Customized double‐shell immobilization device combined with VMAT radiation treatment of basosquamous cell carcinoma of the scalp

Abstract Malignancies with a superficial involvement of the scalp/skull present technical challenges for radiation‐treatment‐planning, such as achieving skin coverage with the prescribed dose and with the desirable conformity, homogeneity, and lower brain dose. We report a radiotherapy treatment technique for a patient diagnosed with diffuse basosquamous cell carcinoma of the scalp and adjacent skull‐bone. This study presents the plan's quality parameters, patient's dosimetry, and patient's outcome. The patient was treated using volume‐modulated‐arc therapy (VMAT) and a double‐shell‐bolus full‐head device (DSBFD) designed for patient immobilization and better skin coverage. A VMAT plan was generated using an Eclipse treatment‐planning system for a prescribed dose of 60 Gy in 30 fractions. The treatment plan was analyzed to determine the conformity index (CI), the homogeneity index (HI), the target‐coverage, and the dose to the organs‐at‐risk (OARs). Skin‐doses were measured using optically stimulated luminescence (OSL) dosimeters. Clinical follow‐up was performed by the radiation oncologist during and after the course of radiotherapy. With regard to planning target volume (PTV) coverage, the V 95 was 99%. The measured and calculated dose to the skin was in the range 100–108% of the prescribed dose. The mean brain‐PTV dose was 711 cGy. The CI and HI were 1.09 and 1.08, respectively. The mean positioning accuracy for the patient over the course of treatment was within 2 mm. The measured accumulated skin dose and planning dose was agreed within 2%. Clinical examination of the patient 6 months after radiotherapy showed good response to the treatment and a 90% reduction in scarring. The DSBFD technique combined with RapidArc treatment was useful in terms of the target dose distribution and coverage. Daily patient alignment was found very precise, reproducible and less time‐consuming.


| INTRODUCTION
Radiotherapy is the only treatment option for basosquamous cell carcinoma of the scalp when surgery is not recommended, such as in cases of large lesions that involve the skull and brain, or owing to cosmetic and reconstructive issues, or when the patient is unwilling to undergo surgery. These extensive scalp lesions present a challenge for radiation therapy because of the convex shape of the target and its proximity to critical structures.
Traditionally, electron-beam techniques have been used for total scalp irradiation. [1][2][3][4][5][6] Electron beams deliver high-surface doses and exhibit rapid dose falloff, which can decrease the radiation exposure of underlying brain tissues; however, their dose distributions are inhomogeneous. Mold brachytherapy has also been widely used for the treatment of superficial skin tumors. [7][8][9][10][11] Both brachytherapy and electron-beam techniques are effective for superficial diseases, but may not provide suitable coverage for the infiltrated disease to the skull and brain due to the heterogeneous nature (i.e., the scalp, skull, and brain) and potentially large depth of the target.
In comparison to static and arcing electron techniques, intensity-modulated radiation therapy (IMRT) has shown 12 superior target coverage of the treatment volume owing to its ability to produce concave or convex dose distributions; however, this method delivers high doses to the brain and eyes. Coplanar Rapi-dArc has been used to treat scalp disease 13 and achieved a better dose conformity than 9-field IMRT. Hu et al. 14 explored the feasibility and efficiency of RapidArc in total scalp irradiation using the target-splitting technique and noncoplanar volume-modulated arc therapy (VMAT) planning, and suggested that the method provided a more promising solution than fixed-beam IMRT; however, conventional RapidArc technique has shown high-mean dose to the brain and low coverage at the skin. 14 A dosimetric comparison between helical tomotherapy (HT) and conventional VMAT showed that HT allows the best target coverage and conformity with the lower mean dose to the brain. 15 Most of the published articles with tomotherapy [16][17][18] techniques claim good target coverage and lowermean dose to the brain in the scalp/skull radiotherapy treatment compared to conventional VMAT and IMRT. The linac-based VMAT which is available almost in every cancer centers has greater potential to improve skin coverage, dose homogeneity, conformity, and reduce the mean dose of the brain with the application of 1-cmthick bolus. Because the patient's skin was included in the clinical target volume (CTV), the physician aimed at the delivery of 100% of the prescribed dose to the skin; this goal was difficult to achieve with the desired dose conformity and homogeneity without the bolus application. Initially the patient's CT simulation was decided for conventional bolus but on the day of the CT simulation, there were several complexes faced and realized including (a) it was difficult to flash the bolus over the whole scalp, (b) placement reproducibility and accuracy every day treatment, (c) patient alignment issue, and (d) time-consuming.
A double-shell-bolus full-head device (DSBFD) was designed for patient-immobilization and better coverage of the CTV especially skin. Here, we report a linac-based radiotherapy treatment technique, VMAT, combined with the DSBFD for a patient diagnosed with diffuse basosquamous cell carcinoma of the scalp and adjacent skull bone. It is worth mentioning to report that the DSBFD device is consistently reproducible, precise, and take short time in every day placement compared to conventional bolus.
Furthermore, the skin dose must be verified with a reliable dosimeter even if 100% skin coverage is predicted by the treatment planning system (TPS). In addition, the patient's daily positioning accuracy during treatment is very important because of the tight margin created by the inclusion of skin in the planning target volume (PTV) and the close proximity of the PTV to the brain.
We present the applied VMAT-based radiotherapy method and the design and manufacture of a head immobilization device, DSBFD, used to achieve 100% dose coverage at the skin and reduce dose to the OAR especially mean dose of the brain. Furthermore, we achieved good planning quality parameters, including the conformity index (CI), homogeneity index (HI), PTV coverage (V 95 , D 95 ), and doses to the organs-at-risk (OARs). In addition, we provide the patient's dosimetry results, including daily and accumulated skin dose measurements with optically stimulated luminescence dosimeters (OSLDs), and discuss in detail the impact of the DSB device on target coverage and patient positioning. Finally, we report on the radiation oncologist's observations of the patient during and after radiotherapy and on the patient's outcome according to the results of the post-treatment MRI.

