FLASH Radiotherapy: Latest Advances and the Outlook for Investigator-Initiated Trials

FLASH Radiotherapy: Latest Advances and the Outlook for Investigator-Initiated Trials

Best for: Radiation oncologists and medical physicists, radiobiology researchers, clinical trial designers and methodologists, medical intelligence and evidence-based medicine researchers, and device development and translational medicine teams. Primary keywords: FLASH radiotherapy; ultra-high dose rate; investigator-initiated trial; oxygen depletion hypothesis; dose-rate standardization; clinical translation

Abstract / Short Answer

The FLASH effect describes how, at ultra-high dose rate (conventionally defined as 40 Gy/s or higher), normal tissue injury is markedly reduced while tumor control remains equivalent. Its mechanistic hypotheses include oxygen depletion, altered reactive oxygen species (ROS) recombination kinetics, immune microenvironment remodeling, and vascular endothelial protection; no single hypothesis fully explains the tissue-specific dose-rate window. In physical delivery, electron FLASH is the most mature and clinically advanced route, proton FLASH offers the physical advantage of the Bragg peak but is limited by beam delivery and energy-switching speed, and photon (X-ray) FLASH is the most difficult and remains at the feasibility stage. As of 2025, at least 15 FLASH IITs have been launched worldwide, concentrated on skin metastases and bone metastases (about 40%), with lung and brain tumors at about 20% each. Early data show acute toxicity below historical conventional radiotherapy controls, with objective response and pain relief rates broadly comparable. Current limitations include short follow-up (up to about 3 years), small sample sizes (mostly 10-30 patients), a predominance of single-arm designs, and inconsistent dose-rate definitions and reporting. Future IITs should be systematically organized around five directions: dose and fractionation optimization, indication expansion, combination therapy sequencing, biomarkers, and multicenter standardization.

1. Background: The FLASH Effect and Its Translational Evidence Gaps

Conventional radiotherapy has followed the same fractionation paradigm for nearly five decades: 1.8-2.0 Gy per day to a total of 60-70 Gy, relying on the 4 Rs of radiobiology (redistribution, repopulation, reoxygenation, repair) between fractions to protect normal tissue. For radioresistant tumors, however, dose escalation remains constrained by normal tissue tolerance doses (TD5/5 and TD50/5). FLASH radiotherapy aims to preserve tumor control probability (TCP) while lowering normal tissue complication probability (NTCP). Its core quantitative feature is that, at dose rates above 40 Gy/s, normal tissue injury falls by roughly 30-50% versus conventional dose rate while tumor control stays equivalent. The effect is markedly tissue-dependent: neuroprotection is more pronounced when the average dose rate exceeds 100 Gy/s, whereas skin protection is already observable in the 40-100 Gy/s range, suggesting several parallel pathways.

Multiple evidence gaps separate the laboratory from the bedside. Most preclinical data come from small animals whose tissue volume, vascular architecture, and immune microenvironment differ fundamentally from humans. Large-animal studies in pigs and dogs have confirmed skin and lung protection, but sample sizes are limited and long-term follow-up is lacking. In addition, definitions of average, pulse, and instantaneous dose rate remain inconsistent, and whether tumor control is compromised by normal tissue sparing is still contested. These gaps make investigator-initiated trials the pivotal link in the evidence chain.

Step 1: Lock the evidence boundary with dual-track retrieval

Before designing a FLASH IIT, researchers should systematically search ClinicalTrials.gov, ChiCTR, the EU Clinical Trials Register, and UMIN-CTR using terms such as FLASH radiotherapy, ultra-high dose rate, and investigator-initiated trial, while defining the boundary of mechanistic and safety evidence from 2019-2025 peer-reviewed literature. The AI guideline retrieval and literature evidence capabilities of QSevidence map registry entries, journal articles, and conference abstracts into a unified evidence table, shortening the path from question to evidence landscape.

Step 2: Identify design elements through evidence grading

Each piece of evidence should be annotated with study type, sample size, dose-rate parameters, and endpoint definitions, and scored for risk of bias such as single-arm versus randomized design and comparability of historical controls. Structured evidence grading helps teams judge which conclusions can support primary endpoint selection and which still require preclinical reinforcement.

Step 3: Output IIT protocol elements in a structured framework

Finally, integrate mechanistic hypotheses, physical feasibility, and regulatory requirements into a protocol framework that specifies dose-rate definitions, endpoint systems, patient selection criteria, and sample size assumptions, producing text ready for ethics committee review and registry submission.

2. Physical Delivery Routes and the Dose-Rate Standardization Debate

The physical premise of FLASH radiotherapy is delivering an average dose rate above 40 Gy/s to the target. Each radiation type faces fundamentally different technical challenges.

