irradiated samples. G2/M blockade significantly increased micronuclei formation (+20%), and cGAS mRNA level (+2.33fold) in FLASH vs CONV irradiated samples. Other markers were not modified. In vivo experiments are ongoing.

Conclusions

These preliminary results support a G2/M-dependent release of DAMPs after FLASH irradiation that might trigger downstream immune response. Experiments are ongoing to characterize this response along with anti-tumor efficacy and normal toxicity in immune-deficient/competent mice.

Background and Aims

To investigate the biological effect of murine ventral skin irradiation with FLASH radiotherapy compared with conventional irradiation.

Methods

Female FvB mice were randomly assigned to three groups: control, conventional (CONV) and FLASH groups. Mice were irradiated at 9 to 19 Gy of CONV (0.1 Gy/s) or FLASH (38.5-600 Gy/s) irradiation using traditional and modified Elekta Synergy linac (6 MeV), respectively. Doses were verified by Gafchromic films positioned under the body. Body weights were recorded every week 1 to 6 weeks after irradiation. Enzyme linked immunosorbent assay (ELISA) were performed in skin tissue lysis and serum samples of the mice for four inflammatory cytokines: tumor necrosis factor-α (TNF-α), interferon-γ (IFN-γ), interleukin-6 (IL-6) and IL-10. Flow cytometry using antibodies for CD3, CD8, CD4 and CD45 in blood were performed pre- and 1-week post irradiation.

Results

A significant increase in weight percentages relative to pre-irradiation were observed in the FLASH group, and the alteration of serum and skin tissue levels of TNF-α, IFN-γ, IL-6 and IL-10 induced by FLASH was mild compared with that of CONV. The CD8+/CD45+ ratio in the blood were higher in the CONV than in FLASH and pre-irradiated ratio. These data indicate that lower inflammatory cytokine levels of serum and skin tissue in FLASH could be the result of minor immune overactivation.

Conclusions

Ultra-high dose rate electron FLASH caused less body weight loss, minor inflammatory cytokine levels of serum and skin tissue, as well as less CD8+/CD45+ ratio in the blood. Thus, electron FLASH irradiation represents a new approach of radiotherapy.

Background and Aims

We have previously found that FLASH-irradiation with a pulsed electron beam (average doserate ≥600Gy/s) was less efficient to sterilize cancer cells in-vitro compared with conventional doserate irradiation (CONV, 0.2Gy/s). In the current work we aimed at investigating the effect for a malignant cell line both in vitro and in vivo.

Methods

Radiation response of melanoma cell line B16_F10 was determined in-vitro by clonogenic assays for an absorbed dose in the range 0-9 Gy comparing FLASH to CONV. In-vivo-response was studied in a syngeneic mice model (C57BL/6J) with subcutaneously injected B16_F10-tumors, irradiated to an absorbed dose of 15, 20 or 25Gy (FLASH and CONV). The tumor growth was quantified by using the relative tumor volume, normalized to unity at the time of irradiation, TV_rel.

Results

The in-vitro results showed a significantly increased survival after FLASH compared with CONV (F-test: p=0.02). Tumor growth curves in-vivo were similar for CONV and FLASH at 15 and 20Gy, but FLASH was relatively less efficient at 25Gy. Four weeks after irradiation with 25Gy, a relative tumor volume of TV_rel<1 was seen in 2/9 mice in the CONV group but in 0/8 mice in the FLASH group. A relative tumor volume of TV_rel<4 was seen in 5/9 mice in the CONV group but 0/8 mice in the FLASH group. Severe skin toxicity was observed in 5/9 vs 0/8.

Conclusions

FLASH may be less efficient than CONV to sterilize malignant cells in-vitro as well as in-vivo. Future work will address the differential response between normal tissue and tumors at higher doses.

Background and Aims

Compared to Standard dose rates, the high dose rates of FLASH radiation can reduce radiotherapy toxicities to normal tissues. We examined the potential of FLASH-proton radiotherapy (F-PRT) to treat murine sarcomas and protect normal epithelial and mesenchymal tissues relative to the effects of standard-proton radiotherapy (S-PRT).

Methods

Mice received 30 or 45 Gy of F-PRT (69-124 Gy/sec) or S-PRT (0.39–0.65 Gy/sec) to their hind legs. Skin, muscle and bone injuries were recorded as acute through chronic macroscopic and/or microscopic observations of radiation-induced damage. Murine skin and bone RNAseq analyses were performed to delineate involved mechanisms. Skin stem cell depletion, inflammatory reaction and TGF-β levels were evaluated, and antitumor efficacy of F-PRT was compared to S-PRT in two murine models of sarcoma.

Results

Fewer severe morbidities were induced by F-PRT, with RNAseq revealing S-PRT to upregulate pathways involved in apoptosis signaling and keratinocyte differentiation in skin, and osteoclast differentiation and chondrocyte development in bone. Accordingly, F-PRT reduced skin injury, stem cell depletion and inflammation; mitigated lymphedema; and decreased myofiber atrophy, bone resorption, hair follicle atrophy, and epidermal hyperplasia. Equipotent control of sarcoma growth was achieved by the radiation modalities. Finally, S-PRT produced higher levels of TGF-β1 in murine skin than did F-PRT, and this finding was corroborated in the skin samples of dogs treated on a F-PRT clinical trial.

