KEY POINTS
- This retrospective single-center study evaluated 15 patients with localized prostate cancer previously treated with online adaptive SBRT using the Varian Ethos platform between June 2024 and December 2025. Three plans were retrospectively generated for each patient using identical CT images, contours, beam geometry, and prescriptions: a standard Ethos Intelligent Optimization Engine v2 (IOE v2) plan, a plan guided by an initial RapidPlan model trained on 20 patients (RP1), and a plan guided by a refined 35-case model (RP2). Importantly, the RP2 training dataset included the same 15 evaluation cases.
- Treatment planning incorporated 36.25 Gy in five fractions to the prostate PTV, with simultaneous boosts to 40 Gy for the intraprostatic boost volume and 42.5 Gy for the dominant intraprostatic lesion microboost. The prostate CTV was expanded by 4 mm, reduced to 3 mm posteriorly, and a separate urethral optimization structure supported urethral sparing. Plans used 9-field or 12-field IMRT configurations, with the same institutional dose constraints across all three optimization strategies.
- Bladder sparing improved significantly with both RapidPlan-guided approaches. Mean bladder dose decreased from 12.05 Gy in reference plans to 11.31 Gy with RP1 and 11.19 Gy with RP2, corresponding to absolute reductions of 0.74 and 0.86 Gy, respectively (both p<0.05). Bladder V15Gy decreased from 35.80% to 33.01% and 32.46%, reductions of 2.80 and 3.35 percentage points (both p<0.05). However, bladder D0.03cc and rectal V18Gy/V25Gy showed no statistically significant improvements.
- Additional dosimetric benefits were observed for the penile bulb and sigmoid colon. Penile bulb mean dose decreased from 4.48 Gy to 3.34 Gy with RP1 and 3.18 Gy with RP2, while penile bulb D0.03cc decreased from 17.96 Gy to 13.86 and 12.76 Gy, respectively (all p<0.05). Sigmoid D0.03cc was reduced from 6.49 Gy to 3.57 and 3.55 Gy (both p<0.05). The refined RP2 model achieved additional significant reductions in penile bulb mean and near-maximum dose compared with RP1, but most other metrics were similar between the two models.
- Improved OAR sparing came at the expense of prostate PTV coverage. Mean PTV_3625 D95 decreased from 36.30 Gy in reference plans to 35.65 Gy with RP1 and 35.71 Gy with RP2, while D98 decreased from 35.65 Gy to 34.78 and 34.93 Gy, respectively (all p<0.05). In contrast, coverage of the 42.5-Gy microboost volume was maintained, with D95 values of 42.32, 42.46, and 42.42 Gy for reference, RP1, and RP2 plans. The intermediate 40-Gy boost structure also showed maintained or slightly improved coverage.
- Urethral high-dose exposure increased despite improvements elsewhere. Urethral D0.03cc increased from 37.81 Gy in reference plans to 38.26 Gy with RP1 and 38.16 Gy with RP2, corresponding to increases of 0.44 and 0.35 Gy (both p<0.05). All plans remained within the institutional 40-Gy urethral D0.03cc constraint, but the urethral optimization PTV volume receiving 105% of the 36.25-Gy prescription increased from 0.36 cm³ to 0.88–0.89 cm³ (both p<0.05), highlighting a potentially important high-dose trade-off.
- Plan monitor units decreased by approximately 17% with RapidPlan guidance, from 3,415 MU in reference plans to 2,817 with RP1 (p<0.001) and 2,836 with RP2 (p=0.007). However, the study did not measure actual treatment delivery time, adaptive workflow efficiency, or clinical toxicity. The small cohort, institution-specific templates, and absence of independent validation for RP2 limit generalizability; incorporating the evaluation patients into RP2 training also prevents treating its results as independent evidence of improved model performance.
CLINICAL TAKEAWAY
RapidPlan guidance can improve selected OAR dose metrics and reduce monitor units during automated adaptive prostate SBRT planning, but these benefits were accompanied by reduced standard prostate PTV coverage and increased urethral near-maximum dose. The findings support further refinement of knowledge-based optimization objectives rather than routine replacement of existing planning templates, with independent validation and careful review of target coverage and urethral dose remaining essential.