Small‐spot intensity‐modulated proton therapy and volumetric‐modulated arc therapies for patients with locally advanced non‐small‐cell lung cancer: A dosimetric comparative study

Abstract Purpose To compare dosimetric performance of volumetric‐modulated arc therapy (VMAT) and small‐spot intensity‐modulated proton therapy for stage III non‐small‐cell lung cancer (NSCLC). Methods and Materials A total of 24 NSCLC patients were retrospectively reviewed; 12 patients received intensity‐modulated proton therapy (IMPT) and the remaining 12 received VMAT. Both plans were generated by delivering prescription doses to clinical target volumes (CTV) on averaged 4D‐CTs. The dose‐volume‐histograms (DVH) band method was used to quantify plan robustness. Software was developed to evaluate interplay effects with randomized starting phases of each field per fraction. DVH indices were compared using Wilcoxon rank sum test. Results Compared with VMAT, IMPT delivered significantly lower cord Dmax, heart Dmean, and lung V5 Gy[ RBE ] with comparable CTV dose homogeneity, and protection of other OARs. In terms of plan robustness, the IMPT plans were statistically better than VMAT plans in heart Dmean, but were statistically worse in CTV dose coverage, cord Dmax, lung Dmean, and V5 Gy[ RBE ]. Other DVH indices were comparable. The IMPT plans still met the standard clinical requirements with interplay effects considered. Conclusions Small‐spot IMPT improves cord, heart, and lung sparing compared to VMAT and achieves clinically acceptable plan robustness at least for the patients included in this study with motion amplitude less than 11 mm. Our study supports the usage of IMPT to treat some lung cancer patients.