Electron FLASH is the most mature route. Modified linear accelerators such as Oriatron eRT6 and FLASHKNiFE bypass or remove beam-flattening systems and raise pulse current, delivering 5-10 Gy in a 1-4 microsecond pulse, with instantaneous dose rate on the order of 10^6 Gy/s and average dose rate above 100 Gy/s. However, electron range is short, with penetration typically under 5 cm, so the approach suits superficial lesions. Proton FLASH exploits the Bragg peak to keep entrance dose low and concentrate dose in the tumor, but beam delivery is the bottleneck: pencil beam scanning has magnet response times near 1 ms and energy switching of 1-2 seconds, far longer than a FLASH pulse, making true whole-target FLASH difficult; a cyclotron achieves an average dose rate of about 30-60 Gy/s at isocenter. Photon (X-ray) FLASH faces an inherent conflict between target dose rate and penetration depth. Carbon nanotube cathodes and laser plasma accelerators are the two frontier routes; the latter can reach 10^12 Gy/s instantaneously but offers pulse repetition below 1 Hz, unstable spectra, and beam pointing jitter of about 5 mrad.

Dose-rate standardization is one of the most prominent current debates. Average dose rate (total dose divided by total irradiation time), pulse dose rate (pulse dose divided by pulse width), and instantaneous dose rate (peak during beam-on) can differ by orders of magnitude, and there is no consensus on which parameter drives the FLASH effect. Available evidence suggests normal lung protection correlates more strongly with pulse dose rate, whereas tumor control equivalence may depend more on average dose rate. The IAEA is promoting full parameter reporting, including average dose rate, pulse dose rate, pulse repetition frequency, dose per pulse, and total irradiation time, and is revising TRS-483 to include dedicated measurement protocols for ultra-high dose rate beams using ion recombination-corrected chambers and real-time scintillator monitoring. For multicenter IITs, a unified reporting format is the prerequisite for comparability.

3. Biological Mechanisms: Oxygen Depletion, ROS Recombination, and Immune Remodeling

Step 1: Quantitatively validating the oxygen depletion hypothesis

The oxygen depletion hypothesis holds that ultra-high dose rate irradiation consumes molecular oxygen in the field within milliseconds, so normal tissue lacks oxygen during the damage fixation stage and sustains less radiation injury. Radiation chemistry models predict that above 40 Gy/s the reaction rate of radicals with oxygen exceeds the rate of oxygen diffusion resupply, dropping local oxygen tension from physiological levels (roughly 20-40 mmHg) toward zero within milliseconds. Oxygen probe imaging, including electron paramagnetic resonance oximetry and phosphorescence lifetime imaging, observed that after FLASH irradiation normal skin oxygen tension fell by more than 80% within 100 ms, whereas conventional dose rate irradiation reduced it by only about 20%. Three controversies remain: the oxygen depletion rate required to trigger protection differs across tissues; some measurements show tumor pO2 falling in parallel, which conflicts with tumor control equivalence; and in vitro experiments under anoxia still show a FLASH effect, implying oxygen-independent mechanisms.

Step 2: Comparing ROS kinetics and DNA damage repair

At conventional dose rate, hydroxyl radicals and hydrated electrons have ample time to react with oxygen, generating superoxide and hydrogen peroxide (H2O2), and long-lived ROS diffuse to the nucleus to cause indirect DNA double-strand breaks (DSBs). Under FLASH conditions radicals are densely packed, so radical-radical recombination such as hydroxyl plus hydroxyl to H2O2 becomes more probable and long-lived ROS yield falls. Experiments show that after FLASH irradiation the number of gamma-H2AX foci, a DSB marker, in normal tissue cells is 30-50% lower than at conventional dose rate, with faster repair kinetics, whereas the difference narrows markedly in tumor cells, possibly because high glutathione levels partly offset the protection. The shift in H2O2 yield also engages downstream signaling: normal cells have more complete catalase and glutathione peroxidase systems that convert H2O2 into adaptive protective signals, while aberrant iron metabolism in tumors amplifies H2O2 through the Fenton reaction into highly toxic hydroxyl radicals, producing a directional difference that spares normal tissue while killing tumor cells.

Step 3: Assessing immune microenvironment remodeling

The impact of FLASH irradiation on the immune microenvironment underpins combination strategies. Preclinical evidence shows that, compared with conventional radiotherapy, FLASH irradiation causes less PD-L1 upregulation on tumor cells, tumor-infiltrating CD8-positive T cells skew toward an effector rather than exhausted phenotype (PD-1-positive, granzyme B-negative), dendritic cell activation is higher, and regulatory T cell infiltration is reduced. Because lymphocytes are highly radiosensitive, the instantaneous nature of FLASH shortens circulating lymphocyte exposure and preserves more effector T cells. These findings come mainly from small-animal models, however, and studies conflict on cytokine profiles such as IFN-gamma, TNF-alpha, IL-10, and TGF-beta, so validation in large animals and early IITs is required.

