Purpose: To simulate treatment delivery with tumor motion by moving the isocenter based on tumor motion waveforms and evaluate the dosimetric impact from the tumor motion variations that differ from the internal tumor volume (ITV) used in volumetric-modulated arc therapy (VMAT) planning.
Methods: A motion-incorporated VMAT planning was applied to simulate plan delivery with actual and simulated tumor motion variations based on 4-dimensional computed tomography (4DCT) acquired at simulation. First, the clinical ITV-based VMAT plans of 9 lung cancer patients with single lesions were segmented into composite plans with multi-beam fields at the control points. Second, tumor motion waveforms were applied to the isocenter. Tumor motion waveforms were created by scaling down the diaphragm motion extracted from the 4DCT images to the tumor motion. The single 4DCT breathing cycle was replicated and concatenated to form a long motion waveform to cover the duration of the 2-4 arc(s) in the clinical plans. Additional tumor motions were simulated by scaling up the tumor waveforms with extra 0.5, 1.0, and 2.0cm motions. Four motion-incorporated plans were created per patient and compared to the clinical VMAT plan, which covered planning tumor volume (PTV) at D95%=100%. The native-motion-incorporated plans were first compared with the clinical VMAT plan as a control, followed by 3 simulated-motion-incorporated plans using D95%=90% as a breach for ITV and PTV.
Results: Using the original 4DCT motion, the mean D95% from all 9 patients is found to be 0.99±0.02. The D95%=90% breach point increases with ITV and PTV, and requires consistent motion variations of 0.5-1.0cm for small tumors (ITV=5-30cc) and ~2.0cm for large tumors (ITV=80-400cc).
Conclusion: This study has demonstrated that large tumor motion variations deviating from simulation can reduce the ITV/PTV coverage substantially. Under clinical conditions, it seems necessary to apply patient breathing monitoring for threshold-based beam gating during treatment delivery.