Conclusions

F-PRT can alleviate radiation-induced damage to both epithelial and mesenchymal tissues without compromise to sarcoma response; continuing investigation will further F-PRT translation to the clinic.

Background and Aims

The normal tissue toxicity sparing and survival benefits of ultra-high dose rate radiation (FLASH) remain poorly understood. We present preliminary results of mouse abdomen FLASH proton radiation from a low-energy proton system (50 MeV) optimized for small animal radiobiological research.

Methods

We radiated 6-7 week old female C57BL/6 mice with partial abdomen radiation using the plateau region of a continuous (unpulsed) cyclotron-generated 50 MeV preclinical proton beam, transmitting through the abdomen, with 1.5cm width of beam via customized vertical and horizontal collimators. Mice were stratified into 3 groups: 1) control/sham radiation (n=8); 2) conventional dose rate (20Gy at ~1Gy/sec, n=19); and 3) FLASH (20Gy at 48-93Gy/sec, n=22). Mice were observed for survival. Colon tissue was harvested at 1-hour post-radiation. H&E and immunohistochemistry was performed for: yH2aX and cleaved caspase-3. Experiments were repeated in triplicate.

Results

Survival was different between FLASH and conventional groups: FLASH (13 days post radiation, 36.4% survival); conventional (15.6% survival, P = 0.04 ) [Figure 1]. One-hour post radiation, lower cleaved caspase-3 IHC staining was seen in the FLASH group versus conventional group, while yH2aX staining was similar in both groups [Figure 2].

Conclusions

Preliminary results of mouse partial abdomen FLASH proton radiation from a 50 MeV beam suggest FLASH proton radiation leads to better survival than conventional dose rate radiation. More studies are ongoing.

Background and Aims

In conventional proton therapy, Bragg peaks are positioned inside the tumor and in a margin surrounding it. This causes uncertainties in LET, RBE and range, endangering critical structures distal from the tumor. To combat this, shoot-through flash beams position the Bragg peaks behind the patient. While potentially losing the dosimetric advantage of Bragg peak plans, the FLASH protective effect may compensate for this. A planning study was performed on 5 neuro cases to evaluate dose-averaged LET_D and robustness for both planning strategies.

Methods

5 neuro cases were planned using four 227 MeV Pencil Beam Scanning proton beams. LET_D was calculated for the clinical plan and the shoot-through plan, applying a 2Gy dose threshold (RayStation 10A & 9AR-IonPG). A FLASH protective factor of 1.5 was used for tissues outside the CTV. Robust evaluation was performed considering movement (0.1 cm, 14 directions) and density uncertainty (±3% throughout entire volume).

Results

Clinical plans showed large LET_D variations compared to shoot-through plans. For shoot-through plans, LET_D merely reflects stopping power ratio differences (Fig. 1,2) and the maximum LET_D in OAR is 2–6 times lower. Although less conformal, shoot-through plans met the same clinical goals as the clinical plans. The shoot-through plans are almost entirely robust to density uncertainties (Fig 3a). Considering both spatial and density uncertainties, the target coverage is less robust while the OAR D2% is in general more robust compared to clinical plans (Fig 3b).

Conclusions

Proton shoot-through flash beams avoid LET_D and range uncertainties, provide adequate target coverage, meet planning constraints and are robust.

Background and Aims

We present the rigorous commissioning of a modified LINAC to deliver ultrahigh dose-rate (UHDR) electron beam and implementation of its model in a widely adopted treatment planning system (TPS) with minimal changes to the clinical setting.

Methods

A Varian Clinac 2100C/D was converted to deliver UHDR beams by withdrawing the target and scattering foil in 10MV x-ray mode. Beam characteristics and stability were quantified by film, Cherenkov, and scintillation imaging. The Geant4 generated beam model was validated with film and implemented in Varian Eclipse TPS. Electron FLASH radiotherapy (eFLASH-RT) plans were generated for representative mammal and human patient cases accounting for complex geometries and anatomical inhomogeneities.

Results

The surface mean-dose-rate at the isocenter was >230Gy/s for all measured fields with adequate long-term stability (deviations of output <7%, symmetry/flatness <2%, spatial shift and FWHM <2mm). The TPS model was validated to clinical accuracy (average error <1.5% for lateral profiles and <2% for percent-depth-dose profiles). Treatments plans were generated and accurately delivered to normal porcine skin surface tissue and a melanoma tumor in a canine’s posterior oral cavity. A human eFLASH-RT plan comparable to a conventional electron plan was achieved by utilizing routine accessories, oblique gantry angle and couch kick.

