Click here to


Are you sure ?

Yes, do it No, cancel

Experimental Quantification of Proton Range Uncertainties Arising From Lateral Inhomogeneities

E Baer1,2*, A Lalonde3 , G Royle1 , H Lu4 , H Bouchard3 , (1) University College London, London, UK, (2) National Physical Laboratory, Teddington, UK, (3) Universite de Montreal, Montreal, Quebec, (4) Massachusetts General Hospital, Boston, MA, USA


(Tuesday, 7/31/2018) 10:00 AM - 10:30 AM

Room: Exhibit Hall | Forum 4

Purpose: One of the largest sources of range uncertainties in particle therapy planning arises from lateral inhomogeneities in the irradiated body, introducing range degradation. Conventional dose calculation algorithms use ray tracing (RT) to calculate the beam range, leading to severe range uncertainties. Monte Carlo (MC) was demonstrated to reduce these uncertainties by accurately simulating particle transport in the patient geometry. The aim of this work is to experimentally quantify range uncertainties arising from lateral inhomogeneities with both RT and MC techniques.

Methods: 12 fresh animal tissue samples are used (7 soft tissues, 5 mixtures of soft tissues and bones), including several soft tissue-bone interfaces which produce lateral inhomogeneities. The water equivalent range (WER) of these samples is measured in a 195 MeV spread out Bragg peak proton beam at 192 spots (112 in soft tissues and 80 in bones) throughout the samples. Additionally, the WER is predicted by using a dual-energy CT scan of the samples as input for two different dose calculation algorithms: conventional RT and MC.

Results: The RMS error between measured and predicted WER over all tissues is 1.49% using RT and 0.76% using MC. Using RT, the mean errors and standard deviations on WER are (0.50+/-0.52)% for soft tissues and (-0.43+/- 2.12)% for samples containing bones. With MC, these values are reduced to (-0.34+/-0.27)% in soft tissues and (-0.30+/-1.03)% in samples containing bones.

Conclusion: This work demonstrates the gain in range accuracy when using MC for dose calculation in the presence of tissue inhomogeneities. A reduction of mean range errors and standard deviations can be achieved in both, homogenous soft tissues and inhomogeneous soft tissue-bone mixtures. If RT is used for range prediction, currently applied clinical margins might not be sufficient for some sites, suggesting the implementation of site specific range margins.


Not Applicable / None Entered.


Not Applicable / None Entered.

Contact Email