Fewer Fractions, Second Courses and the RBE Problem: Particle Therapy in July 2026

Particle therapy has always been sold on physics. July's literature is a reminder that its future will be decided by biology, fractionation and organisation. The month brings long-term prostate data that make the case for shorter proton courses, prospective imaging evidence that the proton RBE is not a constant, consensus on how carbon-ion doses should be translated between centres, and an unusually frank look at the limits of treating the same tissue twice. Behind it all, two reviews ask whether the field can cost its treatments consistently and deliver them fairly.

Prostate: the case for shorter courses

A phase II trial (PPS-002) enrolled 125 men with cT1c–3aN0M0 prostate cancer and treated them with image-guided proton therapy to 63 Gy(RBE) in 21 fractions. At a median follow-up of 71.0 months, the 5-year cumulative incidence of late grade 2 or higher gastrointestinal adverse events was 3.3%, and of genitourinary events 8.1%, meeting the primary endpoint. Biochemical recurrence occurred in three patients and no clinical recurrences were recorded, giving 5-year biochemical recurrence-free survival of 97.6% and overall survival of 97.5%.

The result arrives as the German Society of Radiation Oncology (DEGRO) prostate working group updated its statement on ultrahypofractionation, drawing on phase III randomised trials with at least five years of follow-up: HYPO-RT-PC, PACE-B and a proton therapy trial. All showed non-inferior disease control. The panel also notes the costs: late urinary toxicity in PACE-B was higher with SBRT (26.9% versus 18.3%), and proton SBRT showed slightly increased toxicity on the sexual and urinary EPIC domains. Its advice is that ultrahypofractionation is an effective alternative in low- to intermediate-risk disease, provided patients — particularly those with poor urinary function — are told about the increased genitourinary risk. Taken together, the two papers suggest proton prostate therapy is converging on the same shorter schedules as photons, with toxicity rather than tumour control the deciding variable.

RBE is not a constant

The most scientifically important paper of the month is a prospective correlative study nested in a randomised breast cancer trial comparing protons and photons for regional nodal irradiation. Serial CT scans of 71 patients with complete follow-up (36 proton, 35 photon) were analysed voxel by voxel for lung density change as a function of dose. Proton therapy produced significantly lower TD₅₀ values at both follow-up visits — that is, lung tissue changed at lower doses. In the overall population, TD₅₀ was 34.5 Gy(RBE) for protons against 43.6 Gy for photons at the first follow-up, an RBE estimate of 1.27, falling to 1.24 at the second. In right-sided free-breathing patients the first-follow-up estimate was 1.35.

Crucially, none of this reached the clinic. No patient developed clinically relevant pneumonitis or fibrosis, cough and dyspnoea rates were similar between arms, and density changes did not correlate with symptoms. The estimates nonetheless sit above the fixed RBE of 1.1 used in clinical planning, and the authors argue for continued evaluation of LET- and RBE-based optimisation.

For carbon ions, where RBE is the rationale rather than the problem, an international group published consensus statements on how RBE is handled clinically. Different biophysical models make different RBE predictions, so RBE-weighted doses reported by different centres are not directly comparable — an obstacle to shared protocols, multi-institutional trials and meta-analyses. The authors set out how RBE is currently accounted for and review approaches for translating doses between prescription systems, concluding that transforming plans between any two systems is feasible.

Lung cancer: dosimetry without a quality-of-life dividend

A cohort of patients considered for protons through model-based selection asked whether patients selected for intensity-modulated proton therapy actually feel better. Among 171 IMPT and 71 IMRT patients, mostly with stage III non-small cell lung cancer receiving chemotherapy, quality of life was measured with EQ-5D-5L, EQ-VAS and the EORTC QLQ-C30 global health status. After propensity weighting, differences between modalities were neither statistically significant nor clinically relevant. Mean follow-up was eight months. The authors point out why this should not surprise: patients are selected for protons on predicted heart dose and two-year mortality, not on expected quality-of-life gains.