Conclusions

Treatment planning and accurate delivery of eFLASH-RT were feasible in minimally modified radiation oncology clinical settings. The modifications and open-source TPS model are readily transferable to facilitate clinical translation of eFLASH-RT.

Acknowledgment: This work was supported by the Norris Cotton Cancer Center (grant P30CA023108) and Thayer School of Engineering (seed funding and grant R01EB024498).

Background and Aims

To exemplify the potential and limitations of two approaches to FLASH treatment planning with protons by testing them on two clinically realistic scenarios and different FLASH-specific parameters

Methods

We selected two planning approaches to be delivered with a cyclotron: 3D range modulator (3DRM) and transmission beams(TB). (See Table below for details on the beam delivery parameters.) We associated each planning technique with a disease site and a clinically applied hypofractionation protocol ( 3DRM - liver - 3x25Gy, TB - lung - 3x20Gy). We evaluated the resulting dose distributions for different beam currents (200nA and 800nA at isocentre), two dose rate definitions (dose-averaged dose rate (DADR) and a sliding time window), two minimum dose thresholds and two dose rate thresholds for the FLASH effect (4Gy and 8Gy, 40Gy/s and 100Gy/s, respectively).

Results

Both techniques achieved acceptable dose distributions with a limited number of fields (liver - 1 field, lung - 3 fields) for FLASH proton plans. All combinations of beam intensity, dose rate definition, dose and dose rate threshold we investigated were associated with some level of FLASH dose, suggesting that these disease sites and dosimetric protocols are reasonable candidates for FLASH proton therapy. The figure below shows an example of the results for 200nA and 4Gy and 40Gy/s thresholds.

Conclusions

Treatment planning studies are a useful tool to test candidate disease sites, protocols and planning techniques for proton FLASH. The next step will be to include additional combinations of beam production systems, planning techniques, and patient anatomies.

Background and Aims

This work aims to study transmission proton pencil beam scanning (PBS) FLASH radiotherapy (RT) planning for liver cancer cases based on the parameters of a commercially available proton system under FLASH mode.

Methods

An in-house TPS was developed to perform intensity-modulated proton therapy (IMPT) FLASH RT planning. Single-energy transmission PBS plans of 4.5 Gy x 15 fractions were optimized for seven hepatocellular carcinoma patients, using 2 and 5 fields combined with 1) the highest minimum MU/spot from 100-400, and minimum spot time (MST) of 2 ms; 2) the minimum MU/spot of 100, and MST of 0.5 ms. Then, the 3D average dose rate (ADR) distribution and major OARs' dose metrics were characterized to evaluate the plan quality for different combinations of field numbers and MSTs.

Results

Shorter MST are generally associated with better dose quality while more fields are only associated with better target uniformities. Fewer fields will allow higher OAR FLASH coverage with 2 ms MST compared to the 0.5 ms. For 2-field plans, dose metrics and V_40Gy/s of some OARs have large variations due to selecting beam angles and the distance to the target. The transmission plans yield inferior dose quality to the conventional IMPT plans.

Conclusions

For the challenging hypofractionation liver case with smaller fractional doses (4.5Gy/fraction), using fewer fields can allow higher minimum MU/spot, resulting in higher OARs FLASH dose rate coverages while achieving similar plan quality compared to plans with more fields. Shorter MST can achieve better plan quality and comparable or even better FLASH dose rate coverage.

Background and Aims

With a sizeable patient population, a target largely comprising healthy tissue, and clinically relevant late effects, FLASH whole breast irradiation (WBI) merits consideration. Transmission beams provide a practical way to deliver ultra-high dose rate proton therapy. However, the large WBI volumes make it harder to achieve FLASH dose rates with pencil-beam-scanning (PBS). We therefore performed a simulation study to identify PBS machine characteristics needed for such treatments.

Methods

For a left-sided breast case (861cc) a single-field spot-reduced plan was generated using 250MeV transmission beams. ‘PBS dose rates’ were calculated, considering the total time (including dead-times) to deliver 95% of the dose in each voxel. We varied maximum beam current at isocenter (200, 400, 800nA), energy-layer-wise or spot-wise current, minimum spot duration (0.5, 1, 2ms), and fraction dose (5x5.7Gy, 2x9.74Gy; equivalent BED₃). The percentage of dose delivered above FLASH thresholds was evaluated, considering dose rate thresholds of 40Gy/s and 100Gy/s, and dose thresholds of 4Gy and 8Gy.

Results

For 40Gy/s dose rate threshold, spot-wise currents generally provided >70% of dose delivered above FLASH thresholds, with little dependence on beam current and spot duration (Figure 1). When using energy-layer-wise currents, comparable FLASH dose was achieved only for 9.74Gy fraction dose and 0.5ms minimum spot duration. For 100Gy/s dose rate threshold, substantial FLASH dose was obtained only with extreme machine settings (i.e. 800nA, spot-wise, <=1ms).

Conclusions

Assuming large fields do not necessarily preclude a FLASH effect, FLASH WBI is theoretically achievable, but may require large fraction sizes, and may be (too) demanding for current PBS machines.