Room: Exhibit Hall | Forum 3
Purpose: Treating deep-seated bulky tumors (> 8 cm) with traditional single-field Cerrobend grid blocks have major limitations such as suboptimal target dose and potentially excessive skin toxicity. Heavy traditional grid blocks are a concern for patient safety at various gantry-angles and dosimetric detail is unavailable without a grid-template in userâ€™s planning system. Herein, we propose a simple MLC-based 3D-crossfire technique to provide sufficient target coverage and reduce skin dose.
Methods: Fourteen patients (multiple sites) who underwent conventional single-field grid-therapy were re-planned using a simple MLC-based 3D-crossfire method. Gross tumor volume (GTV) was used to generate a lattice structure of 1 cm diameter and 2 cm center-to-center grid-pattern, mimicking conventional grid-patterns, using commercially available DICOManTM tool. 3D-crossfire grid plans were generated using 6-gantry angle positions at 60-degree spacing with differentially-weighted 10MV/18MV beams in Eclipse using Acuros-XB. Standard Millenium120 (Varian) 5 mm MLC-leaves were fit to grid-patterns with 90-degree collimator rotation. Prescription was 15 Gy in one fraction. Dosimetric parameters evaluated include: mean GTV dose, GTV dose heterogeneities (valley-to-peak dose ratio, VPR) and the skin dose. Additionally, number of monitor units and beam-on time was recorded.
Results: All 3D-Grid plans exhibited excellent target coverage with mean Grid-GTV dose of 13.5 Â± 0.5 Gy and mean VPR of 0.5 Â± 0.1. Comparison with traditional plans showed max and 5 cc skin doses were reduced, on average, by 3.4 Â± 4.0 Gy and 5.3 Â± 2.8 Gy, respectively. Average beam-on time was reduced by 20%.
Conclusion: The simple MLC-based crossfiring grid-therapy approach described here resulted in enhanced target coverage for deep-seated bulky tumors and reduced skin toxicity. This technique can be easily adopted into any radiotherapy center. It provides detailed dosimetry and safe, effective treatments by eliminating the heavy, physical grid-block. Future research includes generating an optimized MLC-based grid-template and further improving the VPR.