Most European proton candidates still face critical access shortages
More than 56,000 theoretically eligible European patients lived in areas with no proton access or critical capacity shortages in 2025.
Latest research
Search the archiveInterpretable AI separated three-year anal cancer relapse risk after chemoradiotherapy
An explainable model separated three-year local relapse rates of 11% versus 39%, but lacks independent external validation.
Danish head and neck risk model generalized to an Australian radiotherapy cohort
The model retained moderate discrimination for recurrence and metastasis, but predicted non-cancer death less reliably, especially in older patients.
Calling it “radiation therapy” reduced treatment preference compared with “radiotherapy”
In a randomized US survey, “radiation therapy” reduced selection of the same treatment by 8.3 percentage points versus “radiotherapy.”
Day-before-surgery radiotherapy matched diclofenac for heterotopic ossification prevention
A single 7 Gy fraction delivered 20–28 hours before hip arthroplasty produced similar heterotopic ossification rates to postoperative diclofenac.
Retroverted uterus altered cervical brachytherapy dosimetry without worsening disease control
Retroverted uterine position increased rectal and bowel dose during cervical brachytherapy, but survival, recurrence and severe toxicity remained comparable.
Simulation-free CBCT workflow enabled urgent radiotherapy within about two hours
The OSCAR workflow delivered the first urgent or palliative fraction without planning CT in approximately 1 hour 50 minutes.
Real-world brain MR-only implementation exposed frequent synthetic CT artifacts
Synthetic CT artifacts were common during brain MR-only implementation, highlighting workflow QA and patient-selection issues not captured by controlled validation.
Commercial AI synthetic CT showed good soft-tissue agreement but important high-density failures
AI synthetic CT preserved soft-tissue HU and geometry well, while bone, air interfaces and implanted devices remained important failure modes.
Radiotherapy-specific nutrition screening tool showed high agreement with established assessments
NRS-RT showed 92.9% sensitivity and 97.2% specificity versus NRS-2002, but requires external validation before clinical adoption.
Twenty-two Gy SRS was associated with better brain metastasis control than 20 Gy
Single-fraction 22 Gy SRS achieved higher one-year local control than 20 Gy, although retrospective dose selection creates important confounding.
Carboplatin-paclitaxel showed comparable outcomes in cisplatin-ineligible head and neck cancer
Retrospective outcomes with carboplatin-paclitaxel were not significantly different from cisplatin, supporting its use when cisplatin is contraindicated.
Photon minibeams enable denser LATTICE patterns in small tumors
Mini-LATTICE created more compact high-dose vertices and greater peak-to-valley separation than conventional LATTICE across six planning cases.
AI-derived LAD dose reproduced survival stratification in RTOG 0617
Automated LAD contours reproduced the association between coronary dose and survival in 460 stage III NSCLC patients, although performance varied by model.
Simulation-free adaptive lung SBRT was feasible in the ONE STOP pilot
Eight of ten patients completed simulation-free single-fraction adaptive lung SBRT, with treatment taking a median 72 minutes.