A Helical tomotherapy as a robust low‐dose treatment alternative for total skin irradiation

Abstract Mycosis fungoides is a disease with manifestation of the skin that has traditionally been treated with electron therapy. In this paper, we present a method of treating the entire skin with megavoltage photons using helical tomotherapy (HT), verified through a phantom study and clinical dosimetric data from our first two treated patients. A whole body phantom was fitted with a wetsuit as bolus, and scanned with computer tomography. We accounted for variations in daily setup using virtual bolus in the treatment plan optimization. Positioning robustness was tested by moving the phantom, and recalculating the dose at different positions. Patient treatments were verified with in vivo film dosimetry and dose reconstruction from daily imaging. Reconstruction of the actual delivered dose to the patients showed similar target dose as the robustness test of the phantom shifted 10 mm in all directions, indicating an appropriate approximation of the anticipated setup variation. In vivo film measurements agreed well with the calculated dose confirming the choice of both virtual and physical bolus parameters. Despite the complexity of the treatment, HT was shown to be a robust and feasible technique for total skin irradiation. We believe that this technique can provide a viable option for Tomotherapy centers without electron beam capability.

lower, which allows re-irradiation. To cover as large an area of the skin as possible, TSEBT is administered with the patient standing on a rotating platform or at several fixed positions at an extended source to skin distance (SSD) of 3-8 m using a beam degrader.
TSEBT offers good short-term remission and few reported cases of severe toxicity. 4 However, it is not possible to irradiate all the cutaneous tissue with this technique, and several patch fields are needed, raising questions regarding over-and underdosage at the field junctions. In addition, lead shielding of genitals, eyes and lips is necessary, making the technique cumbersome.
An alternative mode of treatment is total skin irradiation (TSI) with helical tomotherapy (HT), 6 a technique combining couch translation and continuous gantry rotation. With this technique, 7,8 targets as long as 135 cm can be irradiated in one field. 9 Treatment of longer targets requires the field to be split but still allowing the whole skin to be treated on one occasion. Furthermore, skin folds can be covered by defining them as target in the optimization, and organs such as the eyes, genitals and lips can be avoided. For TSI with HT, the patient can lie down in supine position during the entire treatment as opposed to standing. This technique can be of value for centers without capability of electron treatment of the entire skin, but also for partial irradiation of the skin. A few studies have previously reported on TSI with HT, 7,10,11 In this work, we evaluate the robustness of TSI with HT and implementation of virtual and physical bolus in the form of a wet suit and verify phantom data with clinical data.
The feasibility, deliverability, and assessment of robustness for the first two patients treated at our clinic is described.

2.A | Overview
Several issues regarding patient positioning, treatment planning, and delivery needed to be addressed before commencing clinical TSI.
In order to achieve a geometrically robust treatment plan, a virtual bolus was designed and applied in the optimization. To test the robustness of the treatment plan, a whole body phantom was shifted and recalculated in the planning system for several positions and verified with dose measurements. Since the dose delivery of TSI is extremely complex, given that only tangential beams are used, the dose calculation accuracy of the treatment planning system (TPS) was verified for both surface dose and scattered central dose. During treatment, the patients were fitted with a wet suit of Neoprene, which is a non-tissue equivalent material of unknown electron density and hence the bolus effect of Neoprene needed to be carefully evaluated. In vivo measurements were performed to verify the dose to the skin, on both patients and phantom.

2.B | Patient characteristics
The first patient was a 72-yr-old male diagnosed with MF 2003. He had previously received radiotherapy with kilovoltage x-ray on several occasions and had also been treated with PUVA + Methotrexate, Neotigason, and Targretin. At the time of TSI he had patches and plaques covering more than 10% of the body surface.

