Evaluation of dosimetric advantages of using patient‐specific aperture system with intensity‐modulated proton therapy for the shallow depth tumor

Abstract In this study, we evaluate dosimetric advantages of using patient‐specific aperture system with intensity‐modulated proton therapy (IMPT) for head and neck tumors at the shallow depth. We used four types of patient‐specific aperture system (PSAS) to irradiate shallow regions less than 4 g/cm2 with a sharp lateral penumbra. Ten head and neck IMPT plans with or without aperture were optimized separately with the same 95% prescription dose and same dose constraint for organs at risk (OARs). The plans were compared using dose volume histograms (DVHs), dose distributions, and some dose indexes such as volume receiving 50% of the prescribed dose (V50), mean or maximum dose (Dmean and Dmax) to the OARs. All examples verified in this study had decreased V50 and OAR doses. Average, maximum, and minimum relative reductions of V50 were 15.4%, 38.9%, and 1.0%, respectively. Dmax and Dmean of OARs were decreased by 0.3% to 25.7% and by 1.0% to 46.3%, respectively. The plans with the aperture over more than half of the field showed decreased V50 or OAR dose by more than 10%. The dosimetric advantage of patient‐specific apertures with IMPT was clarified in many cases. The PSAS has some dosimetric advantages for clinical use, and in some cases, it enables to fulfill dose constraints.

and effective spot size. This limitation complicates the treatment planning so correspond to this limitation. One method to irradiate shallower regions less than 4 g/cm 2 is to use an energy absorber (EA); however, the monoenergetic beam with an EA has a large spot size due to scattering and has a limited ability to make sharp lateral penumbra, worsening the dose distribution. Although in some cases the scanning method has some issues with the lateral penumbra, it has some advantages over the passive method and x-ray treatment, and a number of studies using scanning proton beams reported dose comparisons of passive beam and x-ray beam 3,4 or verification of the relative biological effectiveness (RBE). 5 Flexibility of scanning technique allows for various irradiation methods, and these methods can improve the dose distribution. Developments in treatment planning systems (TPS), treatment machines, and treatment techniques permit the use of intensity-modulated proton therapy (IMPT) for various tumor sites. 6,7 Figure 1 shows the archived spot sizes with or without EA that were retrieved from the Hitachi ProBeat III operating in the Nagoya Proton Therapy Center (NPTC), 8 and these data were presented in a previous paper. 9 Without the EA, the maximum spot size in air was 13.8 mm at the isocenter plane and the minimum range was 4 g/cm 2 . Regions below 4 g/cm 2 could be irradiated using EA, but the spot size of minimum range was more than 26.7 mm. These bigger lateral penumbras mean a worse dose distribution. 10,11 To avoid these issues, a variety of useful methods, such as advanced optimize method (contour scanning), 12 improvement of beam line, 13 and various type of aperture or collimator, [14][15][16][17] have been proposed. In the same way, methods to reduce spot size owing to close the bolus or nozzle to patient were also reported. 18,19 These methods enable to achieve a sharp lateral penumbra, better dose distribution, and lower out-of-field dose. Although various useful methods were reported in recent years, we started scanning treatment with a patient-specific aperture system (PSAS) in May 2016 to irradiate shallow regions and to obtain sharp lateral penumbras. The PSAS has some downsides that require patient-specific manufacture and the weight of aperture is a burden to therapists. However, the treatment procedures and QA of the PSAS methods were the same as those with non-PSAS methods. Furthermore, PSAS methods do not require complicate equipment and machine maintenance such as MLC, so it is easy to apply PSAS in a clinical setting. In the previous study, we presented verification results of the PSAS for the spot scanning beam and the results showed that PSAS reduced the lateral penumbra by 30%-70% in the simple case. 9 The PSAS can be used with both single-field uniform dose (SFUD) and IMPT, but the dosimetric advantage of the patient-specific aperture for IMPT has not been clarified. In this study, we clarified the dosimetric advantage of the aperture for IMPT using DVH, dose distribution, and some dose indexes.

