Fractionated Electron FLASH RT: Feasibility and Skin Toxicity in a Porcine Model

Investigators evaluating fractionated electron FLASH radiotherapy report that a standard five-fraction workflow could be implemented in a porcine skin model and compared directly with matched conventional dose-rate electron irradiation, while tracking delivery performance and longitudinal cutaneous injury. In three Yorkshire–Landrace swine, paired FLASH and conventional fields were treated using the same planning and setup approach. The authors outline the practical steps used to deliver ultra-high dose-rate electrons and describe a dose-dependent dermatitis pattern over follow-up. Across the tested regimens, they report technical deliverability with independent verification streams and state that a FLASH-specific sparing effect was not detected under these fractionated conditions. Overall, the report focuses on (1) delivery/quality assurance in a clinically simulated workflow and (2) skin-toxicity outcomes used to compare modalities.

The five-fraction regimen was implemented in three Yorkshire–Landrace swine (43–49 kg at the start of radiation), with six dorsolateral flank fields per animal spanning modality, dose level, and field size combinations. FLASH and conventional irradiations were paired within animals at three fractionation regimens—5 × 6 Gy, 5 × 7 Gy, and 5 × 8 Gy—using circular field diameters of 4, 7, or 10 cm; fractions were delivered over 11 days at 48–72 h intervals. Radiation was delivered with a FLASH-capable Mobetron using 9-MeV electrons at 43.7 cm source-to-surface distance with a 5 cm air gap, lead skin collimation, and 1 cm tissue-equivalent bolus. For FLASH delivery, the article describes four pulses at 90 Hz (total irradiation time 33 ms), with pulse width adjusted to achieve dose at the depth of maximum dose, and reports mean dose rates of 175–246 Gy/s; conventional delivery was described at 8 Gy/min with 30 Hz and nominal 1.2 µs pulse width. Animals were anesthetized with tiletamine–zolazepam and xylazine induction and maintained with isoflurane under 100% oxygen after intubation, with identical peri-procedural conditions for both modalities. The methods are presented as a controlled delivery setup intended to isolate dose rate while holding other conditions constant.

Dosimetry validation and monitoring were described as a multi-instrument approach spanning pre-treatment verification, daily QA, and in vivo measurements during fractionated delivery. For the largest field size and each dose level, the authors report triplicate measurements with GafChromic EBT3 film, thermoluminescent dosimeters, and alanine pellets in the prescribed geometry at the depth of maximum dose, while smaller-field output factors were checked with film. Daily machine and patient QA were performed with a cross-calibrated ionization chamber, and in vivo dose validation used EBT3 film at the skin surface extrapolated to prescription depth via percentage depth dose ratios. Real-time FLASH output tracking was additionally performed via beam current transformer monitoring, and the authors report that prescribed and delivered doses agreed within 3% across the monitored comparisons, alongside matched percentage depth dose curves and dose profiles between FLASH and conventional beams. In aggregate, the study describes convergence across independent measurement streams while delivering a five-fraction regimen in vivo.

Cutaneous toxicity findings were presented as dose dependent across the three regimens, with the authors distinguishing tolerated courses from reactions they characterize as unacceptable at the highest dose level. The article reports that 5 × 6 Gy and 5 × 7 Gy fields reached no more than grade 2 dermatitis and were described as fully resolved by the end of the 22–24 week follow-up; paired analyses showed no statistically significant differences between FLASH and conventional irradiation for peak dermatitis, erythema index, or assessed histologic endpoints at these dose levels. At 5 × 8 Gy, both modalities were associated with more severe reactions, including patchy to confluent moist desquamation; one conventional field progressed to necrosis, and some residual clinical toxicity persisted at study end. Over time, initial dermatitis signs were reported within weeks after the last fraction, with peak dermatitis occurring later as dose increased; improvement also began later at higher dose. The authors note that field size did not significantly affect peak dermatitis or maximum erythema index in their analyses. In discussion, they frame the findings as showing no detectable FLASH sparing effect under the tested fractionated conditions and describe peri-procedural hyperoxia and other physical/biologic conditions as potential modifiers of the FLASH effect. The authors interpret the results as supporting the technical feasibility of fractionated electron FLASH delivery in this setting and state that further work is needed to optimize delivery parameters and better define conditions under which a normal-tissue sparing effect might be observed. They also note that peri-procedural conditions (including oxygenation) and the limited follow-up duration in this model could influence whether a FLASH-specific sparing effect is detected and whether later toxicity is observed.

Key Takeaways:

• The authors report that five-fraction electron FLASH delivery in a porcine skin model was operationally achievable with a clinically simulated workflow, alongside described QA and real-time monitoring infrastructure.
• Skin toxicity was reported to increase with dose across 5 × 6, 5 × 7, and 5 × 8 Gy, with no statistically significant FLASH–conventional differences at the lower two regimens and severe reactions observed at the highest regimen in both modalities.
• The authors note that peri-procedural oxygenation (their animals received 100% oxygen) could have influenced whether a FLASH sparing effect was observed, and that the study’s follow-up duration may have limited detection of later effects such as fibrosis or atrophy.