Interfractional Geometric Variations and Dosimetric Benefits of Stereotactic MRI Guided Online Adaptive Radiotherapy (SMART) of Prostate Bed after Radical Prostatectomy: Post-Hoc Analysis of a Phase II Trial

SIMPLE SUMMARY: Geometric variations of patients receiving stereotactic body radiotherapy (SBRT) after radical prostatectomy were evaluated in this study. With specific bladder filling and rectal emptying patient instructions, CTV shape and volume remain relatively stable and dose coverage is well maintained whilst maintaining OAR dose constraints in most treatment fractions. However, target coverage and OAR doses can be further improved by MR-guided online adaptive planning to account for interfractional geometric variations. ABSTRACT: Purpose: To evaluate geometric variations of patients receiving stereotactic body radiotherapy (SBRT) after radical prostatectomy and the dosimetric benefits of stereotactic MRI guided adaptive radiotherapy (SMART) to compensate for these variations. Materials/Methods: The CTV and OAR were contoured on 55 MRI setup scans of 11 patients treated with an MR-LINAC and enrolled in a phase II trial of post-prostatectomy SBRT. All patients followed institutional bladder and rectum preparation protocols and received five fractions of 6−6.8 Gy to the prostate bed. Interfractional changes in volume were calculated and shape deformation was quantified by the Dice similar coefficient (DSC). Changes in CTV-V95%, bladder and rectum maximum dose, V32.5Gy and V27.5Gy were predicted by recalculating the initial plan on daily MRI. SMART was retrospectively simulated if the predicted dose exceeded pre-set criteria. Results: The CTV volume and shape remained stable with a median volumetric change of 3.0% (IQR −3.0% to 11.5%) and DSC of 0.83 (IQR 0.79 to 0.88). Relatively large volumetric changes in bladder (median −24.5%, IQR −34.6% to 14.5%) and rectum (median 5.4%, IQR − 9.7% to 20.7%) were observed while shape changes were moderate (median DSC of 0.79 and 0.73, respectively). The median CTV-V95% was 98.4% (IQR 94.9% to 99.6%) for the predicted doses. However, SMART would have been deemed beneficial for 78.2% of the 55 fractions based on target undercoverage (16.4%), exceeding OAR constraints (50.9%), or both (10.9%). Simulated SMART improved the dosimetry and met dosimetric criteria in all fractions. Moderate correlations were observed between the CTV-V95% and target DSC (R2 = 0.73) and bladder mean dose versus volumetric changes (R2 = 0.61). Conclusions: Interfractional dosimetric variations resulting from anatomic deformation are commonly encountered with post-prostatectomy RT and can be mitigated with SMART.