Improvement of conformal arc plans by using deformable margin delineation method for stereotactic lung radiotherapy

Abstract Purpose Stereotactic body radiotherapy (SBRT) is an established treatment technique in the management of medically inoperable early stage non–small cell lung cancer (NSCLC). Different techniques such as volumetric modulated arc (VMAT) and three‐dimensional conformal arc (DCA) can be used in SBRT. Previously, it has been shown that VMAT is superior to DCA technique in terms of plan evaluation parameters. However, DCA technique has several advantages such as ease of use and considerable shortening of the treatment time. DCA technique usually results in worse conformity which is not possible to ameliorate by inverse optimization. In this study, we aimed to analyze whether a simple method – deformable margin delineation (DMD) – improves the quality of the DCA technique, reaching similar results to VMAT in terms of plan evaluation parameters. Methods Twenty stage I–II (T1‐2, N0, M0) NSCLC patients were included in this retrospective dosimetric study. Noncoplanar VMAT and conventional DCA plans were generated using 6 MV and 10 MV with flattening filter free (FFF) photon energies. The DCA plan with 6FFF was calculated and 95% of the PTV was covered by the prescription isodose line. Hot dose regions (receiving dose over 100% of prescription dose) outside PTV and cold dose regions (receiving dose under 100% of prescription dose) inside PTV were identified. A new PTV (PTV‐DMD) was delineated by deforming PTV margin with respect to hot and cold spot regions obtained from conventional DCA plans. Dynamic multileaf collimators (MLC) were set to PTV‐DMD beam eye view (BEV) positions and the new DCA plans (DCA‐DMD) with 6FFF were generated. Three‐dimensional (3D) dose calculations were computed for PTV‐DMD volume. However, the prescription isodose was specified and normalized to cover 95% volume of original PTV. Several conformity indices and lung doses were compared for different treatment techniques. Results DCA‐DMD method significantly achieved a superior conformity index (CI), conformity number (CIP addick), gradient index (R50%), isodose at 2 cm (D2 cm) and external index (CΔ) with respect to VMAT and conventional DCA plans (P < 0.05 for all comparisons). CI ranged between 1.00–1.07 (Mean: 1.02); 1.00–1.18 (Mean: 1.06); 1.01–1.23 (Mean 1.08); 1.03–1.29 (Mean: 1.15); 1.04–1.29 (Mean: 1.18) for DCA‐DMD‐6FFF, VMAT‐6FFF, VMAT‐10FFF DCA‐6FFF and DCA‐10FFF respectively. DCA‐DMD‐6FFF technique resulted significantly better CI compared to others (P = 0.002; < 0.001; < 0.001; < 0.001). R50% ranged between 3.22–4.74 (Mean: 3.99); 3.24–5.92 (Mean: 4.15) for DCA‐DMD‐6FFF, VMAT‐6FFF, respectively. DCA‐DMD‐6FFF technique resulted lower intermediate dose spillage compared to VMAT‐6FFF, though the difference was statistically insignificant (P = 0.32). D2 cm ranged between 35.7% and 67.0% (Mean: 53.2%); 42.1%–79.2% (Mean: 57.8%) for DCA‐DMD‐6FFF, VMAT‐6FFF respectively. DCA‐DMD‐6FFF have significantly better and sharp falloff gradient 2 cm away from PTV compared to VMAT‐6FFF (P = 0.009). CΔ ranged between 0.052 and 0.140 (Mean: 0.085); 0,056–0,311 (Mean: 0.120) for DCA‐DMD, VMAT‐6FFF, respectively. DCA‐DMD‐6FFF have significantly improved CΔ (P = 0.002). VMAT‐ V20 Gy, V2.5 Gy and mean lung dose (MLD) indices are calculated to be 4.03%, 23.83%, 3.42 Gy and 4.19%, 27.88%,3.72 Gy, for DCA‐DMD‐6FFF and DCA techniques, respectively. DCA‐DMD‐6FFF achieved superior lung sparing compared to DCA technique. DCA‐DMD‐6FFF method reduced MUs 44% and 33% with respect to VMAT‐6FFF and 10FFF, respectively, without sacrificing dose conformity (P < 0.001; P < 0.001). Conclusions Our results demonstrated that DCA plan evaluation parameters can be ameliorated by using the DMD method. This new method improves DCA plan quality and reaches similar results with VMAT in terms of dosimetric parameters. We believe that DCA‐DMD is a simple and effective technique for SBRT and can be preferred due to shorter treatment and planning time.

