KEY POINTS
- This single-center study from The Royal Marsden evaluated 10 patients with muscle-invasive bladder cancer receiving whole-bladder RT on an Elekta Unity MR-linac. Treatment schedules were 55 Gy in 20 daily fractions or 36 Gy in six weekly fractions, with all patients starting treatment using 5-mm isotropic PTV margins and an empty-bladder preparation protocol.
- Daily treatment used an MR-guided adapt-to-shape workflow with online contour modification and plan optimization. Additional MRI was acquired during treatment to coincide with delivery completion, and the adaptive plan was retrospectively recalculated on this post-treatment anatomy; acceptable coverage required CTV V95% ≥95%, with GTV coverage assessed when that objective was missed.
- A total of 138 fractions were included, with 119 evaluable post-treatment MRI datasets available for dose analysis. Median workflow time from daily MRI to post-treatment MRI was 25.6 minutes (IQR 20.8–30.7) without treatment interruptions and 26.6 minutes (IQR 21.0–34.1) when interruptions were included.
- Before any margin personalization, the standard 5-mm PTV achieved acceptable CTV coverage in 86 of 91 fractions (95%). Coverage was satisfactory in 68/73 fractions (93%) for the 20-fraction schedules and 18/18 fractions (100%) for the six-fraction schedules, suggesting that small margins were feasible for most observed treatments.
- Across the full evaluable dataset, including fractions delivered after margin adaptation, CTV coverage was acceptable in 112/119 fractions (94%). All seven fractions failing the mandatory CTV criterion occurred in the 20-fraction cohort; GTV coverage remained acceptable in six of seven, while the remaining fraction involved multifocal disease and was conservatively classified as unacceptable.
- Three of ten patients required personalized margin expansion, applied across 31 treatment fractions. Two patients received an 8-mm superior margin, while one required 8-mm margins in most directions and 10 mm posteriorly; among evaluable fractions in these patients, satisfactory coverage was 18/20 (90%) before versus 26/28 (93%) after personalization, without evidence that the difference was statistically or causally attributable to expansion.
- Intrafractional bladder filling varied substantially, with median rates of 0.6 cc/min in the 20-fraction cohort versus 0.2 cc/min in the six-fraction cohort. Atypical bowel motion also caused repeated bladder displacement in one patient, requiring 18 adapt-to-position corrections; the analysis did not evaluate accumulated dose, reductions in OAR exposure, toxicity, or oncological outcomes.
CLINICAL TAKEAWAY
A 5-mm isotropic PTV margin appears technically feasible for many patients undergoing whole-bladder MR-guided online adaptive RT when repeated imaging, motion assessment, and selective margin expansion are available. However, clinically relevant undercoverage still occurred, and this small feasibility study cannot establish a universal margin recipe or demonstrate reduced toxicity without accumulated-dose and prospective outcome validation.