Click here to


Are you sure ?

Yes, do it No, cancel

Optimization of Marker-Based VMAT Plans for Lung Tumor Tracking

A Omotayo*, S Venkataraman , B McCurdy , CancerCare Manitoba, Winnipeg, MB


(Sunday, 7/29/2018) 3:00 PM - 6:00 PM

Room: Exhibit Hall

Purpose: Investigating incorporation of real-time tumor tracking visualization parameters into the dosimetric optimization of marker-based modulated arc therapy plans. The optimized plans can be applied during treatment for real-time tumor tracking.

Methods: Lung tumor 3-dimensional motion was simulated and obtained via a dynamic thorax phantom. VMAT lung SBRT plan with two partial arcs on Varian Eclipse TPS (v11.0.31) was created and treatment was delivered on a 6MV Varian EDGE Linac, using an ITV approach. The prescription dose for PTV was 48Gy over four fractions. The VMAT plan was then exported to an inverse planning and optimization tool, MonArc. MonArc utilizes the same progressive resolution optimization (PRO) framework as found on earlier versions of Eclipse TPS. Three fiducial markers were implanted around the tumor for localization. We modified MonArc to introduce marker-based constraints (Hard and Soft). Hard constraint imposes visualization (avoiding MLC aperture blockage) of all markers, while Soft constraint penalizes visualization (discourages MLC blockage) for markers in the beam’s eye-view. Dose distributions from plans with marker-based constraints were then compared to plans without marker-based constraints, to analyze the discrepancies using mean dose to target and organs-at-risk, and dose-volume histograms. Optimized plans were re-imported into Eclipse and doses calculated with AcurosXB (v13.6.23) to ensure dosimetric accuracy.

Results: The mean dose to the PTV was 106%, 108% and 109% for plans with No, Soft and Hard constraints respectively. For organs at risk, mean dose to the spinal cord was 8.9%, 9.7%, and 10.2% respectively. Conformity index of PTV was 1.70, 1.28 and 1.76.

Conclusion: We have demonstrated that dose distributions obtained from imposing marker-based Hard constraints degrade the plan quality, as opposed to discouraging MLC aperture blockage of all markers. We hypothesize that this modified optimization strategy can be used to improve the performance of real-time markerless tracking methods using VMAT delivery.

Funding Support, Disclosures, and Conflict of Interest: Research Manitoba


Optimization, Treatment Planning, Lung


TH- External beam- photons: VMAT dose optimization algorithms

Contact Email