MRI improved LLM-based recurrence prediction after definitive cervical chemoradiotherapy
Multimodal LLM prediction reached an AUC of 0.79, but AI-generated MRI reports showed clinically important discordances requiring expert review.
Multimodal LLM prediction reached an AUC of 0.79, but AI-generated MRI reports showed clinically important discordances requiring expert review.
Five proton centres outline why conventional radiation protection workflows cannot simply be assumed adequate for ultra-high dose-rate FLASH research.
A preclinical proton STAR workflow achieved benchmark-level target coverage, OAR sparing and cardiac gating latency below 30 ms.
In randomized i-STOP, adding SBRT to intermittent ADT reduced 3-year biochemical progression from 60% to 21% in PSMA PET-only oligometastatic prostate cancer. ADT re-initiation was also numerically less frequent, but not significantly.
In the randomized COMPPARE proton cohort, 60 Gy(RBE) in 20 fractions was noninferior to 78 Gy(RBE) in 39 fractions for 3-year bowel function, with similar grade 2+ GI toxicity and biochemical control.
Long-term NRG/RTOG 0815 shows that adding 6 months of ADT to dose-escalated RT reduces biochemical failure, metastases, and prostate cancer mortality, but still does not significantly improve overall survival.
In phase III GETUG P05, brachytherapy boost did not improve 5-year biochemical control over dose-escalated EBRT to 80 Gy in intermediate-risk prostate cancer and increased mild-to-moderate urinary toxicity.
In INTREPID, replacing 6 months of ADT with darolutamide during RT preserved erectile function but failed the prespecified noninferiority endpoint for early PSA suppression. Longer-term oncologic outcomes remain pending
Control, predictability, information and relational continuity emerged as key needs for making awake pediatric proton therapy more manageable.
Across 744 cervical oART fractions, RTT-led adaptation took a median 33 minutes while maintaining ≥95% target D98 in every fraction.
A model using target volume and HyTEC limits estimated automated SRS prescription doses with approximately 1.3% mean absolute error.
After proton reirradiation for recurrent pediatric ependymoma, 5-year survival was 64% with one grade 3 radiation injury and no grade 4+ toxicity.
After repeat spine SBRT, 12-month local failure was 6% with Dmin ≥28 Gy EQD2 versus 36% below that level.
In recurrent IDH-wildtype glioblastoma LMD, comprehensive RT and systemic therapy were independently associated with longer survival.
In randomized phase II I-SABR, four doses of nivolumab added to SABR improved 6-year OS to 69% vs 55% and EFS to 55% vs 32%. The long-term signal is compelling, but OS was a secondary endpoint in a relatively small phase II trial.