2.C | Phantoms and detectors
A number of phantoms and detectors were used in this study.
The density of simulated soft tissue of the phantom is 1.061 g/cm 3 , with a relative electron density of 0.975. The weight is 50 kg and the length 165 cm. The phantom includes relevant organs such as a lung cavity and a synthetic skeleton.
• A TomoTherapy phantom (Accuray Inc., Madison, WI, USA), which is a cylindrical Solid Water (RMI Gammex) phantom with varying density plugs, inserts for an A1SL ion chamber, and a removable midsection for film dosimetry.
• A Delta4 1042 cross-plane PMMA diode array detector with a density of 1.19 g/cm 3 and relative electron density of 1.16 (Scandidos, Uppsala, Sweden).
• Two separate Exradin A1SL ion chamber (Standard Imaging Inc., Middleton, WI, USA).   in head-first supine (HFS) position, and the second from the toes to the upper thigh in feet-first supine (FFS) position. Between scans, the vacuum cushion was rotated 180°and the head and neck immobilization removed. The two scans were performed with a slice thickness of 5 mm and overlapped by approximately 15 cm.

2.D.2 | Patients
The patients were immobilized and scanned following the same pro-  These structures were cropped from the PTV inwards by 5, 15, and 30 mm, where the 30-mm structure were set to completely block the fluence. This procedure prevented all except tangential beams from entering the patient/phantom, thus reducing the dose to deep-lying organs. The aim of planning optimization was to achieve the prescribed dose to cover 60% of the PTV, and a minimum of 95% of the prescribed dose would cover 95% of the PTV.
The shape of the blocking structures was modified until target coverage was deemed acceptable.
The field junction was designed to be robust for uncertainties in patient positioning. A dose gradient was achieved on both CT sets by contouring a junction structure centered at the junction markers in the longitudinal direction. We started with a 4 cm long junction structure and then adjusted the length until coverage was acceptable. The junction structure was set as a target structure, without setting the structure in use and with an overlap priority higher than any other target structure. This achieves a similar effect as cropping the PTV. In combination with optimization with fixed jaws, this pro-

2.E.2 | Patient
The prescribed dose was defined as 12 Gy in six fractions for the first patient whereas the second patient was prescribed 20 Gy in ten fractions. Planning and optimization were performed using similar planning parameters as in the phantom study, with several internal F I G . 1. The anthropomorphic whole body PBU-60 phantom, immobilized by a large vacuum cushion, with and without the wetsuit, showing the thermoplastic mask and support under the knees. Red circles mark the position of the internal reference points for the two plans and the blue line marks the position of the field junction.
blocking structures to prevent dose to internal organs such as bone marrow.

2.F | Virtual bolus
A virtual bolus was used to prevent over-optimization of the fluence in air, due to expansion of the PTV outside the body. The wet suit was replaced by virtual bolus in the optimization since the fit of the suit varied from day to day. Targets very close to the tissue-air border causes the TPS to compensate the fluence to achieve full dose in the build-up region and in the air surrounding the body. If the patient is not perfectly aligned during treatment, the patient may receive a dose well above that prescribed during treatment (Fig. 2). This can be managed by using a virtual bolus.

2.F.1 | Phantom
With the whole body phantom, optimization tests were performed in the TPS using a varying bolus density of 0, 0.4, and 1.0 g/cm 3 .
The thickness of the virtual bolus was 8 mm, that is, the PTV with an additional 3 mm margin, as suggested by Moliner. 13

2.F.2 | Patients
Although the patients were CT-scanned wearing the full wet suit, a virtual bolus of water of specified density was still added in the TPS for two reasons; to account for daily variations caused by the fit of the wet suit and secondly, to replace the unconventional bolus material of neoprene with a material of well-known dosimetric properties. The bolus was applied uniformly over the entire skin.

2.G.1 | Phantom
A 7 mm thick foamed neoprene (polychloroprene), wetsuit (AquaLung Dive, US) was used as a physical bolus for the PBU-60 phantom. A wetsuit was chosen as bolus since it can be made to cover almost the entire body, has a uniform thickness and no metal components. The wetsuit covered the entire phantom except hands, feet and head.