| INTRODUCTION
Lung cancer is the leading cause of cancer death among both men and women in the United States. Non-small-cell lung cancers (NSCLC) account for about 85% of lung cancer cases. 1,2 Radiotherapy combined with chemotherapy is standard treatment for stage III NSCLC patients with unresectable tumors, but the potential toxic effects of radiation limit the feasibility for delivering adequate tumoricidal dose to targets in most patients. 3,4 With photon radiation and concurrent chemotherapy, the long-term results from RTOG 0617 reported 5-year overall survival (5-year OS) of 32.1% (standard dose arm with 60 Gy) and 23% (high dose arm of 74 Gy) for unresectable NSCLC patients. 5 The fact that dose escalation has led to worse overall survival is possibly due to higher cardiac toxicity. 4,6 The improvement of overall survival would require the minimization of incidental radiation dose to critical normal structures.
Volumetric-modulated arc therapy (VMAT) is an advanced form of intensity-modulated radiation therapy (IMRT) that can deliver a precisely sculpted dose distribution using a single or multi-arcs. 7 It has gained popularity in treating lung cancer patients due to its superior dose coverage, decreased radiation-induced pneumonitis, and shorter delivery time compared to conventional static-field IMRT. [8][9][10][11] On the other hand, due to the sharp falloff of dose deposition distal to the Bragg peak, proton therapy has great potential to provide highly conformal tumor target coverage while sparing adjacent organs at risk (OARs), such as heart, lungs, spinal cord, and esophagus. 12,13 Proton therapy is used in three different modalities: passive-scattering proton therapy (PSPT), uniform scanning proton therapy (USPT), and intensity-modulated proton therapy (IMPT).
Recently, Chang et al. 14 published a phase 2 study of high dose PSPT (74 Gy [RBE]) and concurrent chemotherapy for unresectable stage III NSCLC. They reported 5-year OS of 29% with very low rates of toxicities. It seemed that high dose PSPT tended to have better 5-year OS than the high dose photon therapy, but still slightly worse outcomes than the standard dose photon therapy if we compared this clinical trial data to RTOG 0617. Therefore, they suggested the use of IMPT to further improve the dose conformality and reduce doses to nearby OARs. 15,16 Unfortunately, IMPT is subject to increased uncertainties for moving targets compared with PSPT and USPT. [17][18][19] Previous studies used proton pencil beam machines with in-air sigma at the isocenter as large as 6~15 mm (depending on proton energy) to treat NSCLC cancer. 15,16 In this study, we defined these machines as large-spot proton machines compared to the proton pencil beam machines with in-air sigma at the isocenter of 2~6 mm (depending on proton energy), which we defined as small-spot proton machines for the purpose of this study.
There is a concern that IMPT with small-spot size may not be a good option for lung cancer treatments with large motions, due to the concerns of uncertainties and interplay effects. 20 A study by Chang et al. suggested that thoracic malignancies with tumor motion larger than 5 mm may not be safely treated using IMPT. 21 Other studies suggested that IMPT treatment may be used for tumors with motion larger than 5 mm, but it would be negatively impacted by interplay effects, especially for small-spot IMPT. [22][23][24][25][26][27][28][29][30][31][32][33] There are some studies that reported the limited impact of uncertainties and interplay effects in robustly optimized IMPT for stage III NSCLC. 34,35 There are no reports about dosimetric comparison between small-spot IMPT and VMAT for NSCLC patients in term of plan quality, plan robustness in the face of uncertainties, and interplay effects.
IMPT with small-spot sizes has been used to treat non-moving targets for years. However, for moving targets such as lung cancer, previous researchers did demonstrate that small-spot IMPT could improve the treatment plan quality. 36 However, a simulation study showed that small-spot IMPT (σ: 2~4 mm) could be less robust toward motion and interplay effects than large-spot IMPT (σ: 8~17 mm). 28 Larger number of spots will be needed to cover the same target volume if small-spot proton machine was applied, which was also reported in a recent study. 37 In the same study it was stated that interplay effects should be considered before IMPT treatment plan was delivered to lung cancer patients. 37 Majority of the new proton centers being developed are equipped with spot scanning beam with small-spot size (in-air sigma at the isocenter as large as 2~6 mm) only. Currently, there are no studies sharing clinical experience in radiation oncology community concerning the treatment of stage III NSCLC patients with small-spot IMPT. In this study, we reported the procedure implemented at our institution for small-spot IMPT in the treatment of NSCLC patients. The study focused on the evaluation of plan quality, robustness and interplay effects, and compared the dosimetric parameters of small-spot IMPT and VMAT.

2.A | Patient selection
We retrospectively reviewed 12 unresectable stage III NSCLC patients treated with IMPT consecutively between March 2016 and June 2017 at our institution. In addition, we retrospectively reviewed 12 selected stage III NSCLC patients treated by VMAT in the same time period at our institution. All plans used in this work were the clinically applied.
The patients included in this study were carefully selected by experienced physicists from the existing database of treated patients to ensure that the patients from the two treatment groups did not show significant differences in age, motion amplitude, or prescription doses (Table 1). However, the tumor size of patients treated by IMPT was significantly larger than that of patients treated with VMAT. All patients were staged using PET/CT and brain CT scans to rule out metastatic disease. All patients had an Eastern Cooperative Oncology Group (ECOG) performance status ≤2 and were definitively treated with radiation therapy with curative intent. None of the patients had implanted cardiac devices.

2.C | Target and normal tissue definition
Treatment targets were defined as follows. Co-registration with contrast enhanced CT scans and/or PET scans were used in identifying the gross target volume (GTV). The internal gross target volume (IGTV) was designed to encompass the extent of GTV motion in all phases of 4D CT. The clinical target volume (CTV) was formed by isotropic expansion of the IGTV by 5-10 mm (typically 7-8 mm).
The value of margin expansions were based on the pathology of tumors and determined by experienced radiation oncologists. The CTV was adjusted based on patterns of potential tumor extent and anatomic boundaries such as vertebral body, chest wall, and heart, etc. Planning target volumes (PTVs) formed by 5 mm uniform expansion of CTVs were used for plan optimization and evaluation in VMAT. All normal tissues were contoured on the 4D averaged CT.
CT artifacts were overridden using HU values sampled nearby.

