KEY POINTS
- This retrospective single-center study analyzed 775 head and neck reirradiation plans delivered on the same TrueBeam STx platform between 2015 and 2025. Treatment evolved from predominantly fixed-field IMRT in 2015–2019 to VMAT in 2020–2025, while Eclipse underwent a major upgrade from version 11.0.31 to 15.6.06.
- Because actual prescriptions differed between patients, absolute dose metrics were analytically normalized to a nominal 60 Gy in 30 fractions for longitudinal comparison. These values represent current-course planning metrics and should not be interpreted as cumulative EQD2 or composite reirradiation dose.
- Target coverage improved substantially over the decade: mean PTV D95 increased from 96.4% to 99.6%, D98 from 95.6% to 98.8%, while hotspot D2 decreased from 108.5% to 104.0%. All four primary target endpoints showed significant year effects after multiplicity adjustment.
- Dose fall-off became progressively steeper. GI20 decreased from 6.2 to 3.9, GI30 from 4.9 to 2.8, GI40 from 3.6 to 2.2, and GI50 from 3.0 to 1.8, with all temporal trends significant at p<0.001.
- OAR doses also fell markedly: brainstem near-maximum dose decreased from 18.7 to 7.2 Gy, spinal cord from 17.8 to 5.9 Gy, oral cavity mean dose from 31.2 to 15.9 Gy, and mean dose to dysphagia- and aspiration-related structures from 30.2 to 16.5 Gy.
- Pooled VMAT plans appeared substantially better than earlier IMRT plans: D95 was 98.8% versus 97.1%, D2 105.4% versus 107.6%, brainstem dose 12.7 versus 17.9 Gy, and spinal cord dose 9.7 versus 16.2 Gy. However, after adjustment for calendar year, the independent modality effect was attenuated for D95 and D2, and continued improvements were seen within both treatment eras.
- A major confounder was changing case selection: mean PTV volume fell from 700.2 cc to 250.4 cc, an approximately 64% reduction. Temporal associations for coverage, hotspot control and selected OAR doses persisted after adjustment for log-transformed PTV volume, but tumor site, prior dose, treatment interval, surgery, target geometry and cumulative dose could not be fully accounted for.
CLINICAL TAKEAWAY
Head and neck reirradiation planning became substantially more conformal over a decade, with better target coverage, lower hotspots, sharper gradients and markedly lower current-course OAR doses. These gains should not be attributed to VMAT alone: target volumes became much smaller while the planning system, optimization strategy, patient selection and institutional experience all evolved, and the study provides no evidence that the dosimetric improvements translated into better toxicity or disease-control outcomes.