2.G.2 | Patients
For the patients, a hood, gloves, and socks of neoprene were also added. In addition, patient #2 had a 5 mm water equivalent bolus   One strip from each sheet was irradiated with 2 Gy at a depth of 1.5 cm in Solid Water and used as a dose reference.

Patients
To assess the patient dose to the skin at treatment, we performed in vivo dosimetry with EBT3 film at the first fraction. At least 20 film strips of 1 × 1.5 cm 2 were taped on several positions on the patients' skin. A reference irradiation was performed at 2 Gy in solid water at 1.5 cm depth with a minimum of 20 cm backscatter.

2.I.2 | Bolus measurement
The bolus effect of the neoprene wet suit fitted on the PBU phantom was quantified by paired film measurements, where film where placed beneath the wetsuit for the first measurements and replaced for the second measurement without wetsuit. In addition, a strip of film was placed on a 20 cm thick Solid Water slab and irradiated with and without a 200 × 200 × 7 mm 3 square of neoprene to measure the buildup effect of neoprene. We compared the two measured groups using Wilcoxon signed-rank test.

2.I.3 | Film evaluation
Prior to each film measurement, a strip of film from the same sheet as that used for measuring was irradiated with 2 Gy at depth of 1.5 cm in Solid Water with at least 20 cm backscatter and the TomoTherapy set in verification mode, that is fixed gantry with no couch travel. The films were scanned with an Epson 4990 flatbed scanner at least 24 h after exposure, and evaluated using the FilmQA Pro software using the reference film strip for dose normalization. The films strips were covered with a glass sheet, and scanned with a 16-bit pixel value, and 5 × 5 mm region of interests (ROIs) for averaging. The same evaluation procedure was used for both phantom and patient measurements.

2.J | Ion chamber measurements
The optimized plan, restricted to only tangential irradiation was delivered to the cylindrical Tomotherapy phantom, to verify the accuracy of the dose calculation algorithm of the TomoTherapy TPS, at depths far from the main interaction sites. The plan was optimized with the phantom surface as target, to 4 Gy per fraction, and with margins and a virtual bolus specification identical to those used for the whole body phantom. The depth dose was measured using two A1SL ion chambers at several positions in the phantom and compared to the dose calculated by the TPS.

2.K | Diode array measurements
Dose verification was also performed using the Delta4 diode array detector placed at several locations to cover the entire irradiation volume of the treatment plan. The measured dose was compared to the planned dose using gamma evaluation. 14 Quality control (QC) acceptance criterion was set to 90% pass rate using 2 mm distance to agreement, 3% dose difference, and global dose normalization.
The dose delivery across the junction was verified by irradiating both plans using the Delta4 detector in the same measurement session. For both plans, we positioned the Delta4 at the lateral and sagittal green laser position and longitudinally in the plan junction markers, due to the red to green laser separation limit of 15 cm for.
The distance from the longitudinal green laser position to the Delta4 was measured in the DQA module and applied at setup. After irradiation of the upper plan, the detector was rotated and aligned to the lasers for the lower plan and subsequently irradiated in the same measurement session. The planned dose for the upper and lower body was manually summed using Python.

3.A | Phantom
Doses to OARs are presented in Table 1. The optimization time for 500 iterations ranged between 4 and 6 h with a GPU-assisted dose calculation engine. The beam on-times for the final plan were 31 and 19 min, for the upper and lower body, respectively. In the optimization, some adjustment of the blocking structure was required to compensate for the flat back of the phantom (see Fig. 3). This adjustment resulted in higher dose to the lungs of the phantom, due to the thin thorax wall of the PBU-60 phantom (6 mm). For the patients, this was corrected for by immobilizing the back in a rounded position.
Verification of the dose to the surface of the whole body phantom using EBT3 film agreed well with the dose to the PBU-60 phantom calculated without the virtual bolus. The results indicate that the dose calculated in the TPS provides a good approximation of the delivered skin dose. When using the virtual bolus and neoprene for build-up, the average dose difference between TPS and film measurements was −0.6% (SD = 3%; Fig. 4). The paired measurements, with and without wet suit, on the PBU-60 phantom showed a significantly higher surface dose with the 7 mm neoprene bolus than irradiation without bolus (Wilcoxon signed-rank test, P < 0.05;

