Randomised Gains in the Mouth, FLASH Near the Heart, and the Gantry-less Room: Particle Therapy in September 2026

September delivers what proton therapy's critics have long asked for: a randomised trial in which protons improved a clinically graded endpoint. The setting — unilateral postoperative head and neck cancer — is narrow, and the benefit is in acute rather than late toxicity, but it is a genuine randomised result. Around it, the month's literature extends FLASH into a more demanding anatomical setting, adds matched carbon-ion data in lung cancer, and gives the most coherent account yet of how upright, gantry-less treatment rooms could make particle therapy cheaper to build.

A randomised result in head and neck cancer

In a multicentre randomised phase II trial (NCT02923570), adults with resected, well-lateralised head and neck cancers requiring unilateral adjuvant radiotherapy were assigned 1:1 to protons or IMRT, to 60 Gy (RBE) in 30 fractions with elective regions treated to 54 Gy. Ninety-eight patients were analysable (54 proton, 44 IMRT), with a median follow-up of 27.2 months. The primary endpoint, clinician-reported acute grade 2 or higher oral mucositis within three months, occurred in 7.4% of proton patients against 22.7% with IMRT. Acute grade 2 or higher dysgeusia fell from 31.8% to 9.3%. Rates of grade 3 or higher dermatitis and grade 2 or higher dysphagia were similar, late toxicities were uncommon with no grade 3 events, and three-year local control, progression-free and overall survival were high and indistinguishable between arms.

The trial's caveat is in patient-reported outcomes, which mirrored the clinician-graded findings but with wide confidence intervals and differences that did not reach significance. Even so, this is the kind of evidence payers have demanded: a randomised comparison, in a defined indication, in which protons reduced a toxicity patients feel during treatment without compromising cancer control.

FLASH reaches the chest wall

FAST-01 showed that proton FLASH was feasible for painful bone metastases in the limbs. FAST-02 moved closer to critical organs. In this single-arm trial, adults with one to three painful non-vertebral thoracic bone metastases received a single 8 Gy fraction at 40 Gy/s or more. Twelve patients were enrolled and ten treated, each at one site in a rib or the scapula. All completed treatment as planned, with an average time on the table of 14.6 minutes. At a median follow-up of 189 days there were no grade 2 or higher acute or late adverse events at least possibly related to FLASH. Six patients had a complete pain response, three a partial response and one stable pain, although three reported pain flares by day 11. With ten patients and no control arm, FAST-02 cannot establish a FLASH sparing effect; what it establishes is that the workflow can be run safely near the lungs and heart, which is the necessary next step.

Carbon ions in lung cancer, and rare tumours

A propensity-matched retrospective study compared definitive carbon-ion and photon radiotherapy in 292 patients (146 per group) with stage III non-small cell lung cancer treated as part of multimodality therapy between 2016 and 2023. At a median follow-up of 32.5 months, carbon ions gave significantly longer median locoregional recurrence-free survival (57.7 versus 25.0 months; HR 0.59) and progression-free survival (17.7 versus 16.7 months; HR 0.70), with fewer acute grade 2 or higher radiation-related adverse events. Overall survival and distant metastasis-free survival were similar. The authors are clear that only a randomised trial can confirm the locoregional advantage, but the data strengthen the case for one.

For rare sarcomas, the Japan Carbon-Ion Radiation Oncology Study Group reported 10 patients with unresectable intermediate-category solitary fibrous tumour treated with 64.0–70.4 Gy in 16 fractions. At a median follow-up of 49 months there were no local recurrences, four-year overall survival and local control were both 100%, and progression-free survival was 67%, with two patients developing distant metastases. There was no grade 3 or higher toxicity, and quality of life was maintained. A systematic review of 32 studies on locally recurrent rectal cancer, meanwhile, concluded that SBRT and particle therapy allow substantially higher doses without more toxicity in unresectable or inoperable recurrences, with SBRT suited to small recurrences and particle therapy to larger ones and complex shapes spanning several anatomical compartments.