2.A | Scanning delivery system with PSAS
The spot size of the proton beam increases because of scattering by the materials in the beam nozzle and the distance from the exit of beam transport system to the isocenter. Each spot is controlled by scanning magnet and the maximum field size is 30 9 30 cm 2 in the NPTC. Ninety-five energies are available for the scanning treatment that are ranging from 71.6 to 221.4 MeV, resulting in water equivalent penetration depths of 4-30.6 g/cm 2 at intervals of 0.1 g/cm 2 at low energy beam and 0.6 g/cm 2 at high energy beam. The irradiation system is able to use a range shifter from 0.1 to 0.5 g/cm 2 for fine adjustment of the range of high or middle energy beams. The lateral penumbra is generally affected by the distance from the collimator to the surface, depth, beam energy, and energy absorber thickness.
Thus, we designed four types of PSAS with different field sizes and distances from the isocenter to put the aperture closer to the patient's body surface. Small PSASs allowed access over the patient's body or patient's immobilization devices for small tumors. The EA had 4 g/cm 2 water equivalent thickness. The maximum water equivalent penetration with PSAS was restricted to 15 g/cm 2 because the patient-specific aperture was made from 3-cm thick brass. Figure 2 and Table 1 showed schematic view and fundamental parameters of the four types of PSASs. The field sizes of the large PSASs (Types 1 and 2) and small PSASs (Types 3 and 4) were 25 9 25 cm 2 and 10 9 10 cm 2 , respectively. The distance from the isocenter to the aperture was 150-345 mm. As Types 2 and 4 had the same isocenter-to-aperture distance, the spot sizes were the same and the largest spot size was 18.9 mm. Type 1 shows the largest spot size that is 26.7 mm.   Table 2. All PSAS fields including without aperture fields listed in Table 2 were using the EA. We used a worst case optimization 22 for dose optimization to obtain the robust plan.
Parameters of the worst case optimization were 3 mm for x, y, z direction and 3.5% for range uncertainty, 23 so 9 scenarios were considered to the target and OARs. The value of 3 mm is the uncertainty of patient setup for HN tumor and machine variability in NPTC. The value of 3.5% is range uncertainty resulting from uncertainties of the range calculation, the acquisition of CT number, and CT number-Stopping power conversion table. CT images were acquired and reconstructed with 1 mm slices. In the aperture field, the aperture margin was required to assure the marginal dose of the target. The aperture margin was computed by expanding of the maximum outline of the target from the beam's eye view and the margin set to the same value as the spot spacing. The spot spacing was affected by spot size, so the aperture margin was being from 7.2 mm to 11.2 mm at isocenter.

2.C | Plan evaluation
We compared with-and without aperture plans using dose volume histogram (DVH), dose distribution, and some dose indexes. As mentioned earlier, we made with-or without aperture plans with the same target dose so that the 95% dose to the target was equal. In this study, we used relative reduction of some dose indexes such as volume receiving 50% of the prescribed dose (V 50 ) and maximum dose (D max ) or mean dose (D mean ) to the OARs. V 50 was used for estimate to the out-of-field dose. The relevant OARs of each plan were varied so we evaluated the relevant OARs as described in Table 2. Lens, optic nerve, brain stem, chiasm, and eye were analyzed using D max . Parotid, tongue and brain were analyzed by D mean .

| DISCUSSION
This study investigated the dosimetric advantages of the patient-specific aperture to HN IMPT. The patient-specific aperture reduced Plan dose constraints of ID 2, 6, 7 only fulfilled with the aperture plan and other plans can reduce unnecessary out-of-field dose. The decrease rate of D max in some cases was dramatic but other cases showed small effects because OARs were close to the CTV. In this study, we optimized plans while prioritizing the target dose so many T A B L E 2 Plan information with fields, gantry angle, prescribed dose, CTV volume, and OAR.  Fig. 3.

ID
F I G . 5. Relative reductions of V 50 , D max , and D mean OAR doses between with and without aperture.
limitations, it is assumed that a movable nozzle and MLC system would improve the lateral penumbra and treatment throughput.
Some additional methods have been investigated, for example, to use dynamic collimation with MLC (layer-by-layer collimation) 24 or dynamic collimation (spot-by-spot collimation). 16,25 These methods will come into practical and might improve the dose distribution more. But at present, PSAS system showed dosimetric advantage for shallow region treatment, and in some case, it enables to fulfill dose constraints.

| CONCLUSION
We herein reported improvement of dose distribution by using patient-specific apertures with IMPT for shallow depth tumor. The PSAS has some dosimetric advantages for clinical use and is easy to use because it does not require complex machine or control mechanism. Using the PSAS has some demerits from the viewpoint of patient throughput and usability; however, it is useful for clinical application.

ACKNOWLEDGMENTS
We thank the members of NPTC and Hitachi Ltd. who cooperated with this study.

CONF LICT OF I NTEREST
The authors have no conflicts of interest directly relevant to the content of this article.