KEY POINTS
- This retrospective planning study reoptimized treatments from 10 patients with solitary recurrent glioblastoma previously treated using fractionated CyberKnife stereotactic radiotherapy. Target volumes ranged from 2.9 to 33.1 cm³, with regimens including 22 Gy in one fraction, 9 Gy × 3, 8 Gy × 4, and 5–6 Gy × 5.
- The experimental plans used a 250 MeV single-energy proton beam, three or four fields, a universal range shifter, patient-specific range compensators, and brass apertures to sharpen the lateral penumbra. Minimum spot weights of 300 or 400 monitor units were evaluated.
- Bragg peak FLASH produced a better conformity index than CyberKnife (1.4 ± 0.2 vs 1.9 ± 0.3; p<0.01) and reduced mean optic chiasm maximum dose from 25.2% to 17.1% of prescription dose (p<0.01).
- Dose gradients were not significantly different: GI50 was 2.9 ± 0.5 versus 2.7 ± 0.2 (p=0.09), and GI30 was 6.2 ± 1.8 versus 5.5 ± 0.5 (p=0.19). FLASH plans had a higher GTV maximum dose (146.6% vs 131.1%; p<0.01), reflecting reduced modulation and dose homogeneity.
- At 400 monitor units per spot and a five-Gy dose threshold, the volume receiving at least 40 Gy/s reached 85.8% ± 13.8% of normal brain and 95.8% ± 6.8% of the GTV. Coverage was lower when the two-Gy threshold was applied.
- Estimated delivery time was 0.47 ± 0.15 seconds per field at the higher monitor-unit setting, corresponding to approximately 1.41 seconds for a three-field plan, compared with 1,020 ± 584 seconds for CyberKnife delivery.
- No patient received proton FLASH, and no clinical toxicity or tumor-control outcomes were assessed. Modelled normal-tissue complication probabilities were similar, but the biological FLASH effect, aperture-related geometric uncertainty, deliverability, and performance in larger targets remain unvalidated.
CLINICAL TAKEAWAY
Single-energy Bragg peak FLASH could potentially provide highly conformal and exceptionally rapid proton reirradiation for small recurrent glioblastoma targets. This remains a technical proof of concept: neither the proposed normal-tissue FLASH effect nor clinical safety and efficacy have been demonstrated.