Conclusions: Our results demonstrated that DCA plan evaluation parameters can be ameliorated by using the DMD method. This new method improves DCA plan quality and reaches similar results with VMAT in terms of dosimetric parameters. We believe that DCA-DMD is a simple and effective technique for SBRT and can be preferred due to shorter treatment and planning time.

| INTRODUCTION
SBRT is the delivery of a curative radiation dose to a visible gross tumor in a very precise way, using image guidance generally in 1 to 5 fractions. [1][2][3][4][5][6][7] Early studies have shown that SBRT is an effective and well-tolerated treatment for early stage inoperable non-small cell lung cancer (NSCLC) patients. [6][7][8][9] SBRT can be delivered with 4 different techniques; three dimensional conformal multiple static beams (3DC) with coplanar or noncoplanar fields, three-dimensional conformal arc (DCA), intensity modulated radiotherapy (IMRT) and volumetric modulated arc therapy (VMAT). Each method has different advantages and disadvantages.
The DCA technique widely replaced 3DC techniques with its advantage of using large number of beam directions and shorter treatment time. 10,11 Moreover, DCA plans have better conformity in three-dimensional complex target volume shapes, converging to quasi-sphere form can result because of better DCA conformity than 3DC plans. 12 Moreover, since the dynamic field shape encompasses the target volume, DCA can avoid interplay effect because of shorter delivery time and continuous dynamic field openings during treatment delivery. 12 Despite the interplay effect concern of intrafractional target volume motion, coplanar and noncoplanar inversely optimized IMRT techniques are also used safely in SBRT treatments. [13][14][15] However, it is largely replaced by VMAT due to the shorter treatment delivery time and improved target dose conformity. [16][17][18] Recent removal of flattening filter from the beam generation module increased dose rates 2.5 to 4 times for different photons energies. This led to significant shortening of the treatment delivery time for both DCA and VMAT techniques. 19,20 FFF-based techniques recently became a standard treatment for SBRT. [21][22][23][24][25] It has also been shown that VMAT-FFF has led to better conformity parameters with shorter treatment delivery time than 3DC, DCA, IMRT, and VMAT techniques. [22][23][24][25][26][27] There are similarities between DCA and VMAT techniques. Both techniques use arc method, and treatment times are significantly short. VMAT technique results in better conformity due to use of inverse optimization method during planning but with the cost of a longer time for planning process and quality assurance. However, it is easy to generate DCA plan but difficult to achieve high dose conformity for complex shaped target volumes compared to VMAT.

2.C | VMAT
The VMAT plans were created, using commercial RapidArc â module in Eclipse TM TPS with progressive resolution optimization (PRO3) (v. 13.6.2) solution. The PRO3 module was mainly based on direct aperture optimization approach varying with multileaf collimators (MLC), gantry speed and dose rate on each control point (CP). 38 The PRO3 module proceeded through four phases at the same time.
The full collection of 178 CPs was optimized in all four phases while Collimator rotation angles of arcs were 10°and 350°, respectively, with 0 mm MLC margin to the outline of PTV. Arc entrance through the contralateral healthy lung was restricted as much as possible.
AAA (v 13.6.2) was used in order to obtain three-dimensional dose distributions for evaluation of 6 MV and 10 MV with FFF plans.  Fig. 1).

2.D |
Firstly, a conventional DCA plan with 6FFF was generated and 100% of prescription isodose line covering the 95% V PTV was normalized. The prescription isodose lines were specified with covering the 95% volume of PTV (V PTV ) which were normalized to 70%-85% of isodose.
We identified dose regions outside the PTV receiving doses over

2.F | Conformity index (CI)
The RTOG conformity index is defined as ratio of prescription isodose volume (V Rx ) to the PTV volume. 33 Ideal value of CI is unity and generally it is greater than one.

2.G | Conformity Paddick index or conformity number (CI Paddick )
A new conformity index (CI Paddick ) was proposed by Paddick 39 as it does not produce false perfect scores. 43 CI Paddick denoted as where TV, PIV are target volume and prescribed isodose volume, respectively, and TV PIV is the volume of target covered by prescription isodose. 39  equal to the prescribed reference dose. 44 Ideal value of CI Paddick is unity and generally less than one.

2.H | Gradient index (GI)
The ratio of 50% prescription isodose volume to the PTV volume is R 50% . 33 2.I | Intermediate dose spillage location at 2 cm (D 2 cm ) Intermediate dose spillage location is defined as the maximum dose in percentage of dose prescribed at 2 cm away from PTV in any direction (D 2 cm ). 33

2.J | Homogeneity index (HI)
The dose homogeneity of PTV, 41 is described as where D 2% , D 50%, and D 98% are the dose values by 2%, 50% and 98% volume of PTV, respectively

2.K | External index (EI)
The external index describes the exposure ratio of health tissue, 35,42 described as: where PI is prescription isodose, V PI denotes total tissue volume received prescribed dose and PTV PI denotes planning target volume received prescribed dose.

2.L | Organs at risk dosimetric evaluation
Volume of 20 Gy, 2.5 Gy and mean dose of lungs (V 20 , V 2.5 , and D mean ) were investigated. As previously described, the tumors inves-

| RESULTS
The

| DISCUSSION
SBRT has been shown to be a precise and efficient dose delivery method for early stage lung cancer. Still, there is significant variability in terms of treatment techniques among institutions worldwide. [6][7][8][9]35 Historically, static 3DC treatment was one of the first techniques used in lung SBRT. 1

| CONCLUSIONS
This study indicates that DCA plans can be improved by using the DMD method. This method overcomes the problems of conventional DCA technique such as hot spot doses adjacent to normal tissue, nonconformal coverage around PTV and hotspot shift out of PTV.
Furthermore, DCA with DMD methods lead to similar, if not better, results in terms of dosimetric parameters in comparison to VMAT. It is strongly believed that DCA-DMD is an efficient and cost-effective technique for SBRT plans.