Thoracic tumor treatment course assessment based on 4D dose accumulation for scanned proton therapy

Purpose or Objective
With the increase of proton therapy facilities worldwide featuring Pencil Beam Scanning (PBS) as their only treatment modality, PBS is on the way of becoming the standard for proton therapy. However, for some indications in the thoracic region PBS is not widely used due to uncertainties in the planned dose, which can be caused by combined effects of setup errors, range uncertainty, interplay effect, breathing irregularity, anatomical variations, delivery machine uncertainties, etc. By performing pre-treatment plan robustness evaluation that includes these effects, it is evident that actual delivered fractional dose at any instance is highly uncertain to predict. 4D dose accumulation is able to control some of the uncertainties that are affecting pretreatment evaluation of the plan quality. Therefore, the purpose of this proof-of-concept study is to investigate the
feasibility of monitoring and assessing the quality of delivered treatment fractions throughout the treatment course.

Material and Methods
4D dose accumulation is performed by utilizing (1) delivery machine log files (IBA, Belgium), (2) breathing pattern records (ANZAI, Japan) and (3) planning 4DCT scans or repeated 4D control CT scans (Siemens, Germany). Dose computation is performed in the RayStation (RaySearch, Sweden) treatment planning system (TPS). For every spot that is delivered during a particular fraction, the spot energy, position, dose (as charge) and absolute time of delivery is retrieved from the machine
log file using a dedicated script. Patient’s breathing pattern is analyzed and inhale peaks are determined. Subsequently, all breathing cycles are divided in 10 phases and each phase is associated with absolute time. PBS spots are split in 10 groups according to their corresponding phase and written to 10 treatment sub-plans (DICOM), where every sub-plan corresponds to a particular phase of the 4DCT. Using scripting capabilities of the TPS, sub-plans are imported for dose computation. Eventually dose warping to the reference phase is performed to estimate the delivered fractional dose. Data sets used for the proof-of-concept were not collected during the same treatment fraction.

Results
By using the described method the timeline of a PBS delivery can be correlated with patient’s breathing pattern as shown in Figure 1. Computation of log based sub-plans on 4DCT results in an accumulated fractional 4D dose distribution as shown in Figure 2. Based on the exemplary case, the method allows to assess the conformity between planned and delivered doses (i.e.,
CTV V95 has dropped to 96.7% from nominal 100%).

Conclusion
The availability of a real-time 4D dose accumulation based treatment assessment tool allows to assess the quality of the delivered dose during progress of the treatment course and to take appropriate actions, as for example, plan adaptation, in cases of significant deviations. It is foreseen to extend this study for a full treatment course of a broader population of patients.