ESTRO 37 Abstract book

ESTRO 37

S488

changes. Mitigating those uncertainties is the strength of robust IMPT.

Poster: Physics track: Treatment planning: applications

PO-0911 Proton therapy of head and neck cancer: comparing PTV-based and robustly optimized IMPT with VMAT S.S. Almberg 1 , A.E. Aarberg 2 , J. Frengen 1 , S. Danielsen 1 1 St. Olavs University Hospital Trondheim, Radiotherapy Department, Trondheim, Norway 2 Norwegian University of Science and Technology, Department of Physics, Trondheim, Norway Purpose or Objective Radiation therapy plays a major role in curative treatment of head and neck cancers. However, the treatment is associated with considerable side-effects, and techniques that can reduce the treatment-related toxicities are of great interest. In this feasibility study, Intensity modulated proton therapy (IMPT) plans were compared to clinically used Volumetric modulated arc therapy (VMAT) plans. Robustness against anatomy changes was also evaluated. Material and Methods IMPT plans were made using a three-field technique for 12 oropharynx cancer patients. Two IMPT plans were made for each patient; one PTV-based (IMPT-PTV) and one robustly optimized (IMPT-Robust). All plans were designed to treat high-, intermediate- and low risk target volumes to different dose levels (68 Gy, 60 Gy and 54 Gy, respectively) using simultaneously integrated boost. The same (or better) target dose coverage as the clinical VMAT plans was required for the IMPT plans. The mean parotid dose was reduced as much as possible without losing target coverage. Normal tissue complication probability (NTCP) values for xerostomia was calculated and employed for comparison of the different treatment plans. A new CT taken halfway through the treatment was used to recalculate the dose distributions and evaluate the robustness of the IMPT plans. All plans were made in RayStation v.5.0. Results Compared to VMAT, IMPT significantly reduced the mean dose to the parotid glands, with a corresponding reduction in the NTCP. Compared to VMAT, IMPT decreased the NTCP for xerostomia by > 10 % for one or both parotids for 9 out of 12 patients. Robustly optimized IMPT plans provided the lowest mean parotid dose and the lowest NTCP values. Recalculation of the IMPT treatment plans on a repeat CT taken halfway through the treatment showed that the IMPT-Robust plans were indeed the most robust regarding target dose coverage: of the 12 patients included in the study, 11 and 6 patients still had adequate target coverage for the IMPT- Robust and IMPT-PTV plans, respectively. Conclusion In this work, IMPT showed to be beneficial compared to VMAT for oropharynx cancer patients by reducing th e dose to the parotid glands, and subsequently reducing the risk for xerostomia. However, as protons are more sensitive to set-up uncertainties and anatomy changes than photons, robustly optimized IMPT plans should be preferred over PTV-based IMPT. Finally, as this work evaluated only a limited set of parameters, more

Figure 1: Schematics of conformal (robust) 4D optimization (yellow boxes: treatment planning; red boxes: simulated course of treatment). A conformal 4D optimization was calculated for conventional IMPT on PTV and robust IMPT. The resulting treatment plan was used in 4D dose calculations on 5 consecutive weekly 4DCTs from 5 fractions. Robustness was tested by comparing the accumulated 4D dose distributions for conventional and robust IMPT, considering OAR sparing and target coverage. Results Our method was tested on a patient case with a lung tumor in direct proximity to the heart, with a motion amplitude > 2 cm and a target dose of 9.4 GyE per fraction. Treatment planning focused on heart sparing, following D 15cc < 6.4 GyE and maximal point dose < 7.6 GyE constraints for the single fractions. Compared to conventional IMPT, robust IMPT reduced the D 15cc value from 4.1 GyE to 3.6 GyE. Moreover, only robust IMPT kept the maximal point dose below the constraint with 6.9 GyE compared to 8.4 GyE with conventional IMPT. Despite the superior OAR sparing target coverage stays on a similar level (conventional IMPT: V95 = 65.7 %, D95 = 74.1 %; robust IMPT: V95 = 71.1 %, D95 = 67.7 %). Furthermore, robust IMPT produces a dose distribution which is more uniform and has less cold spots (see figure 2).

Figure 2: Axial dose cuts comparing planned dose with accumulated dose of 5 weekly fractions for conventional IMPT (a,c) and robust IMPT (b,d). Conclusion An initial patient simulation shows the feasibility and the importance for robust 4D optimization methods for the treatment of moving targets for carbon ions. Only robust IMPT could respect both heart dose constraints over the course of treatment. The conventional IMPT achieves good results for a single fraction, but fails facing uncertainties of the consecutive ones, especially range