Breast without surgery, and BNCT re-irradiation

A small prospective study tested partial-breast proton therapy without lumpectomy, at 62.4 Gy (RBE) in 26 fractions, for women aged 40–70 with cT1N0M0, oestrogen-receptor-positive, HER2-negative invasive ductal carcinoma. Eleven patients had completed at least five years of follow-up, with a median of 82 months. Tumours shrank promptly after treatment, and no in-field recurrence was observed in patients followed for more than five years. Dermatitis was grade 1 in 55% and grade 2 in 45%, with none at grade 3 or higher. The authors present this as preliminary, but the idea of a radiation-only alternative to partial mastectomy for carefully defined low-risk disease is striking.

Re-irradiation of recurrent oral squamous cell carcinoma with accelerator-based BNCT produced an objective response rate of 80% in 45 patients. Two-year overall survival was 51.1%, but local control was 31.8% and progression-free survival 22.2%. Grade 3 or higher acute mucositis affected 49%, and late grade 3 or higher osteoradionecrosis 10%. Achieving a complete response within 90 days was associated with better survival, and tumour volume warranted further study for risk stratification. BNCT is an option for selected patients, the authors conclude, but local control remains its weak point.

Better plans and safer delivery

Technical progress this month centres on delivering more sophisticated plans efficiently and verifying them. For spot-scanning proton arc therapy, all existing optimisation algorithms were designed for cyclotron systems. A new dynamic-programming algorithm for synchrotron systems merges adjacent energy layers while preserving plan quality; across liver, head and neck, lung and prostate cases it reduced acceleration cycles from an average of 154.75 to 54.25 and saved 210.47 seconds of delivery time, a relative reduction of 57.37%.

In oesophageal cancer, a retrospective analysis of 192 patients treated with IMPT, 85 of whom were replanned, examined adaptive replanning triggered by target underdose or cardiac overdose. Using deformable dose accumulation, adaptation increased ITV D98% by 1.9 percentage points in the underdose subgroup, and in the cardiac overdose subgroup reduced mean heart dose by 1.0 Gy and modelled two-year mortality risk by 1.5 percentage points, without raising lung dose.

The ESTRO-EPTN (European Particle Therapy Network) published consensus guidelines on independent dose calculation and log file-based patient-specific quality assurance in scanned particle therapy, presenting the two as complementary and, when validated, able to substantially reduce reliance on phantom measurements — with minimum requirements for vendors on log files. A PTCOG-led panel of 15 radiation oncologists produced international consensus contouring guidelines for skull base chordoma; agreement was highest for preoperative gross tumour volume (mean Dice 0.85) and lowest for residual disease (mean Dice 0.45), and the panel set out recommended intermediate- and high-dose CTV definitions.

The case for the gantry-less room

The most strategic discussion of the month concerns upright treatment. An ESTRO physics workshop of 22 participants — physicists, a radiation oncologist, academics and industry representatives — concluded that gantry-less radiotherapy with upright positioning holds significant promise for simplifying delivery, improving access and reducing cost, but that imaging and immobilisation remain crucial challenges, upright patient data are scarce, and clinical and technical evidence is the most critical gap. The group published a priority-ranked wish-list for vendors.

An editorial in the British Journal of Radiology assessed upright particle therapy against Thomas Kuhn's criteria for scientific change, arguing that removing rotating gantries should reduce mechanical complexity, weight, maintenance and capital, construction and operating costs, while preliminary studies suggest suitable setup reproducibility and comparable plan quality. A pilot study in 11 healthy volunteers adds a practical piece: real-time optical respiratory guidance improved breathing-task performance and regularity, and posture changed how breathing was organised, with greater abdominal dominance supine than upright.

What September adds up to

A randomised trial shows protons reducing a clinically graded toxicity in a defined head and neck setting. FAST-02 shows FLASH can be delivered near the thorax's critical organs, and matched carbon-ion data in lung cancer justify randomised testing. The upright consensus points to a future in which particle therapy is cheaper to build. The next step is the same in every case: larger prospective trials that turn these signals into practice-changing evidence.