3389 - Assessment of CT Simulation-Based Prostate Radiotherapy Plan Robustness Using Serial Cone-Beam CT
Presenter(s)
S. Targ1, V. Sato1, N. J. Zhou1, J. Shi2, D. E. Go3, and H. J. No3; 1Southern California Permanente Medical Group, Los Angeles, CA, 2Kaiser Permanente Research, Pasadena, CA, 3Southern California Permanente Medical Group, Ontario, CA
Purpose/Objective(s): CT simulation planning is standard for definitive prostate radiotherapy (RT), though anatomic variation during treatment may compromise Planning Target Volume (PTV) coverage and organ at risk (OAR) dose. We evaluated longitudinal plan robustness using serial cone-beam computed tomography (CBCT) to assess whether first-fraction CBCT can risk-stratify patients for adaptive evaluation.
Materials/Methods: We retrospectively reviewed 40 patients treated with definitive RT for early-stage prostate cancer (70 Gray in 28 fractions) at one institution from 12/2021 to 12/2023. The delivered simulation plan was evaluated on CBCT images acquired at fraction 1 (CBCT1), mid-treatment (CBCT2), and final fraction (CBCT3). Acceptable coverage was defined as PTV D95 =95% and V70 =90%. Paired comparisons used Wilcoxon signed-rank tests and McNemar’s exact test. Rectum and bladder volumes were recorded at simulation and CBCT1–3. Predictive discrimination was evaluated with receiver operating characteristic (ROC) analysis, with optimal cutpoint determined by Youden index. A =20% threshold was selected for clinical interpretability and tested using logistic regression and exact methods.
Results: Nine of 40 patients (22.5%) maintained acceptable PTV coverage from the first through final CBCT evaluation. Relative to simulation, rectum and bladder dose metrics increased during treatment (p < 0.001). Rectum V65 =25% constraint violations emerged over treatment (p = 0.004), whereas bladder V65 =50% violations did not increase (p = 1.000). Failure to meet PTV coverage on CBCT1 was strongly associated with subsequent failure on CBCT2 and CBCT3 (OR 18.0, 95% CI 2.0–164.3, p = 0.002). Absolute rectal volume deviation at CBCT1 demonstrated discrimination for subsequent PTV coverage failure (AUC 0.783). The optimal ROC-derived cutpoint was 18.9%. A =20% rectal volume deviation at CBCT1 identified a high-risk subgroup for persistent PTV undercoverage (70.0% vs 10.0%, p = 0.020). Greater maximal rectal deviation across images was associated with lower minimum PTV D95 observed during treatment (p = 0.012). Bladder volume change was not independently associated with PTV coverage degradation.
Conclusion: CT simulation–based prostate RT plans frequently fail to maintain PTV coverage across treatment when assessed on subsequent CBCT imaging. Rectal instability appears to contribute to this effect, whereas bladder variation was not associated with PTV coverage degradation. A =20% rectal deviation on first-fraction CBCT identifies a high-risk subgroup for persistent PTV undercoverage and supports a selective adaptive RT triage strategy.