2.D | Treatment planning
IMPT treatment planning generally followed the treatment planning guidelines recommended by the Particle Therapy Co-Operative Group (PTCOG) Thoracic and Lymphoma Subcommittee. 20 The proton beam scanning machine for IMPT treatment was commissioned to have an energy-dependent spot size (in-air σ) of 2 mm to 6 mm and a fixed spot spacing of 5 mm was chosen in treatment planning.
Discrete proton energies (from 71.3 to 228.8 MeV) were selected to minimize the ripple in the spread out Bragg peak (SOBP) dose distributions along the beam direction. VMAT treatment was administered using CLINAC machines (Varian Medical System, Palo Alto, CA, USA).
All IMPT plans were generated on the averaged 4D CT with IGTV density override (HU = 50). During the initial spot arrangement, an additional 7 mm margin expansion based on the PTV was used in the IMPT planning to ensure that there was at least one spot outside of the PTV to generate a possible homogeneous dose distribution within the PTV.
In most cases, two or three beams were used in IMPT. Beam directions in IMPT were chosen by dosimetrists with the help of experienced physicists if needed to minimize the impact of motion and spare normal tissues. Ten of twelve IMPT plans required single field optimization (SFO). If the SFO plan could not meet dosimetric and robustness requirements, a multiple field optimization (MFO) plan using robust optimization was generated. The final plan was chosen by an experienced radiation oncologist after careful evaluation of plan quality, plan robustness, and interplay effects.
For IMPT plans, two verification plans were generated by recalculating the dose on the exhale and inhale 4D CT phases (without the density override) to evaluate the impact of respiratory motion.
The original plan was adjusted until the verification and original plan dose distributions met all the required dose volume constraints (Table 2), plan robustness quantification thresholds, and the prescription criteria (see 2.F subsection).
In VMAT treatment planning, PTV was used for plan optimization. We applied photon optimizer (PO) model in the Eclipse™ for VMAT optimization, and analytical anisotropic algorithm (AAA) model for dose calculation. For target coverage, PTV high V 100% was at least 95% of prescription dose, and PTV high D 0.03 cc was not more than 110% of prescription dose. Most commonly, two or three arcs were used.

2.E | Plan quality evaluation
We calculated CTV D 95% , D 5% (the dose level covering at least 95% and 5% of the structure volume with the highest dose respectively), T A B L E 1 Patient characteristics between the two treatment groups.

2.F | Robustness quantification
To evaluate the robustness of IMPT and VMAT plans, we used the The field information and delivery durations of IMPT and VMAT plans can be found in the supplementary material (Tables S2 and S3).
Iso-layer repainting was used to mitigate the impact of interplay effects. 35,37,39 If the respiratory motion amplitude was less than For IMPT plans, we developed software to calculate the dose under the influence of interplay effects. 29,30,40,41 In the software, time-dependent spot delivery parameters, 4D CTs, and the time spent in each phase during the 4D CT simulations were used 39,40,[42][43][44] to calculate the dose delivered in a patient with interplay effects considered.
We randomized the starting phase of each field per fraction to effectively mitigate the impact of the starting phase. 40 (Table 4).

3.C | Interplay effect
Interplay effects were only considered for the IMPT plans as shown in Figs. 3(a)-3(d). The median values of CTV D 95% , D 2 cc , and D 5% -

CONF LICTS OF INTEREST
The authors declare no conflict of interest.

SUPPORTING INFORMATION
Additional supporting information may be found online in the Supporting Information section at the end of the article. Table S1. Comparison of the doses to non-target tissues between small-spot IMPT and large-spot IMPT.  Data S1. Comparison of IMPT plan quality with different spot sizes.
Data S2. Comparison with reported large spot size IMPT results.