Feasibility and Benefit of Reduced CTV-to-PTV Margins In Prostate VMAT Using Monte Carlo–Based Geometric Perturbation
Abstract
Purpose
The MIRAGE trial demonstrated reduced acute toxicity in intact prostate radiotherapy in MRI arm, partially attributed to the use of smaller PTV margins (2 mm versus 4 mm with CT guidance). Motivated by these findings, this study retrospectively evaluates whether CTV-to-PTV margins can be further reduced while maintaining acceptable target coverage.
Methods
Fifty intact prostate cancer patients treated with clinical VMAT plans were retrospectively analyzed. In contrast to conventional margin criteria (3 mm in most directions and 2 mm posteriorly), the PTV was defined using a reduced anisotropic margin of [0, 0, 2] mm in the LR, AP, and SI directions. Geometric uncertainty and robustness were modeled following van Herk et al. (2000). Random setup errors (σ = 1–2 mm) were simulated via dose blurring, and systematic errors (Σ) were evaluated using rigid dose shifts. Robustness was assessed using population-based criteria requiring ≥90% of patients to meet minimum CTV coverage and population TCP loss <1%. TCP and NTCP were evaluated using HyTEC parameters.
Results
VMAT plans exhibited a dosimetric buffer beyond the PTV to the 90% isodose surface. Acceptable coverage (≥90% of patients) was maintained for systematic uncertainties within an ellipsoidal tolerance region with semi-axes of [1.6, 1.3, 2.3] mm in the X, Y, and Z directions. For high-risk disease, acceptable TCP was preserved within a smaller systematic-shift tolerance ellipsoid with semi-axes of [1.3, 1.0, 1.9] mm. Concurrently, OAR doses and NTCP decreased for both rectum and bladder with margin reduction.
Conclusion
For prostate VMAT plans using a [0, 0, 2] mm CTV-to-PTV margin, acceptable target coverage was maintained within a finite tolerance to geometric uncertainty. These results suggest that margin reduction can be dosimetrically feasible in modern inverse-planned prostate radiotherapy, with preserved TCP and improved OAR sparing as reflected by reduced NTCP.