Dosimetric impact of random spot positioning errors in intensity modulated proton therapy plans of small and large volume tumors

Abstract

Objective To study dosimetric impact of random spot positioning errors on the clinical pencil beam scanning proton therapy plans. Methods and materials IMPT plans of 10 patients who underwent proton therapy for tumors in brain or pelvic regions representing small and large volumes, respectively, were included in the study. Spot positioning errors of 1 mm, −1 mm or ±1 mm were introduced in these clinical plans by modifying the geometrical co-ordinates of proton spots using a script in the MATLAB programming environment. Positioning errors were simulated to certain numbers of (20%, 40%, 60%, 80%) randomly chosen spots in each layer of these treatment plans. Treatment plans with simulated errors were then imported back to the Raystation (Version 7) treatment planning system and the resultant dose distribution was calculated using Monte-Carlo dose calculation algorithm. Dosimetric plan evaluation parameters for target and critical organs of nominal treatment plans delivered for clinical treatments were compared with that of positioning error simulated treatment plans. For targets, D95% and D2% were used for the analysis. Dose received by optic nerve, chiasm, brainstem, rectum, sigmoid, and bowel were analyzed using relevant plan evaluation parameters depending on the critical structure. In case of intracranial lesions, the dose received by 0.03 cm3 volume (D0.03 cm3) was analyzed for optic nerve, chiasm and brainstem. In rectum, the volume of it receiving a dose of 65 Gy(RBE) (V65) and 40 Gy(RBE) (V40) were compared between the nominal and error introduced plans. Similarly, V65 and V63 were analyzed for Sigmoid and V50 and V15 were analyzed for bowel. Results The maximum dose variation in PTV D95% (1.88 %) was observed in a brain plan in which the target volume was the smallest (2.7 cm3) among all 10 plans included in the study. This variation in D95% drops down to 0.3% for a sacral chordoma plan in which the PTV volume is significantly higher at 672 cm3. The maximum difference in OARs in terms of absolute dose (D0.03 cm3) was found in left optic nerve (9.81%) and the minimum difference was observed in brainstem (2.48%). Overall, the magnitude of dose errors in chordoma plans were less significant in comparison to brain plans. Conclusion The dosimetric impact of different error scenarios in spot positioning becomes more prominent for treatment plans involving smaller target volume compared to plans involving larger target volumes. Advances in knowledge Provides information on the dosimetric impact of various possible spot positioning errors and its dependence on the tumor volume in intensity modulated proton therapy.