Impact of Bladder Filling on an MR-Only Pelvis Radiotherapy Workflow
Morris E, Moncion M, Sevak P, Weiss S, Garbarino K, Renisch S, and Glide-Hurst C. Impact of bladder filling on an MR-only pelvis radiotherapy workflow. Med Phys 2017; 44(6):2770.
Purpose: As MR-SIM emerges as a primary treatment planning modality for prostate cancer, it becomes critical to quantify the uncertainties introduced for MR-only planning. This work characterized temporal intra-fractional changes between the prostate, seminal vesicles (SVs), and organs at risk (OAR) in response to bladder filling conditions, including geometric, dosimetric, and local deformation assessments. Methods: Serial T2- weighted and mDIXON sequences (3-6 timepoints/subject, 1.5T and 3.0T, 25 evaluable timepoints) were acquired in 6 subjects using a fixed bladder filling protocol (bladder void, 20 oz water consumed pre-imaging, 10 oz mid-session). Population-based synthetic CTs were generated by combining modelbased bony segmentation with voxel classification (muscle, adipose, spongy, and compact bone) of mDIXON. Rectal air was assigned to -350 HU. Delineations for each timepoint were performed by one physician including prostate, SVs, bladder, rectum, and penile bulb. VMAT treatment plans were derived on the fullest bladder (reference) to meet QUANTEC criteria. Dose was recalculated using fixed monitor units on other timepoints and dosimetric differences were quantified. Local deformations of the prostate and SVs were assessed using multiresolution deformable image registration using T2 images. Results: In an empty bladder condition, PTV coverage (volume receiving 95%) reduced significantly for all cases (13.0 ± 8.2%, range: 4.3- 27.2%), likely due to shifts of prostate/SVs relative to full bladder conditions. Rectal dose differences were patient-specific; empty bladders tended to decrease most rectal metrics although QUANTEC criteria were still met for all but 1 case. Total vector displacements analyzed via displacement vector fields revealed that V90% wasprostates, with local effects observed near the bladder/prostate interface. Conclusion: For MR-only planning in the pelvis, bladder status may significantly impact target and OAR locations, dosimetry, and target deformations. Given long MRI acquisition times, proper management of filling status will be an important consideration in MR-only workflows.