An interactive plan and model evolution method for knowledge-based pelvic VMAT planning

To test if a RapidPlan DVH estimation model and its training plans can be improved interactively through a closed-loop evolution process.

Eighty-one manual plans (P₀ ) that were used to configure an initial rectal RapidPlan model (M₀ ) were reoptimized using M₀ (closed-loop), yielding 81 P₁ plans. The 75 improved P₁ (P₁₊ ) and the remaining 6 P₀ were used to configure model M₁ . The 81 training plans were reoptimized again using M₁ , producing 23 P₂ plans that were superior to both their P₀ and P₁ forms (P₂₊ ). Hence, the knowledge base of model M₂ composed of 6 P₀ , 52 P₁₊ , and 23 P₂₊ . Models were tested dosimetrically on 30 VMAT validation cases (P_v ) that were not used for training, yielding P_v (M₀ ), P_v (M₁ ), and P_v (M₂ ) respectively. The 30 P_v were also optimized by M_{2_new} as trained by the library of M₂ and 30 P_v (M₀ ).

Based on comparable target dose coverage, the first closed-loop reoptimization significantly (P < 0.01) reduced the 81 training plans' mean dose to femoral head, urinary bladder, and small bowel by 2.65 Gy/15.63%, 2.06 Gy/8.11%, and 1.47 Gy/6.31% respectively, which were further reduced significantly (P < 0.01) in the second closed-loop reoptimization by 0.04 Gy/0.28%, 0.18 Gy/0.77%, 0.22 Gy/1.01% respectively. However, open-loop VMAT validations displayed more complex and intertwined plan quality changes: mean dose to urinary bladder and small bowel decreased monotonically using M₁ (by 0.34 Gy/1.47%, 0.25 Gy/1.13%) and M₂ (by 0.36 Gy/1.56%, 0.30 Gy/1.36%) than using M₀ . However, mean dose to femoral head increased by 0.81 Gy/6.64% (M₁ ) and 0.91 Gy/7.46% (M₂ ) than using M₀ . The overfitting problem was relieved by applying model M_{2_new} .

The RapidPlan model and its constituent plans can improve each other interactively through a closed-loop evolution process. Incorporating new patients into the original training library can improve the RapidPlan model and the upcoming plans interactively.