Treating the same tissue twice

Re-irradiation of recurrent head and neck squamous cell carcinoma received international consensus recommendations in The Lancet Oncology, endorsed by the Reirradiation Collaborative Group, the ESTRO Reirradiation Focus Group and ASTRO. Developed through a structured process with formal agreement assessment, the document covers patient selection, imaging, target delineation, planning, dose accumulation and toxicity management, and names proton therapy among the modern techniques that have widened the feasibility of re-treatment.

Clinical data from a prospective single-institution registry show what proton re-irradiation can offer. Sixty-five patients with recurrent or second primary head and neck cancer received pencil-beam scanning IMPT, with a median dose of 66 Gy(RBE) in 30 fractions. Median overall survival was 17.2 months, with a 12-month rate of 67.1%; median progression-free survival was 10.1 months. Grade 3 or higher oral mucositis occurred in 7.7%, with no grade 4–5 acute events, and late osteoradionecrosis affected 6.2% — though median follow-up was only 9.3 months. Ninety-two per cent completed treatment. On multivariable analysis, T3–T4 stage was the only variable that retained significance for progression-free survival.

A second course of boron neutron capture therapy tells a harder story. Thirty-three patients with recurrent head and neck cancer at the Southern TOHOKU BNCT Research Center received BNCT twice, a median of 4.1 months apart. Response rates were 81.8% to the first course and 70% to the second, and median survival after the second course was 33.6 months. But local control after the second course was 38.3% at one year and 22.3% at two, and late grade 3 or higher toxicity occurred in 60.6% of patients, with one grade 5 aspiration pneumonia. The authors conclude that a second BNCT course offers limited benefit at a high cost in late toxicity and should be reserved for carefully selected patients without alternatives.

Carbon ions: sequencing and simplification

A single-centre series of 41 patients with hepatocellular carcinoma explored how carbon-ion radiotherapy fits alongside modern immunotherapy (atezolizumab plus bevacizumab, or durvalumab plus tremelimumab). When immunotherapy was used for recurrence after carbon ions, 14 of 20 patients responded, responders having low tumour burden and preserved liver function. When curative-intent carbon-ion therapy followed immunotherapy, 4 of 10 patients achieved a clinical complete response and became drug-free, all of them with solitary intrahepatic tumours. Carbon ions used to mop up residual lesions after an insufficient immunotherapy response, by contrast, rarely converted to complete response.

On the engineering side, a dosimetric study of 12 volunteers scanned supine and upright found that upright carbon-ion plans for prostate cancer matched supine plans: mean absolute differences in five CTV dose metrics were all under 0.15 Gy, 0.29% of the 51.6 Gy prescription, with equivalent organ sparing and robustness. The authors note the potential for reduced installation costs and simpler platform management.

The geometric case for gantries is evolving too. A scoping review of 74 studies on particle arc therapy found maintained target coverage, a median 6.3% relative improvement in conformity index and a median 10.5% relative reduction in integral dose against conventional particle techniques, sometimes at the cost of a larger low-dose bath. Carbon-ion arc studies reported redistribution of high LET from healthy tissue into the tumour. The evidence, however, is almost entirely in silico.

Costs and access

A systematic review of particle therapy costing studies published between 2000 and 2024 found just 12, with methods and assumptions so heterogeneous that cost estimates varied widely, even though quality scores were high. The authors propose costing recommendations aligned with accounting principles as a basis for standardised guidelines — a prerequisite for credible reimbursement decisions.

Access is the other half of the economic question. A scoping review of 41 studies on paediatric radiotherapy infrastructure found that proton therapy, along with IMRT and VMAT, remains largely concentrated in high-income countries, while many low- and middle-income countries rely on limited equipment, with Morocco presented as an illustrative case.

What July adds up to

The month's evidence pulls in two directions. Shorter proton courses for prostate cancer, upright carbon delivery and arc techniques promise a field that is cheaper and simpler to run. At the same time, in-patient RBE measurements, divergent carbon-ion models and high late toxicity after repeat BNCT show how much remains biologically uncertain. Robust costing standards and honest re-irradiation guidance are the tools that will let payers and clinicians weigh those two directions sensibly.