KEY POINTS
- This technical note described the development and clinical implementation of a RapidPlan knowledge-based planning model for lung SBRT in Eclipse v15.6 using 161 coplanar VMAT SBRT plans calculated with Acuros XB, 2.0 mm grid, and 6 MV FFF beams.
- After Model Analytics review, 4 plans were removed for atypical geometric or dosimetric behavior, leaving 157 plans for final DVH estimation and prediction.
- The training set included multiple SBRT prescriptions: 30 Gy × 1, 34 Gy × 1, 12 Gy × 4, 10 Gy × 5, 18 Gy × 3, and 7.5 Gy × 8, preserving prediction across dose levels rather than normalizing to one regimen.
- The optimization template used explicit target priorities plus a PTV gEUD objective with a = 20, alongside gradient-focused constraints for D2cm, ribs, chest wall, skin, and normal tissue objective behavior.
- In prospective validation on 21 patients, 11 plans (52%) met all clinical acceptance criteria and were equal or superior to approved clinical plans on first RapidPlan optimization; after minor localized adjustments, 15 plans (71.4%) were ultimately equal or superior.
CLINICAL TAKEAWAY
This paper is most useful as a reproducible implementation blueprint for centers building a lung SBRT RapidPlan model, especially where steep dose gradients and chest wall or rib interfaces drive manual iteration. The gEUD-based template appears practical and locally adjustable, with most corrections taking 5–15 minutes when first-pass plans failed. But this is a technical note without clinical outcome, toxicity, or prospective randomized comparison, so it supports workflow standardization rather than proving superior patient benefit.