Robustness evaluation in proton therapy treatment planning: a systematic review

Purpose/Objective:
S4687 Physics - Optimisation, algorithms and applications for ion beam treatment planning ESTR0 2024
Given the limitations in proton therapy (PT) of using target and organ at risk volumes expansions, both in terms of
plan optimization and evaluation, robust planning has gained high visibility in publications and is part of clinical
practice in many centers. However, there is currently no consensus on the methods and parameters to be used for
robust optimization or robustness evaluation (RE). As a first step of the consensus process, the purpose of this work
a is to perform a systematic review of the literature regarding RE.
Material/Methods:
The systematic review followed the methodology proposed by Stoll et al [1]. One expert performed a search on
PubMed. A second expert independently performed a search on Web of Science. In case of disagreement between
experts or between databases (i.e., a paper was proposed and accepted from a certain database but not proposed
by the other database), a third expert took the final decision. In this way the bias of using a single database was
minimized. Figure 1 shows the keywords used for the search on databases and the inclusion/exclusion criteria.
Results:
The results of the systematic review are shown in Figure 2 and Table 1.
S4688 Physics - Optimisation, algorithms and applications for ion beam treatment planning ESTR0 2024
Specifically, Table 1 provides a synthetic summary of the robustness methods used in publications focusing on a
particular tumor site. It can be observed that most of the selected publications implemented robustness evaluation
against setup and range errors (around 85%). About 21% of the papers report the implementation of some form of
probabilistic evaluation. Most sources of uncertainties are well represented, except perhaps radiobiological
effectiveness and linear energy transfer (LET) which amounts to only 8% of the selected publications. Intra-fraction
motion and interplay effect are considered in 85% and 46% of lung studies, respectively, and in 33% of studies related
to the treatment of targets in the mediastinum.
Site Publ
Worstcase
Prob SE RE Inter-F Intra-F IP RBE-LET
# (% of publications per clinical site)
Head and neck 21 76 33 90 90 57 0 0 5
Brain and skull 9 56 44 100 78 33 0 0 22
Breast 5 80 20 100 100 40 60 0 20
Lung 26 85 23 88 88 27 85 46 8
Esophagus 9 89 11 89 89 56 78 33 0
Mediastinum 3 67 0 67 67 67 33 33 0
Abdomen and
pelvis
19 74 0 74 74 42 32 26 5
All sites 92 77 21 87 85 42 63* 34* 8
Table 1 Summary of the robustness parameters used in clinical studies according to clinical sites. Extensive summary
can be found in the appendix. Publ: Publications; Prob: Probabilistic; SE: Setup Errors; Inter-F: Inter-Fraction; Intra-F:
Intra-Fraction; IP: Interplay effect
S4689 Physics - Optimisation, algorithms and applications for ion beam treatment planning ESTR0 2024
*Head and neck, Brain and skull were excluded from the total to compute the percentage
This systematic review found that setup and range uncertainties were widely considered in both clinical applications
and methodological papers. Geometrical uncertainties, mainly related to patient positioning in the treatment room,
were frequently addressed, encompassing factors such as gantry and couch rotation precision, imaging system
resolution, and reliability of immobilization devices. These uncertainties were typically represented by a shift in the
isocenter, incorporating the combined effect of various sources of uncertainty. Regarding range uncertainties, the
reviewed clinical papers treated them as systematic uncertainties and evaluated them at their maximum magnitude
(±3%/±3.5%). Furthermore, most clinical studies did not consider the mitigation effect of fractionation and of random
errors in the evaluation of plan robustness. Only a few studies considered its impact on robustness analysis and few
of them are clinically applicable due to the computational efforts required. Nonetheless, it is worth mentioning that
the actual methods provided by commercial TPSs do not explicitly account for these sources of mitigation, making
the robustness evaluation a limited representation of the treatment course reality.
Conclusion:
Robustness evaluation has become an important area of research in PT. Various approaches, ranging from pragmatic
to comprehensive and computationally intensive methods, have been proposed for evaluating the robustness of
treatment plans. Many studies in PT incorporate robustness evaluation, and clinical centers are increasingly adopting
these strategies with the help of robustness evaluation tools integrated into commercial TPS. However, there are
challenges in achieving harmonization and consistent reporting of robustness evaluation.
Keywords: Robusteness evaluation, proton therapy
References:
[1] C.R.T. Stoll, S. Izadi, S. Fowler, P. Green, J. Suls, G.A. Colditz, The value of a second reviewer for study selection in
systematic reviews, Res. Synth. Methods. 10 (2019) 539–545