2415 - Application of Photon-Derived Worst-Case Robustness Criteria to Prostate Cancer Proton Therapy Planning

K. Siwakoti¹, A. P. Dalton², J. A. Maas³, A. M. McDonald², S. Marcrom⁴, R. A. Cardan², J. Harms², J. B. Fiveash⁵, and A. J. Kole²; ¹Department of Medicine, University of Alabama at Birmingham, Birmingham, AL, ²Department of Radiation Oncology, University of Alabama at Birmingham, Birmingham, AL, ³Legacy Medical Group-Radiation Oncology at Good Samaritan, Portland, OR, ⁴University of Alabama at Birmingham Department of Radiation Oncology, Birmingham, AL, ⁵University of Alabama at Birmingham, Birmingham, AL

Purpose/Objective(s): Robustness evaluation is critical for proton beam therapy (PBT) planning, but ideal robustness criteria have not been clearly defined. We previously characterized x-ray intensity modulated radiation therapy (IMRT) robustness among patients with prostate cancer and identified CTV V100% > 90% and CTV V95% > 99% as minimal CTV worst-case coverage thresholds (Maas et al, PRO, 2024). Here we present a follow up study evaluating PBT robustness using similar benchmarks and the dosimetric impact of plan modification to meet robustness goals. Materials/

Methods: Patients with prostate adenocarcinoma treated with PBT from 2021-2022 at a single institution were retrospectively evaluated. Treatment was delivered to 70 GyE in 28 fractions to the prostate alone. SpaceOAR was present in all patients. Robustness evaluation was calculated with 5 mm setup error in each of 6 cardinal directions and range uncertainty of +/-3% (12 uncertainty scenarios). Nominal and worst-case data were collected for CTV V100%, CTV V95%, rectum V70 Gy, and bladder V60 Gy. Clinically delivered (“Clinical”) plans were evaluated for passing the minimal worst-case CTV coverage benchmark. If the benchmark was unmet, plans were modified (“Benchmark” plans) to meet both CTV V100% and V95% goals. Clinical plans were re-normalized (if global hot spot remained <107%) or re-optimized until worst-case CTV and hot spot goals were met. Two tailed paired t-tests with alpha set at 0.05 were used to compare dosimetric parameters of Clinical and Benchmark plans.

Results: 32 patients were evaluated with a median age of 71 years and a median PSA of 6.84 ng/mL at diagnosis. Most patients were clinical T1c (68.8%), favorable-intermediate risk group (56.3%) and not on androgen deprivation therapy (74.2%). For Clinical plans, the mean nominal CTV V100% was 98.8% and mean worst-case CTV V100% was 79.8%. The mean nominal and worst-case CTV V95% were 100% and 98.9%, respectively. A worst-case CTV V100% > 90% and CTV V95% > 99% were achieved in 50% and 53% of plans respectively; only 31% met both IMRT-derived criteria. Plan renormalization (16 patients) or reoptimization (6 patients) was conducted with all modified plans meeting CTV V100% and CTV V95% thresholds with global hot spot of <107%. Comparing Clinical and Benchmark plans, nominal rectum V70 Gy increased from 0.72 cc to 0.92 cc, and nominal bladder V60 Gy increased from 5.9% to 6.0% (P < 0.05 for each parameter). Overall plan hot spot between Clinical and Benchmark plans increased from 104.5 to 105.5% (P < 0.05).

Conclusion: When compared to an IMRT-derived benchmark for robustness coverage, PBT plans were less robust, with only 1/3 of patients meeting minimal worst-case CTV V100% and CTV V95% goals. However, with re-normalization or re-optimization, all plans met the worst-case CTV benchmark with minimal expected clinical impact on plan hot spot or organ at risk dosimetry. Consideration should be made to adopt these benchmark criteria for men receiving PBT to the prostate.