Head and neck MR-Linac use remains limited despite widespread access in specialist survey
Among 36 MR-Linac consortium experts, 86% reported access to the technology, but only 28% treated head and neck cancer on the platform.
Among 36 MR-Linac consortium experts, 86% reported access to the technology, but only 28% treated head and neck cancer on the platform.
Gadoxetate-enhanced MRI at 20–30 minutes improved liver tumor visibility and increased interobserver contour agreement from Dice 0.84 to 0.92.
In 50 pancreatic SBRT fractions, expert recontouring identified duodenal dose-limit exceedances in two patients despite largely preserved target coverage.
After 50 Gy in five fractions, local progression was 15% at one year, while distant progression remained the dominant failure pattern.
Per-fraction MR radiomics identified distinct tumor and healthy-liver trajectories during SBRT that were not explained by volume or dose alone.
A deep-learning model ordered first-to-last fraction images with 95% accuracy, while smaller longitudinal changes were associated with later biochemical recurrence.
The 20-patient ARIA study will test 5×8.5 Gy MR-guided adaptive RT with gating for cT1–2N0M0 glottic cancer.
Among 20 mostly metastatic GEP-NET patients, MR-guided ablative RT achieved 95% local control with acute grade 3 toxicity in 10%.
Accumulated gastrointestinal doses were generally lower than planned, while higher small-bowel dose showed exploratory associations with diarrhoea after MR-guided SABR.
Median online adaptive MR-guided treatment time fell 28% despite increasing complexity, with 78% of late-era fractions completed within 60 minutes.
Calibrated uncertainty thresholds kept gamma-defined dose-modelling failure below 10% while accepting 45–77% of irradiation segments.
Hydrogel spacers showed no meaningful loss of geometry during prostate SBRT, supporting stable rectal separation across the treatment course.
A hybrid deep learning–registration workflow reduced median daily prostate MRgRT contouring time from 7.8 to 3.1 minutes while maintaining high contour agreement.
ETLD tracked 106,000 cine-MRI frames with submillimetre error, while integrated segmentation achieved a mean global Dice score of 87.7%.