4. Global FLASH-IIT Landscape and Design Elements

As of 2025, at least 15 FLASH IITs have been launched worldwide, with distinct patterns in indications, device types, and geography. By indication, skin metastases and bone metastasis pain relief together account for about 40%, reflecting the maturity and low implementation threshold of electron devices for superficial treatment; lung cancer (early non-small cell lung cancer) and brain tumors (neurocognitive protection during whole-brain radiotherapy) account for about 20% each; head and neck and pelvic projects are fewer, reflecting the difficulty of deep delivery in maintaining penetration and dose rate. By device, electrons represent over 60%, protons about 25%, and photons only a few feasibility studies. Geographically, Europe launched the earliest projects through institutions such as Geneva University Hospitals (FAST-01), North America is dense in proton FLASH, and projects in Asia, including China and Japan, are growing quickly but concentrate on electron FLASH for skin and bone indications.

Step 1: Fix primary endpoints and toxicity assessment tools

Almost all IITs use acute toxicity as the primary endpoint, graded by CTCAE v5.0 with the LENT-SOMA scale for late effects, and patient-reported outcome (PRO) instruments such as EORTC QLQ-C30 to capture quality of life. Tumor control secondary endpoints vary by indication: skin cancer trials largely use RECIST 1.1, whereas bone metastasis trials emphasize pain relief rate and quality of life scores.

Step 2: Define patient selection and dose-rate reporting

Inclusion criteria typically require expected survival of at least 3 months, a Karnofsky performance status of at least 70, and lesion geometry compatible with the beam. Some brain tumor trials exclude prior whole-brain radiotherapy to reduce confounding. On dose rate, most trials enroll at an average dose rate of 40 Gy/s or higher, yet pulse and instantaneous dose rates are reported inconsistently, limiting cross-trial comparison.

Step 3: Evaluate statistical design and strength of evidence

Most current IITs are single-arm phase I or phase I/II designs with 10-30 patients; only a few use randomized controls. Differences between single-arm designs and historical controls, including data era, treatment technique, and patient selection criteria, can introduce selection and time-trend bias. Small samples widen efficacy confidence intervals and preclude meaningful subgroup analysis. These limitations point collectively to the need for multicenter, randomized, long-term follow-up designs.

5. Future IIT Directions

Step 1: Optimize dose and fractionation

The FLASH effect is most pronounced at larger single doses, while the interval between fractions may alter the magnitude of protection through the 4 Rs. A three-arm randomized trial is advisable: conventional fractionation (for example 30 x 2 Gy), single-session FLASH (8-12 Gy), and fractionated FLASH (2 x 6 Gy or 3 x 5 Gy), with grade 2 or higher acute toxicity by CTCAE v5.0 as the primary endpoint and late fibrosis, local control, and PRO as secondary endpoints. A dose-rate escalation trial should compare 40, 100, and 200 Gy/s within one indication to map the toxicity gradient, and the protocol should report the full dose-rate parameter set using IAEA-recommended protocols.

Step 2: Extend indications to high-risk populations

Pediatric tumors, re-irradiation, and oligometastases have the greatest translational potential. Pediatric trials must address late effects on neurocognition, endocrine function, and second primary cancer risk, and should include a dose-escalation safety lead-in with an independent data monitoring committee, with long-term follow-up of Wechsler intelligence scales, endocrine panels, and echocardiography. For re-irradiation, given that conventional re-irradiation carries a 15-30% risk of radionecrosis, single-session FLASH of 8-12 Gy or fractionated 3 x 5 Gy may be used, with radionecrosis incidence at 6 months confirmed by dynamic contrast-enhanced and perfusion MRI as the primary endpoint; prior dose distributions should be fused by deformable registration to build cumulative dose-volume histogram constraints. Oligometastatic disease may explore single-session stereotactic FLASH of 20-30 Gy, prioritizing lung, liver, and bone lesions, with 1-year local control non-inferiority as the primary endpoint.

