Dosimetric Evaluation of Prostate SBRT Using Deformable Image Registration
Abstract
Purpose
Adaptive radiotherapy (ART) is designed to improve target coverage and normal tissue sparing compared with non-ART treatment. While in prostate SBRT differences between adaptive and non-adaptive plans are often modest, certain fractions exhibit clinically significant under-coverage without plan re-optimization. This study quantified the dosimetric impact of these under-dosed fractions by accumulating fractional doses on daily anatomy using deformable image registration to the reference planning CT.
Methods
From a retrospective institutional prostate SBRT cohort delivered on a Varian Ethos platform (250 fractions), nine fractions were identified with severe under-coverage on scheduled plans, showing notable inter-fractional variability in a subset of patients. Three patients had D95%<50%, four patients had one fraction each with D95%=50–80%, and one patient had two fractions with D95%=75–80%. For each affected patient, the full treatment course was exported, and fractional doses were mapped to the reference planning CT using deformable image registration to calculate cumulative dose. Reconstructed prostate doses were compared with the planned doses to quantify deviations.
Results
Our results showed that, despite occasional large fractional anatomical variations and severe under-coverage, cumulative prostate dose remained close to the prescription dose, with differences ≤ 3% of prescription dose (Dp=37Gy). In one patient, accumulated D95% was ~81% of Dp because Ethos automatically shifted the treatment isocenter to the prostate centroid. Dose reconstruction without this auto-isocenter shift restored the cumulative dose to the prescription dose. It should be noted that if multiple fractions exhibited severe under-coverage, larger cumulative deviations could occur.
Conclusion
Despite occasional large anatomical variations, cumulative prostate dose was largely preserved except for multiple fractions with severe under-coverage. These findings highlight the potential utility of adaptive planning and call for flexible workflows that allow free switch between IGRT and ART based on clinical needs to improve treatment efficiency while maintaining dosimetric fidelity.