TORPEdO
ForOropharyngeal SCC requiring concurrent chemo-RT with bilateral neck treatment
No difference vs IMRT
Mean scores similar at 3/12/24 mo post RT; no effect size reported in source
TL;DRNo mean UW-QoL physical composite difference IMPT vs IMRT at 3/12/24 mo post RT, 205 pts.
Both arms were prescribed the same 70 Gy / 56 Gy in 33 fractions under identical constraints, so this tests IMPT under photon-derived objectives, not its dosimetric ceiling. The physical composite (saliva, taste, chewing, swallowing) separates at no timepoint from week 6, weakening the QoL case for referring unselected bilateral-neck OPSCC pts to protons.
In oropharyngeal SCC needing bilateral-neck chemoradiotherapy, patient-reported QoL alone does not support a proton referral; it does not speak to unilateral-neck, RT-alone, or reirradiation pts, where the sparing case differs.
The referral decision is what moves: with identical prescriptions (70 Gy / 56 Gy in 33 fractions) and matched constraints, IMPT showed no mean UW-QoL physical composite advantage at any timepoint from week 6. That argues for model-based selection of individual patients over categorical proton referral in bilateral-neck OPSCC.
+1 more figure
9 details
Multicentre phase 3 RCT, 2:1 randomisation to IMPT vs IMRT, 205 pts recruited. Stratified by T-stage, N-stage, p16 status and smoking history. This ESTRO 2026 presentation reports the longitudinal HR-QoL analysis only.
Oropharyngeal SCC requiring concurrent chemo-radiotherapy including bilateral neck treatment. p16 status was a stratification factor, not an exclusion, so both HPV-driven and HPV-negative disease are represented.
70 Gy / 56 Gy in 33 fractions over 6.5 weeks in both arms, IMPT vs IMRT. Same dose, same schedule, same constraints, so delivery technique is the only variable.
Concurrent cisplatin 100mg/m2 on D1 and D22, identical in both arms. The systemic backbone is fixed, so nothing here reads on regimen choice.
Co-primary (clinician): CTCAE grade 3 weight loss (≥20% decrease from baseline) or gastrostomy dependence at 12 months post RT. Co-primary (patient): UW-QoL physical composite of saliva, taste, chewing, swallowing, appearance and speech at 12 months post CRT.
No difference in mean UW-QoL physical composite between arms at 3, 12 and 24 months post RT, with similar trajectories from week 6 post CRT and similar results across multiple PRO instruments. Scores fell at end of treatment then recovered, most stabilising from 12 months. No effect sizes given in source.
Non-randomised proton series in oropharynx have reported lower gastrostomy dependence and xerostomia than photon comparators, and that expectation is what this trial was built to test. The randomised patient-reported comparison does not reproduce a separation. No cross-trial numbers are in the source.
A composite of six domains dilutes a benefit confined to one, xerostomia being the obvious candidate. 90% CIs and no stated non-inferiority margin mean this is an absence of difference, not demonstrated equivalence. The commentary point that UK proton centres are early on the learning curve is untestable from the source.
The clean part of this design, matched dose and matched constraints, is also what bounds the answer: IMPT was planned to objectives written for photons, so the trial measures what protons deliver under photon rules rather than what they can achieve when pushed. It supports selecting patients by individual sparing benefit rather than referring bilateral-neck OPSCC to protons as a class.
Randomised phase 3 PRO co-primary shows no arm difference, supporting IMRT as standard. Clinician-reported co-primary and effect sizes absent from source.
- Clinician co-primary (weight loss / gastrostomy) result not yet reported
- Whether proton-experienced centres would show a QoL difference
- HR-QoL beyond 2 years; follow-up ongoing to 5 years
📚 Sources · 🐦 2 tweets · 📄 1 paper
Day FOUR of #ESTRO26 Coverage by OncoAlert 🚨
— OncoAlert (@OncoAlert) May 18, 2026
Health-related quality of life in the phase III trial of Toxicity Reduction using Proton Beam Therapy for Oropharyngeal Cancer (TORPEdO;CRUK/18/010) Presented by Matthew Tyler🇬🇧 #RadOnc ☢️
TORPEdO, a multicentre phase 3… pic.twitter.com/ZP6yK7RThL
TORPEdO. Misma planificación + constraints idénticas y centros UK noveles probablemente limitaron el potencial de #IMPT.
— Amadeo Wals (@AmadeoWals) May 18, 2026
Centros con alta experiencia se siguen viendo ventajas clínicas . La QA rigurosa del UK es una fortaleza, pero no maximiza la diferencia.#ESTRO26 #HNCSM https://t.co/rASp3QDIk1
The longer read
TORPEdO asks whether the dosimetric advantage of intensity modulated proton therapy converts into something an oropharyngeal cancer survivor can feel, and the patient-reported half of that answer is no separation. The structure of the trial is what makes the null informative and also what bounds it. Both arms were prescribed 70 Gy / 56 Gy in 33 fractions over 6.5 weeks with concurrent cisplatin on the same schedule, so delivery technique is the only variable. That is a clean modality question. It also means IMPT was held to planning objectives written for photons, and the commentary from the meeting floor made exactly this point: when constraints are matched rather than pushed, the proton plan is permitted to be better on paper but never asked to be, and the mean dose reduction that would drive a salivary or swallowing signal may not be realised at all.
The endpoint choice matters as much as the technique. The UW-QoL physical composite pools saliva, taste, chewing, swallowing, appearance and speech into a single score. A composite is efficient for a 205-patient trial and it dilutes: an improvement confined to one domain, xerostomia being the obvious candidate for a proton effect, is averaged against five domains protons have no particular reason to change. A domain-level analysis could still show what the composite cannot. That said, the presentation reported similar outcomes across several PRO instruments rather than this one alone, which makes a large hidden domain effect less likely, though not a modest one.
There is also the question of what a negative patient-reported co-primary licenses. The trial was built as a toxicity-reduction study, and its clinician-reported co-primary, grade 3 weight loss or gastrostomy dependence at 12 months, was not part of this presentation. A trial can miss on patient-reported physical function and still move feeding-tube dependence; those endpoints are not obliged to agree. Until that co-primary reads out, the defensible summary is that protons did not change how these patients said they felt, not that protons did not reduce toxicity.
Two further caveats shape how much confidence to move. Nothing in the source states a non-inferiority margin, and the intervals shown are 90%, so this is an absence of demonstrated difference in a modestly sized trial rather than established equivalence. And the UK proton programme is young; the trial's quality assurance rigour cuts both ways, guaranteeing acceptable plans while standardising away the centre-level variation experienced programmes use to their advantage.
For a radiation oncologist deciding whom to refer, the practical read is against unselected referral of bilateral-neck oropharyngeal patients and in favour of a model-based, individual-benefit approach: identify the patients whose photon plan genuinely cannot spare the structures that matter, rather than treating protons as categorically gentler. Follow-up runs to 5 years, and late endpoints are where a modality difference in this disease would be expected to appear if it exists.