Step 3: Clarify combination sequencing and biomarkers

Sequencing of immunotherapy is the most critical unknown. Preclinical work suggests giving immune checkpoint inhibitors 24-48 hours after FLASH may maximize synergy by reducing immunosuppressive cell infiltration in normal tissue, whereas concurrent dosing might blunt efficacy through immune cell protection. A randomized trial should compare, in PD-L1-positive (TPS at least 50%) non-small cell lung cancer or head and neck squamous cell carcinoma, three arms: FLASH plus pembrolizumab 24 hours after FLASH, FLASH plus concurrent pembrolizumab, and conventional radiotherapy plus pembrolizumab, with objective response rate by RECIST 1.1 as the primary endpoint and dynamic monitoring of peripheral PD-L1-positive CTLA-4-positive T cells, ctDNA clearance, and tumor-infiltrating lymphocyte density.

Step 4: Build multicenter standardization and registry research

Moving FLASH from IITs to confirmatory trials requires unified data collection standards for dose-rate definitions, toxicity assessment, PRO instruments, and imaging parameters, plus shared preclinical data and biobanks and a prospective real-world patient registry. Designing multicenter randomized controlled trials from positive IIT hypotheses is the practical path to accelerate evidence generation and regulatory dialogue.

6. Conclusion and Outlook

FLASH radiotherapy has moved from proof of concept into early clinical exploration. Existing IITs provide essential safety data and initial support for normal tissue sparing with equivalent tumor control, but short follow-up, small samples, and predominantly single-arm designs mean they are not yet sufficient to change clinical practice. Future research should prioritize dose and fractionation optimization, combination immunotherapy, pediatric tumors, re-irradiation, and oligometastases, underpinned by multicenter standardization networks and unified reporting. Throughout this process, rigorous evidence retrieval, grading, and structured synthesis are the prerequisites for high-quality decisions at the intersection of mechanistic hypotheses, physical feasibility, and regulatory requirements, which is precisely the value QSevidence delivers to radiation oncology researchers through AI guideline retrieval, literature evidence synthesis, and structured evidence generation.

References

1. Favaudon V, Caplier L, Monceau V, et al. Ultrahigh dose-rate FLASH irradiation increases the differential response between normal and tumor tissue in mice. Sci Transl Med. 2014;6(245):245ra93.
2. Bourhis J, Sozzi WJ, Jorge PG, et al. Treatment of a first patient with FLASH-radiotherapy. Radiother Oncol. 2019;139:18-22.
3. Gaide O, Herrera F, Sozzi W, et al. Comparison of ultra-high versus conventional dose rate radiotherapy in a patient with cutaneous lymphoma. Radiother Oncol. 2022;171:1-9.
4. Vozenin MC, Hendry JH, Limoli CL. Biological benefits of ultra-high dose rate FLASH radiotherapy: sleeping beauty awoken. Clin Oncol. 2019;31(7):407-415.
5. Montay-Gruel P, Acharya MM, Petersson K, et al. Long-term neurocognitive benefits of FLASH radiotherapy driven by reduced reactive oxygen species. Proc Natl Acad Sci USA. 2019;116(22):10943-10951.
6. Diffenderfer ES, Verginadis II, Kim MM, et al. Design, implementation, and in vivo validation of a novel proton FLASH radiation therapy system. Int J Radiat Oncol Biol Phys. 2020;106(2):440-448.
7. Wilson JD, Hammond EM, Higgins GS, Petersson K. Ultra-high dose rate (FLASH) radiotherapy: silver bullet or fool's gold? Front Oncol. 2020;9:1563.
8. Esplen N, Mendonca MS, Bazalova-Carter M. Physics and biology of ultrahigh dose-rate (FLASH) radiotherapy: a topical review. Phys Med Biol. 2020;65(23):23TR03.
9. Schuler E, Acharya M, Montay-Gruel P, et al. Ultra-high dose rate electron beams and the FLASH effect: from preclinical evidence to a new radiotherapy paradigm. Med Phys. 2022;49(3):e1-e22.
10. Lin B, Gao F, Yang Y, et al. FLASH radiotherapy: history and future. Front Oncol. 2021;11:644400.
11. FLASH radiotherapy: consensus report on dose-rate reporting and terminology. IAEA TRS-483 revision working group and FLASH Consortium, 2023.
12. QSevidence (qsevidence.com): AI guideline retrieval, literature evidence synthesis, and structured evidence generation tools.

Medical Disclaimer

This article is based on published preclinical research, early clinical trial registry records, and expert consensus, and is intended for medical education, research methodology, and clinical trial design reference only. It does not constitute any diagnostic or radiotherapy recommendation. FLASH radiotherapy described here remains under clinical investigation, and its safety, efficacy, indication scope, and device regulatory status are not yet finally established; the dose-rate parameters and fractionation schemes discussed must not be applied directly to patient treatment. Clinical and research decisions must be made by qualified radiation oncologists, medical physicists, and ethics review bodies in accordance with current regulations and individual circumstances.