2.A | Patient history and diagnosis
A 53-year-old woman presented in our department with a typical history of an old burn on her head that occurred at the age of 5 and was followed by skin grafting at that time. | 85 basosquamous cell carcinoma. CT of the head and neck performed prior to treatment showed focal skin thickening and ulceration of the left higher scalp region with an underlying aggressive bony lesion of the parietal bone that was consistent with the patient's squamous cell carcinoma. The metastatic workup was negative. The lesion was inoperable owing to its large size and its adherence to bone, and a complete excision with wide negative margins was not achievable in this location. As a result, the patient was scheduled for external-beam radiotherapy with 60 Gy, to be delivered in 30 fractionations. The patient tolerated the treatment well and showed no signs of severe toxicity. response to systemic chemotherapy and radiotherapy. 19 However, radiotherapy has been shown to be useful in the inoperable primary or recurrent tumors.

2.C | Design and construction of immobilization device
The main objective of the immobilization device for the patient's head was to achieve 100% skin coverage and ensure an accurate and reproducible daily setup. Another aim of this design was to treat the patient in the prone position, owing to the location of the lesion.
The device comprised unique anterior and posterior shells designed to cover the patient's head and secure the 1-cm-thick bolus flush against the target area. The design and fabrication process of the shells are described below.

2.C.1 | Anterior shell (mask)
This shell was used for the immobilization of the face when the patient was in the prone position. The mask covered the face from the forehead to the chin and extended laterally to the level of the ears ( Fig. 1).
i The patient was placed in the supine position on a comfortable headrest that was selected to fit her head and neck as shown in superior sternal notch. Anterior scout images were acquired using a CT scanner to confirm the position. The midline was marked to guide the therapist during the shell construction process and ensure the patient's alignment was maintained.
iii A three-point shell was placed in a hot water bath until translucent.
Subsequently, the shell was patted dry and stretched over the patient's face. A good molding over the entire face was ensured and special care was given to the forehead, nose, chin, and cheeks.
Bolus was sandwiched between the shells at the forehead area.
iv After the mask was set, it was removed from the patient.
v The mask was cut to include the entire face area to the level of the ears.

2.C.2 | Posterior shell
This shell was designed for the immobilization of the posterior aspect of the head and neck area and was used as a repositioning guide during treatment.
i The anterior shell was placed on the prone headrest. The patient was then placed in a prone position with her arms by her sides. A mattress and an ankle rest placed under her ankles were used for comfort. The posterior shell is shown in Fig. 2.
ii A posterior scout image was obtained to verify the patient's alignment.
iii The anterior mask was fixed to the prone headrest (a glue gun was used).
iv A three-point shell was placed in a hot water bath until translucent. The shell was then patted dry and stretched over the patient's posterior head and neck area. A good molding of the shell over the neck and skull area was ensured.
v When the posterior shell was set and completely dried, the bolus was added. The thickness and area of the bolus were determined by the radiation oncologist and verified by the medical physicist.
During application, a better attachment to the shell was achieved by removing the skin of the bolus and placing its sticky side facing the shell. We ensured that there were no gaps or overlap resulting from the irregular shape of the area to be covered.
vi A thermoplastic sheet was placed in a hot water bath. Next, the sheet was positioned over the bolus and attached to the posterior shell (the bolus was permanently sandwiched between the shells). As advised by the medical physicist and the radiation oncologist, the therapist performed a CT scan of the patient in the treatment position (see Section 2.D).

