TU‐E‐141‐01: Using Respiratory Motion as a Degree of Freedom Through Weighted Phase‐Correlated 4D Planning

Abstract

Purpose: We investigate a novel 4D treatment planning strategy that exploits respiratory motion as an additional degree of freedom. The key idea is to deliver more (less) fluence when the tumor target is out‐of‐line (in‐line) with an organ‐at‐risk (OAR). We have developed our weighted‐phase‐correlated (WPC) 4D treatment planning paradigm on a commercial platform (Eclipse, Varian) and designed it for an “open‐gate” delivery, where beam is on throughout the respiratory cycle. Methods: 4DCT data were collected from twenty lung SBRT patients who exhibited >= 5mm tumor motion despite the use of abdominal compression. For each patient, (i)f × θ 3D‐conformal plans were created and corresponding 3D dose distributions were calculated, where θ = number of respiratory phases (10) and f = number of fields (7 – 12) (ii) For each beam, the dose from nine phases was registered to the reference phase (end‐exhale) using a GPU‐based B‐spline deformable registration, NiftyReg. (iv) using dynamic‐penalized‐likelihood optimization, fluence weights were optimized over all dose distributions. This key step ensured that for each motion‐induced state of the anatomy, an optimal fluence would be delivered. (v) A summed fluence map was imported back into Eclipse and an MLC leaf sequence was generated. In order to ensure deliverability, a maximum leaf velocity constraint (v <= 3.5 cm/s) was imposed. Each WPC‐4D plan was compared with the corresponding, clinically‐delivered, ITV‐based plan. Results: In all 20 patients, WPC‐4D plans gave PTV conformity comparable to ITV‐based plans. The percent dose‐sparing (maximum, average) achieved using WPC‐4D compared to ITV‐based planning: Spinal Cord Dmax=(66,20); Esophagus Dmax=(94,29), Heart Dmax=(84,30); Heart Dmean=(82,35); Lung V20=(27,7); Lung Dmean=(27,13). Conclusion: We have developed a novel 4D‐planning solution that exploits rather than trying to mitigate respiratory motion. Initial results indicate that this approach achieves significant dose‐sparing in OARs compared to conventional motion‐management (abdominal compression) + ITV‐based planning. This research was partially supported by Varian Medical Systems, Palo Alto, CA