"In traditional proton therapy, we treat the target by making sure that we have full dose coverage of the target geometry," he explained. This can be achieved by placing the plateau region of a spread-out Bragg peak (SOBP) over the target, with the distal edge of the SOBP aligned with the edge of the tumour. However, at the distal edge of the proton beam, proton LET is at a maximum, thereby placing the high-LET region within normal tissue. Likewise, in situations where two opposed and overlapping lateral beams are used, high-LET protons are deposited in two regions of normal tissue.
While the relative biological effectiveness (RBE) of a proton beam is conventionally taken as 1.1, proton RBE actually increases with increasing LET. "This LET effect is wasted in normal tissue," said Carabe-Fernandez. "There's also the question of whether the decrease in dose to an organ-at-risk [OAR] may be negated by an increase in LET and the associated increasing biological effect."
So is there an alternative proton treatment planning approach that could exploit these spatial variations of RBE? Carabe-Fernandez suggested that by moving the high-LET region into the target itself, it may be possible to decrease the required prescribed dose without reducing biological effectiveness – essentially "exchanging dose for LET". This shift should also reduce both the LET and the dose to surrounding normal tissues and OARs.
The next task is planning the treatment so as to bring the high-LET protons into the target. Carabe-Fernandez proposed a split-target approach, in which the clinical target volume (CTV) is divided into two, and two proton beam fields are targeted to place their distal edges at the internal edge of each "half target" – effectively meeting in the centre of the tumour. Taking this idea a step further, the CTV could be split into four sections targeted with four fields, still aimed at the centre. "If you carry on splitting the target, to seven fields, you end up focusing the LET further in your target," he explained.
Comparing dose distributions from a standard full-target proton therapy plan with those from 2-, 4- and 7-field split-target plans revealed that as the tumour is divided into an increasing number of targets, dose to the target is maintained while the high-LET region shifts from outside to inside the target. Dose-volume histograms also revealed a reduction in overall dose to normal tissue.
For a 2-field split target, for example, the increased LET enables a reduction of the dose-per-fraction by 9% while keeping the same radiobiological-weighted dose to the target. The 4-field and 7-field approaches, meanwhile, enable dose reductions of 11% and 12%, respectively.
Proton arcs
So how far can we take this approach? Carabe-Fernandez suggested that it's possible to move from discrete beams to continuous beam delivery, describing the technique as "proton-modulated arc therapy", or PMAT. He then examined the feasibility of performing PMAT using proton pencil-beam scanning (PBS) delivery.
If multiple energy layers need to be delivered at each angle, the gantry cannot rotate continuously and PMAT is not feasible in PBS mode, he explained. However, it should be possible to use PBS if a mono-energetic beam is employed and the gantry rotation is used to paint the dose. Nevertheless, patients are not spheres and a single proton energy will not be sufficient to cover a target within an irregular body shape.
As such, Carabe-Fernandez proposed the use of two proton beam energies delivered over subsequent gantry rotations. He presented an example comparing PMAT and PBS treatment plans for a brain tumour located near several critical structures. Both plans were designed to deliver 79.2 Gy over 44 fractions. The PMAT plan used two arcs: the first delivering a constant energy of 113.2 MeV over a full 180° arc; the second delivering an energy of 110.2 MeV, over specified sections of the returning 180 to 0° arc. Carabe-Fernandez and colleagues are currently working on single mono-energetic solutions that will help to substantially reduce the number of monitor units and time required to treat the patient compared to multiple-beam PBS treatments.
Even when using just two proton beam energies, the PMAT plan achieved 90% target coverage. Dose-volume histograms showed that the CTV coverage was very similar to the PBS plan. Doses to OARs, in particular the brain stem and optic chiasm, were greatly reduced for the PMAT plan compared with the PBS plan.
As expected, LET in the target was higher for the PMAT plan than the PBS plan. Carabe-Fernandez noted, however, that some of the OARs also had higher LET. He concluded that while full RBE-based proton plan optimization might yet not be realistic, LET-guided treatment planning is achievable. He also noted that when performing LET-based treatment planning, consideration must be given to the method used to calculate LET, as different Monte Carlo implementations can result in deviations in calculated values.
"PMAT is an interesting option that might allow simultaneous dose and LET painting of a target while delivering the dose in an efficient manner," he concluded. "Other advantages of PMAT include a lower number of monitor units and reduced room time, which could increase the patient throughput."
Related stories
• LET painting improves tumour control
• LET guides proton planning
• Can LET painting overcome hypoxia?
• LET mapping: a treatment planning tool