2.D | CT scanning and simulations
CT simulation of the patient is required for radiation dose calculations.
The patient was immobilized using the DSB device. The patient was scanned in the prone position using a Siemens CT scanner (Somatom Sensation Open, Siemens Medical Solutions, Erlangen, Germany) under the standard brain imaging protocol (120 kVp and 120 mAs).
Images were reconstructed with a voxel size of 1 × 1 × 2 mm 3 . 2.E | RapidArc treatment planning and delivery The delineation of the CTV and the PTV was performed by the radiation oncologist using the Eclipse TPS (version 11.0, Varian Medical Systems, Palo Alto, CA, USA). For the CTV-to-PTV expansion, a 3 mm margin was applied. The OARs, including the eyes, lenses, optic nerves, optic chiasm, and brain-PTV, were delineated by the medical physicist and reviewed by radiation oncologists.
The CT images and OAR contours are shown in Fig. 3(a). The

2.F | Plan evaluation
The approved treatment plan was evaluated according to standard planning quality parameters, including the CI, HI, dose-volume histograms (DVHs), the mean dose to the skin in the PTV, V 95 , D 95 , the mean brain dose and doses to the OARs (eyes, lenses, optic nerves, optic chiasm, and brain). Dose fall of characteristic beyond the PTV extending into normal tissue was evaluated from the DVH curve of the Brain-PTV structure. For the dose fall off trend, the parameters including D 5% and D 20% for the Brain-PTV structure were closely looked by the RO during the plan approval.

2.F.1 | Conformity index (CI)
The CI of the treatment plan was defined using the following equation 13 : where V PTV is the treatment volume of the identified isodose, V TV is the volume of the PTV, and TV PV is the volume of the PTV inside the identified isodose.

2.F.2 | Homogeneity index (HI)
The HI was defined using the following formula: where D 98% , D 50% , and D 2% are the doses received by 98%, 50%, and 2% of the PTV volume, respectively.

2.H | Patient dosimetry with OSLDs
As the disease was most severe and intense at the skin of the patient, skin dose monitoring was very important. In vivo dosimetry for radiotherapy patients is necessary to verify that the dose deliv-

3.A.1 | PTV coverage
The 95% of the prescribed dose distribution in colorshad is shown in Fig. 3(b). The conformity and homogeneity indices for the VMAT plan were 1.09 and 1.08, respectively. Quantitative evaluations were performed using cumulative DVH analysis (Fig. 4). The dose was received by 1% volume of the PTV (D 1% ) values of 6240 cGy (104%). The V 95 was 99% and the PTV mean dose was 6075 cGy.
These dosimetric parameters show that the dose distribution within the PTV was very uniform, homogenous, and conformal. The mean skin dose measured within the PTV was 6239 cGy. From the brain-PTV curve (Fig. 5), the dose as a function of volume reduces sharply; this is a clear indication of the sharp dose reduction from the edge of the target to the normal tissues. The parameters including D 5% and D 20% for the Brain-PTV structure were 2946 and 993 cGy, respectively. The good dosimetric outcome is attributed to the DSB device because the optimizer software required less effort to cover the skin. The application and implementation of the 1-cm-thick bolus DSB reduced the mean brain dose and improved the dose distribution within the PTV while keeping the mean brain dose reasonably lower.

3.A.2 | Doses to OARs
The plan quality parameters including PTV, OAR, and normal tissues are listed in Table 1.

3.C | Patient dosimetry with OSLD
The OSLD-A reading obtained after each fractionation is plotted versus the number of fractions in Fig. 6. The average measured skin dose for location A was 204 ± 4.4 cGy per fraction. The plot shows that the skin dose at location A measured throughout the treatment was higher than the prescribed dose by~2%. Since, the tumor was at the skin, the objective was to irradiate the skin with 100% pre-    One of the major factors affecting the accuracy of treatment is patient setup error; therefore, the use of an immobilization device is imperative. The immobilization device must provide adequate rigidity to ensure maximum immobilization, the alignment accuracy of our immobilization device was measured to be within 2 mm throughout the treatment.
The doses at the skin lesion measured with OSLDs, including daily fractions and total accumulated fractions, was in good agreement with the calculated dose. The daily dose measurements at the skin lesion throughout the 30 fractions showed more than 100% of the prescribed dose. The agreement between the total measured accumulated dose and the calculated planning dose was within 2%.
The daily patient dosimetry resulted in greater confidence that the skin lesion was properly aligned and received the prescribed dose. It is worthy to mention that the CBCT doses (approximately 0.35% of the prescribed dose) were ignored because of its low magnitude compared to the prescribed therapeutic dose. The patient outcome, including lesion resolution and disease control, was reasonably good.
The difference between the lesion/tumor size before and after radiotherapy can be observed from the pre-and post-radiotherapy images in Fig. 8(a)-(c).

| CONCLUSIONS
The use of the DSB device was determined to significantly aid with the coverage of a complicated (concave) shape and a superficial target (skin lesion). Our objective of 100% coverage of the skin was successfully achieved. Furthermore, the implementation of the DSB improved dose homogeneity and conformity in the PTV and showed reduction in mean dose to the brain compared to the published data. [13][14][15][16][17][18][19]22,23 The patient alignment accuracy in all three directions was within 2 mm throughout the treatment.
The dose measurements at the skin lesion with OSLDs, including daily and total accumulated doses, were useful for determining dose accuracy and the level of agreement with the planned dose. Also, it gives a strong clue regarding the patient alignment throughout the treatment duration in the form of dose accuracy at the skin. The daily dose measurements at the skin lesion were achieved 100% of the prescribed dose throughout the treatment. The agreement between the measured dose and the planning dose was within 2%.

| DECLARATIONS
Ethics approval and consent to participate

Consent for publication
Available on request.