Positioning of head and neck patients for proton therapy using proton range probes: a proof of concept study

Abstract

To exploit the full potential of proton therapy, accurate and on-line methods to verify the patient positioning and the proton range during the treatment are desirable. Here we propose and validate an innovative technique for determining patient misalignment uncertainties through the use of a small number of low dose, carefully selected proton pencil beams ('range probes') (RP) with sufficient energy that their residual Bragg peak (BP) position and shape can be measured on exit. Since any change of the patient orientation in relation to these beams will result in changes of the density heterogeneities through which they pass, our hypothesis is that patient misalignments can be deduced from measured changes in Bragg curve (BC) shape and range. As such, a simple and robust methodology has been developed that estimates average proton range and range dilution of the detected residual BC, in order to locate range probe positions with optimal prediction power for detecting misalignments. The validation of this RP based approach has been split into two phases. First we retrospectively investigate its potential to detect translational patient misalignments under real clinical conditions. Second, we test it for determining rotational errors of an anthropomorphic phantom that was systematically rotated using an in-house developed high precision motion stage. Simulations of RPs in these two scenarios show that this approach could potentially predict translational errors to lower than1.5 mm and rotational errors to smaller than 1° using only three or five RPs positions respectively.

Technical Note: First report on an in vivo range probing quality control procedure for scanned proton beam therapy in head and neck cancer patients

“…As demonstrated by Farace et al, residual setup errors would cause increased range errors along areas of high heterogeneity index Hi values. Effects of residual misalignment and density interface have also been investigated by Hammi et al 14,24 Such patterns were not observed in our data set along, for example, the spine (see Fig. 4), however to some extent could have been a contributing factor towards mean relative range error variation between 1st and 2nd QC based on the reduced data set of Patient 7.…”

“…In practice, noise on the proximal side of the peak has limited impact on the fitting since gradients on the proximal side are much lower than on the distal side of the peak. Additional metrics for quantifying range probing measurements have been investigated and proposed in literature, such as the weighted mean range and range dilution 24 . Future improvements of the RP QC procedure could be warranted by investigating applicability of additional metrics and introducing them into result reports.…”

Experimental helium‐beam radiography with a high‐energy beam: Water‐equivalent thickness calibration and first image‐quality results

“…This application was suggested for proton radiography early on 14 and over the past 10 years, several studies based on irradiations of a few isolated pencil beams called range probes, or on full radiographs using single-proton tracking were published. [15][16][17][18] These applications of iRAD could potentially be performed at an imaging dose to the patient that is about 10 times lower than for X-ray radiography under the condition of same density resolution and pixel size, referring to studies about proton imaging. 6,14,18,19 For example, the application of daily iRAD would not lead to an increase of dose if the daily orthogonal X-rays for patient alignment can be replaced by iRAD.…”

“…Second, i Rad could simultaneously be used for the alignment during patient setup. This application was suggested for proton radiography early on 14 and over the past 10 years, several studies based on irradiations of a few isolated pencil beams called range probes, or on full radiographs using single‐proton tracking were published 15–18 …”

Classification of various sources of error in range assessment using proton radiography and neural networks in head and neck cancer patients

“…In previous studies, PR was used with the aim to identify individual sources of error (Farace et al 2016a(Farace et al , 2016b). In addition, some studies investigated experimentally different body sites to perform range probing for positioning purposes with a head phantom (Hammi et al 2018a(Hammi et al , 2018b. Furthermore, the effects of setup errors in range assessment have been evaluated, concluding that they could be distinguished from other sources of error affecting the proton range (Deffet et al 2017).…”