3.B | Patient
Based on the experience from the immobilization of the PBU-60 phantom, the patients were immobilized with the back in a laterally curved position to better facilitate tangential irradiation in the optimization.
For both patients, the six-first fractions were recalculated based on daily MVCT images and compared to the robustness calculations performed with the PBU-60 phantom (Fig. 8). In addition, in vivo film dosimetry corresponded well with dose calculated in TPS, with a mean difference from TPS of 5.3% (SD = 11.9%) and 1.5% (SD = 9.0%) for patient 1 and 2 respectively (Fig. 9) Both patients could put the wetsuit on within a few minutes, with no notable effort. The fit for patient 2 was not optimal which was compensated for by taping air gaps to achieve a snug fit. Figure 10 shows F I G . 5. Measurements with and without wet suit using electron beam therapy 3 film on the PBU-60 phantom presented as a boxand-whisker plot. The median (red line), as well as first and third quartile (box) and 1.5 times past the interquartile range (outer line) is plotted with outliers (black points). Dose is presented as percentage of prescribed dose.
skin delivery, despite the fact that neoprene is not a standard material, and the lack of water equivalence in the material. The results quantify the difference with using neoprene as bolus and stands in contrast to other studies that did not find it necessary to use bolus for skin irradiation. 8,16 The dosimetric advantage to a non-bolus treatment is clear, any attempt at optimizing or deliver photons to the skin performs better with bolus added.

4.B | Robustness
The robustness test showed that calculated and delivered dose cor-

4.C | Comparison to standard treatment
TSEBT is today regarded as the standard treatment of mucosis fungoides and in comparison, TSI using HT is a lengthy and complex Doses to organs at risk are generally below clinically used dose constraints, but a comparison with TSEBT is not possible as, to the best of the authors' knowledge, doses to organs at risk have not been published for TSEBT. Slightly higher doses are expected to deep-lying organs in TSI with HT compared to TSEBT, which is a trade-off. In contrast, high robustness and a homogeneous target coverage can be achieved on a single treatment occasion using Tomotherapy.
Similar to results reported by Buglione et al. 16 we believe TSI with TT to be a complement to electron treatment and in certain cases where treatment with Tomotherapy could be beneficial. In addition, since TSI with HT is an image guided technique, problems that may arise during treatment can be evaluated by dose recalculation or re-optimization of the treatment plan. Previously treated areas and organs at risks can be avoided, and simultaneous integrated boost to for example, plaque areas can be implemented.
The results from this study can be of use when treating patients with partial irradiation of large areas, especially of convex shape such as the scalp 20 or melanoma. 21 Furthermore, this technique may be an alternative to centers where electron therapy is not available.

4.D | Film dosimetry
The EBT3 film is an established and appropriate dosimetry system for surface dose measurement. [22][23][24] It has a low angle dependence and stable response over a wide dose and energy range, especially when used with the FilmQApro scanning system, where all color channels can be evaluated. The largest uncertainty stems from positioning accuracy, that is, the problem to correctly assess the points of measurement of the films in the TPS for correct dose comparison.
F I G . 9. Measured dose with electron beam therapy 3 film for patient 1 and 2 at the first fraction. Dose is plotted as the difference to prescribed fraction dose. Data are shown as a box-and-whisker plot, where the red line shows the mean of 23 measurements, median (red line), mean (green triangle), 1 SD (box), and 95% confidence interval (outer line) as well as outliers (black point).

4.E | Conclusions
The presented technique was shown to be feasible and robust to deliver for both phantoms and for two individual patients. We believe that TSI with tomotherapy may an alternative for centers without electron beam capability, if a more homogenous dose is desirable, or for partial skin irradiation were electron therapy for any reason is not feasible.

CONF LICT OF I NTEREST
Department of corresponding author have an ongoing research agreement with Accuray Inc. which includes funding.