Assessing the cause and effects of objective function variation approach on volumetric modulated arc radiosurgery of...

To demonstrate the cause and effects of dose gradients in normal brain and resulting plan metrics using varied objective functions for generating efficient volumetric modulated arc radiosurgery plans for solitary brain metastasis.

Normal tissue and dose volume objective function (NTO and DVO) variations were applied to the virtual phantom and 23 solitary brain metastasis (SBM) cases to plan 20 Gray (Gy) doses. Initial testing on the phantom used eight variations manipulating NTO priority (50, 100, 150, 200), end-dose (25% and 10%), and DVO target dose maximum (25 and 27 Gy). Two strategies were then tested on SBM cases. First strategy (S1) used NTO priority = 100, end-dose = 25%, target maximum dose < 25 Gy. The second strategy (S2) tightened NTO and relaxed DVO by changing the corresponding parameters to 200 and 10%, and 27 Gy. Plans were analyzed comparing and correlating relevant metrics.

For the phantom case, the mean multi-leaf collimator aperture area (AA) variation between the highest NTO priority S2 plan and lowest priority S1 plan was about 46%, which brought substantial changes to 12 Gy spill outside tumor (V12), efficiency index, and prescription isodose level (PIDL). SBM cases showed about 14% lower PIDL and associated V12 reduction of about 0.7 cm3 for S2 than S1. V12 reduction was explained by AA reduction (r ≥ 0.966). For the studied PTVs, monitor units per Gy (MU/Gy) and AA product were lower for S2 indicating one or either factor was lower.

Stringent NTO and relaxed DVO functions cause aperture shrinkage to create dose gradients for overall plan efficacy.