Dosimetric Evaluation of Commercially Available Flat vs. Self-Produced 3D-Conformal Silicone Boluses for the Head and Neck Region

Background: Boluses are routinely used in radiotherapy to modify surface doses. Nevertheless, considerable dose discrepancies may occur in some cases due to fit inaccuracy of commercially available standard flat boluses. Moreover, due to the simple geometric design of conventional boluses, also surrounding healthy skin areas may be unintentionally covered, resulting in the unwanted dose buildup. With the fused deposition modeling (FDM) technique, there is a simple and possibly cost-effective way to solve these problems in routine clinical practice. This paper presents a procedure of self-manufacturing bespoke patient-specific silicone boluses and the evaluation of buildup and fit accuracy in comparison to standard rectangular commercially available silicone boluses.

Methods: 3D-conformal silicone boluses were custom-built to cover the surgical scar region of 25 patients who received adjuvant radiotherapy of head and neck cancer at the University Hospital Würzburg. During a standard CT-based planning procedure, a 5-mm-thick 3D bolus contour was generated to cover the radiopaque marked surgical scar with an additional safety margin. From these digital contours, molds were 3D printed and poured with silicone. Dose measurements for both types of boluses were performed with radiochromic films (EBT3) at three points per patient-at least one aimed to be in the high-dose area (scar) and one in the lower-dose area (spared healthy skin). Surface-bolus distance, which ideally should not be present, was determined from cone-beam CT performed for positioning control. The dosimetric influence of surface-bolus distance was also determined on slab phantom for different field sizes. The trial was performed with hardware that may be routinely available in every radiotherapy department, with the exception of the 3D printer. The required number of patients was determined based on the results of preparatory measurements with the help of the statistical consultancy of the University of Würzburg. The number of measuring points represents the total number of patients.

Results: In the high-dose area of the scar, there was a significantly better intended dose buildup of 2.45% (95%CI 0.0014-0.0477, p = 0.038, N = 30) in favor of a 3D-conformal bolus. Median distances between the body surface and bolus differed significantly between 3D-conformal and commercially available boluses (3.5 vs. 7.9 mm, p = 0.001). The surface dose at the slab phantom did not differ between commercially available and 3D-conformal boluses. Increasing the surface-bolus distance from 5 to 10 mm decreased the surface dose by approximately 2% and 11% in the 6 × 6- and 3 × 3-cm² fields, respectively. In comparison to the commercially available bolus, an unintended dose buildup in the healthy skin areas was reduced by 25.9% (95%CI 19.5-32.3, p < 0.01, N = 37) using the 3D-conformal bolus limited to the region surrounding the surgical scar.

Conclusions: Using 3D-conformal boluses allows a comparison to the commercially available boluses' dose buildup in the covered areas. Smaller field size is prone to a larger surface-bolus distance effect. Higher conformity of 3D-conformal boluses reduces this effect. This may be especially relevant for volumetric modulated arc therapy (VMAT) and intensity-modulated radiotherapy (IMRT) techniques with a huge number of smaller fields. High conformity of 3D-conformal boluses reduces an unintended dose buildup in healthy skin. The limiting factor in the conformity of 3D-conformal boluses in our setting was the immobilization mask, which was produced primarily for the 3D boluses. The mask itself limited tight contact of subsequently produced 3D-conformal boluses to the mask-covered body areas. In this respect, bolus adjustment before mask fabrication will be done in the future setting.