A simple method to account for skin dose enhancement during treatment planning of VMAT treatments of patients in contact with immobilization equipment

Abstract Purpose The ability to accurately predict skin doses and thereby design radiotherapy treatments that balance the likelihood of skin reactions against other treatment objectives is especially important when hypofractionated prescription regimes are used. However, calculations of skin dose provided by many commercial radiotherapy treatment planning systems are known to be inaccurate, especially if the presence of immobilization equipment is not accurately taken into account. This study proposes a simple method by which the accuracy of skin dose calculations can be substantially improved, to allow informed evaluation of volumetric modulated arc therapy (VMAT) treatment plans. Method A simple method was developed whereby dose calculation is split into grid regions, each with a correction factor which determines MU scaling for skin dose calculation. Correction factors were derived from film measurements made using a geometrically simple phantom in partial contact with a vacuum immobilization device. This method was applied to two different test treatments, planned for delivery to a humanoid phantom with a hypofractionated stereotactic body radiotherapy technique, and results were verified using film measurements of surface dose. Results Compared to the measured values, calculations of skin dose volumes corresponding to different grade tissue reactions were greatly improved through use of the method employed in this study. In some cases, the accuracy of skin dose evaluation improved by 76% and brought values to within 3% of those measured. Conclusion The method of skin dose calculation in this study is simple, can be made as accurate as the user requires and is applicable for various immobilization systems. This concept has been verified through use on SBRT lung treatment plans and will aid clinicians in predicting skin response in patients.


| INTRODUCTION
As radiotherapy treatment complexity and precision increases, so does the need to ensure that patients are in reproducible and stable positions during treatment delivery. Immobilization equipment (vac bags, thermoplastic masks, belts etc.) is increasingly being used during radiotherapy treatment and commercial radiotherapy treatment planning systems (TPS) are known to produce inaccurate calculations of surface dose from megavoltage photon beams, 1-7 even without the involvement of immobilization equipment.
Immobilization equipment can produce a skin dose enhancement effect from megavoltage photon treatments due to the buildup of scattered and photo-electrons. 8 For example, skin dose from a 6 MV photon beam can be increased by up to 22% and 43% due to 2.5 and 10 cm thick vacuum bags, respectively. 9,10 Carbon fiber couch tops contribute significantly to skin dose for VMAT plans with one study reporting skin doses as high as 81% of the prescription dose in regions of couch involvement. 11 Thermoplastic immobilization for breasts has been demonstrated to produce an increase in skin dose upwards of 62%. 12 Overall, the dose enhancement effects of couches and immobilization equipment are well documented 13,14 but TPS-specific tools of management thereof are not, especially for VMAT.
Undesired dose enhancements are difficult to avoid and are especially important when VMAT treatments are used due to the substantial proportion of arc delivery through couch, vac bag, or other immobilization systems 14 effects that may have traditionally been avoided through careful selection of beam arrangement in static gantry treatments. This is exacerbated when treatments are delivered via hypofractionated regimes such as stereotactic ablative body radiotherapy (SABR or SBRT) and stereotactic radiosurgery (SRS) due to the increased fractionated dose delivery to "early responding" skin. 15 SBRT treatments have produced grade 3-4 skin reactions in lung cases 16,17 and grade 1-2 reactions in spine cases. 18 Treatment planning systems cannot, of course, provide accurate predictions of skin dose if the effects of the immobilization systems are not included. Previous studies have attempted to account for the effects of immobilization equipment through several methods not limited to (a) creating an artificial bolus structure in the TPS, 19 (b) contouring fixation devices/couch in the TPS either with an appropriate assigned density or according to their tissue-equivalent thickness [20][21][22]  was then sent to Pinnacle 3® 9.10 (Koninklijke Philips N.V., Amsterdam, the Netherlands) and a 4 × 4 cm 2 360°arc was planned and delivered to the phantom. A dose calculation grid that encompasses the whole patient was used with a resolution of 2.5 mm. 24 1 × 1 cm 2 contours were added allowing for calculation of skin dose at points of interest. These contours were mirrored laterally, to correspond to five horizontal regions, g i , in the ant/post directions, as shown in Fig. 1, that broadly represents anterior, posterior, and lateral regions as well as regions in between with and without vac bag involvement.
Eight pieces of Garchromic ™ EBT 3 film (Ashland Specialty Ingredients, Bridgewater, NJ, USA) were used to measure the surface dose at the eight locations specified in the treatment plan when the immobilized CIRS phantom was irradiated using the 4 × 4 cm 2 arc.
Care was taken to minimize air gaps between the film and phantom surface, by cutting sufficiently small (1 × 1 cm 2 ) pieces of film. After irradiation, the films were scanned on a Perfection ® V850 Pro scanner (Epson ® , North Ryde, NSW, Australia) and analyzed in the red and green channels through ImageJ (NIH, USA) using a rational dose calibration curve according to the methods of Micke et al. 25 Skin dose correction factors (c gi ) were then defined for each of the horizontal regions g i shown in Fig. 1, as the ratio of the mean surface dose measured using film, D film;gi , to the mean surface dose calculated by the TPS of the film contour, D TPS;gi The effective depth of measurement of EBT film is 0.2 mm (i.e., within the film/skin contour) and PDD correction factors can be applied to film readings for a particular depth of interest based on the buildup PDD. 7 However, the magnitude of these corrections is largely diminished when surface film is also exposed to exit dose from another beam or opposite side of a VMAT arc. 7 The corrections are also field size-dependent 7  fractions, according to recommendations of RTOG 0618. 27 Organs at risk were also contoured in a realistic manner and given constraints according to RTOG 0618. Like the CIRS phantom reference measurements, the ART phantoms skin structure was created using the established method of contouring a 5-mm thick ring of tissue beneath the external surface contour. 1,20 After the VMAT treatment plans had been created and the resulting doses calculated, the skin doses calculated by the TPS were modified according to the following procedure. The calculation grid used for the VMAT dose calculations was restricted to cover only one of the regions, g 1 to g 5 (as shown in Fig. 2), at a time. The number of monitor units (MU) used in the treatment were scaled with the correction factor identified using the film measurements on the CIRS phantom, for the region being calculated. The volumes of skin exposed, v x;gi , to doses of x Gy (in this case, 2.6, 6.5 and 13 Gy, corresponding to a Grade 1, 1-2, and 2-3 reactions, respectively, for 3 fraction SBRT 28,29 ) were recorded and summed over the calculation regions g i to calculate V x according to eq. (2). For example, if the region g i had a correction factor of c gi and the treatment had a prescription of y MU then for calculation of grid g i , the number of monitor units was changed to c gi Â y and v x;gi was extracted accordingly.

| RESULTS
Values of c gi calculated from the example 4 × 4 cm 2 reference arc are listed in Table 1 Table 2 and show no significant difference between measured and calculated dose for g 1 , g 2 , and g 3 but for g 4  Skin values of V 2.6 Gy , V 6.5 Gy and V 13 Gy for the two SBRT lung plans delivered on the ART phantom are shown in Table 3. The    can be applied to most immobilization systems and has the potential to assist in predicting and avoiding skin reactions, especially from hypofractionated VMAT treatments.

CONF LICTS OF INTEREST
The authors have no other relevant conflicts of interest to disclose.