Modified Dynamic Conformal Arcs With Forward Planning for Radiosurgery of Small Brain Metastasis: Each Double Arc and Different To-and-Fro Leaf Margins to Optimize Dose Gradient Inside and Outside the Gross Tumor Boundary

Abstract

Dynamic conformal arcs (DCA) are a widely used technique for stereotactic radiosurgery (SRS) of brain metastases (BM) using a micro-multileaf collimator (mMLC), while the planning design and method considerably vary among institutions. In the usual forward planning of DCA, the steepness of the dose gradient outside and inside the gross tumor volume (GTV) boundary is simply defined by the leaf margin (LM) setting to the target volume edge. The dose fall-off outside the small GTV tends to be excessively precipitous, especially with an MLC of 2.5-mm leaf width, which is predisposed to the insufficient coverage of microscopic brain invasion and other inherent inaccuracies. Meanwhile, insufficient dose increase inside the GTV boundary, i.e., less inhomogeneous GTV dose, likely leads to inferior and less sustainable tumor response. The more inhomogeneous GTV dose is prone to the steeper dose gradient outside the GTV and vice versa. Herein, we describe an alternative simply modified DCA (mDCA) planning that was uniquely devised to optimize the dose gradient outside and inside the GTV boundary for further enhancing and consolidating local control of small BM. For a succinct exemplification, a 10-mm spherical target was assumed as a GTV for DCA planning using a 2.5-mm mMLC. The benchmark plan was generated by adding a 0-mm LM to the GTV edge by assigning a single fraction of 30 Gy to the isocenter, in which the GTV coverage by 24 Gy with 80% isodose surface (IDS) was 96%, i.e., D_96%, while the coverage of GTV + isotropic 2 mm volume by 18 Gy with 60% IDS was 70%, with the D_98% being 12 Gy with 40% IDS, viz., too steep dose fall-off outside the GTV boundary. Alternatively, the increase of LM with or without decreasing the isocenter dose enables the increase of the GTV + 2 mm coverage by 18 Gy while resulting in an inadequate GTV dose with either a less inhomogeneous dose or an excessive marginal dose. Meanwhile, in the newly devised mDCA planning, every single arc was converted to a double to-and-fro arc with different LM settings under the same spatial arrangement, which enabled GTV + 2 mm volume coverage with 18 Gy while preserving the GTV marginal dose and inhomogeneity similar to those for the benchmark plan. Additionally, the different collimator angle (CA) setting for the to-and-fro arcs led to further trimming of the dose conformity. The limitations of general forward planning with only adjusting the LM for every single arc were demonstrated, which can be a contributing factor for local tumor progression of small BM. Alternatively, the mDCA with each double to-and-fro arc and different LM and CA settings enables optimization of the dose gradient both outside and inside the GTV boundary according to the planners' intent, e.g., moderate dose spillage margin outside the GTV and steep dose increase inside the GTV boundary.

Keywords: brain invasion; brain metastasis; dose distribution; dose gradient; dose inhomogeneity; dynamic conformal arcs; forward planning; micro-multileaf collimator; small tumor; stereotactic radiosurgery.

Figure 1. Benchmark plan: Arc arrangement, target definition, and dose distributions.

The images show an arc arrangement (A); target definition (B, C, F); an enlarged view of B (C); dose distributions (D, E, G, H); coronal views (A, F-H); and axial views (B-E). (A) The arc arrangement consists of one coplanar arc and two non-coplanar arcs with each arc length of 140º and collimator angle of 0º, unless otherwise specified, to divide the cranial hemisphere evenly. (C, F) The gross tumor volume (GTV) was defined as a sphere of 10 mm in diameter that was automatically generated with the treatment planning system, which is susceptible to the voxel size of computerized tomographic images with a 1.25-mm slice thickness. The GTV + 2 mm volume for evaluation is generated by adding an isotropic 2-mm margin to the GTV. In the coronal view, both the GTV and GTV + 2 mm volume shapes are awkwardly irregular (arrows in F). (D, E, G, H) Represented % isodoses are normalized to 100% at the isocenter of 30 Gy. (D, E) The dose spillage near the GTV boundary, e.g., 60 and 80% isodose lines, in the axial views is unevenly and disproportionately distributed (arrows in D) compared to those for the coronal view (G, H).

Figure 2. Alternative plan A: Dose distributions and dose-volume histograms compared to those for the benchmark plan.

The images show dose distributions (A-D); dose-volume histograms (DVH) (E); axial views (A, B); coronal views (C, D); benchmark plan (A, C); and alternative plan A (B, D). (A-D) Represented isodoses are normalized to 100% at the isocenter dose of 30 Gy for the benchmark plan, in which 26.67 Gy is assigned to the isocenter dose for alternative plan A. (E) The GTV + 2 mm volume coverage with 18 Gy isodose increased to 97% (a), and the GTV coverage with 24 Gy increased to 98% (b) in alternative plan A. However, the maximum (c) and median (d) doses of the GTV decreased in alternative plan A. The GTV + 10 mm volume represents total irradiated isodose volumes, including GTV. LM X: Leaf margin X mm; GTV: Gross tumor volume

Figure 3. Alternative plan B: Dose distributions and dose-volume histograms compared to those for the benchmark plan.

The images show dose distributions (A-D); DVH (E); axial views (A, B); coronal views (C, D); benchmark plan (A, C); and alternative plan B (B, D). (A-D) Represented isodoses are normalized to 100% at the isocenter dose of 30 Gy for the benchmark plan and alternative plan B. (E) The GTV + 2 mm volume coverage with 18 Gy increased to 95% (a), and the GTV coverage with 24 Gy become too excessive, (b) with the GTV D_98% increasing to 26.3 Gy in alternative plan B. Meanwhile, the maximum dose is preserved. LM X: Leaf margin X mm; GTV: Gross tumor volume; DVH: Dose-volume histograms; D_98%: A minimum dose encompassing at least 98% of the object volume

Figure 4. Modified dynamic conformal arcs: Dose distributions and dose-volume histograms compared to those for the benchmark plan.

The images show dose distributions (A-D); DVH (E); axial views (A, B); coronal views (C, D); benchmark plan (A, C); and modified dynamic conformal arcs (mDCA) (B, D). (A-D) Represented isodoses are normalized to 100% at the isocenter dose of 30 Gy for the benchmark plan and mDCA. (E) The GTV + 2 mm volume coverage with 18 Gy increased to 98% (a), and the GTV coverage with 24 Gy (b) and the median and maximum doses are preserved in the mDCA. (A, B) The disproportionate distributions of 60% and 80% IDS are improved in the mDCA. LM 0: Leaf margin 0 mm; GTV: Gross tumor volume; DVH: Dose-volume histograms

Figure 5. Comparative view of four differently optimized plans for dynamic conformal arcs.

The images show defined target volumes (A); dose distributions (B-E); benchmark plan (B); alternative plan A (C); alternative plan B (D); and modified dynamic conformal arcs (mDCA) (E). (B-E) Representative % isodoses are normalized to 100% at the isocenter of 30 Gy for the benchmark plan. GTV: Gross tumor volume